Baldwin Locomotive Transition –  or the lack thereof

Happy Holidays from VDD! This post was written by guest author Nathaniel Rana.

For some background, I’ve been coding a comprehensive railroad DC rotating equipment simulator which has required some in-depth technical information and research to the point I’m regularly consulting electrical engineers. I’ve gotten in the habit of learning and unlearning information on the regular, the most recent of which is the process by which Baldwin diesel locomotives with Westinghouse electrical gear handled motor back EMF. Rather than utilizing transition in addition to motor field shunts such as with GE and EMD’s rotating equipment, Westinghouse equipment in nearly all Baldwin locomotives (as well as the vast majority of Lima-Hamilton and potentially early Fairbanks-Morse low-horsepower switchers) exclusively handled motor back EMF by means of motor field shunts.

For the longest time I wondered what, if anything, made Westinghouse special. I didn’t have the means to ensure an actual answer until recently where, armed with maintenance bulletins and a better understanding of motor properties, I could finally sort this out; I had wrongly assumed that Baldwin simply took the hit to high speed tractive effort that neglecting transition would normally get you. This article will attempt to explain why Baldwin did not utilize motor transition.

Motor Background:

So, to start: Whereas EMD and GE primarily offered the same traction motor between their switchers and road power, Westinghouse used two different models for their postwar offerings. The first being the 362, smaller of the pair, was initially used on VOs and good for some 200-250 horsepower. The larger, later motor was, the 370, first used on the massive multi-engine Essl locomotive project and nominally rated for 750 horsepower, with later uprating to 900.

Fig 1 – 362D from a locomotive at SMS.

The 362 was not particularly special; it was a four pole, DC unit that was relatively light – 5,520 pounds bare motor with pinion (FM Bulletin SEC.411.5A) versus 7,000 pounds same for a GE 752 (ALCO, TP-503 1st edition). The primary use case was light road-switchers and switchers limited to 1,200 horsepower for traction and below. Of note is the high internal resistance – two to three times that of the 752 or D7 series.

Fig – 2 – 370 motor

The 370, however, had a unique quirk as far as railroad rotating equipment went; it was the only six pole DC motor used in significant capacity for US railroad diesels. Additionally, it was capable of more horsepower than the GE 752 on introduction – 750 versus 600 or so – despite similar weights: a complete motor weighed 6620 lbs, 7400 with the gear and gearcase (Westinghouse Maintenance Instruction for Traction Motor, Type 370). It is worth mentioning that this motor also has the lowest internal resistance I’ve seen so far, slightly below that of the 752.

A quick note on motor behavior – torque of a motor is a function of amperage through the armature, and the amps drawn by a motor relies on resistance (as well as internal structure but that’s much harder to quantify). Additionally, back EMF is a function of resistance and efficiency curves – manually playing with these in my simulator it became obvious that the 362 sucked down relatively few amps and put out a ton of back EMF for any voltage applied to the motor at speed. As a result, one could expect to note that applying motor field shunts or transitions would become necessary at much lower speeds than competitor’s equipment. Let’s see how that holds up in practice.

The 362:

Fig 4 – S12 switcher approximate tractive effort

Immediately of note is the bump at 21mph. Having seen these charts before, I’d assumed this to be a quirk of the drawing or just the effect of motor field shunting, since before now I only had third-party diagrams to go off of. However, one of the available modifications caught my attention:

Fig -5

It specifically calls out this speed as 22.7mph – the “unloading speed” – where full horsepower utilization of the engine is no longer possible as the generator cannot safely output any higher voltage to counteract back EMF (This is why alternator-equipped DC traction locomotives had greatly simplified transition sequences – significantly more voltage available to cram into the motors). Bizarre to see it as an option, though while they could potentially benefit from transition, the application of it was less than critical on a switcher so perhaps the S12 is a poor example. So – what about something road-capable?

The Lima-Hamilton 1200 horsepower road switcher is one of very, very few 362-equipped locomotives that also received proper transition. The full sequence is two steps of field shunting followed by motor transition as the unit accelerated, per the operations manual; maximum voltage for the generator applied to the L-H engine is similar vs. Baldwin units at 1000 volts before unloading occurs.

Fig 6 – L-H 1200hp R/S TE curve

The “unloading” speed for full parallel operation is a much more reasonable 40mph. The BLH RS-12s NYC purchased following this order behaved similarly, evidently with special-order transition. In short: 362 equipped locomotives didn’t make transition because they tended to not need the high speed TE given the intended duties, but there’s no special quirks about the motors that made it truly unnecessary. Full horsepower utilization could likely have been achieved up to 60mph with an additional step of field shunting in full parallel. This leaves us with the larger road-exclusive traction motor to look at.

The 370:

 So let’s repeat what we did above, and look for the unloading corner on TE charts – I’ll use the AS-16 here, with the date of the chart being an early 1952 model giving us a unit equipped with four steps of motor field shunting, and motors permanently wired in series-parallel similar to nearly all other four-motor Baldwin units.

Fig 7 – AS16 approximate tractive effort

No corner – maybe it’s a simplified chart? From seeing wiring diagrams of AS-616 locomotives I know that they’re permanently wired in series parallel, with two groups of motors wired in three series which should net us a higher back EMF, and there’s a chart in this manual too.

And here’s where it gets goofy.

Fig 8 – AS616 approximate tractive effort

Prominently visible at 40mph is our unloading corner, and also featured is Baldwin’s cheaty way of preserving high speed acceleration. Rather than making transition, the optional system would completely short two motor fields – one motor per parallel branch – lowering the resistance in the whole system (and neutralizing all the back emf contributed from those two motors) and, therefore, acting like another step of field shunting beyond what the motors could normally handle. We can also see that the B dashed line is significantly closer to full horsepower utilization (Though a corner can still be seen, this can potentially be attributed to the remaining armature resistance on the idled motors).

Thus we can see – the AS-16 chart is not simplified. An identical, corner-less curve is visible on the RF-16 chart; no other four motor with 370s was offered by 1952. Evidently the available maximum voltage and internal properties of the 370 made transition genuinely unnecessary on Baldwin four axle road-capable locomotives.

Thanks to Nathaniel for putting together this post. All material above was provided by the author. Any questions/comments please drop us a line and we will pass them along!

The Alaska Railroad SD70MAC-HEP, The Summer Passenger, Winter Freight Locomotive

Introduction – November 2021 – This was written by Jay Boggess in 2006 and originally titled “Adapting a Freight Loco into a Passenger Loco” and was presented at the Locomotive Maintenance Officers Association (LMOA) at their 2006 annual meeting.  Back then, he was a manager with the Alaska Railroad.  Before that, Jay worked at EMD for 22 years, which is now called Progress Rail.  He has added some footnotes and clarifications after the 15 years since this was originally published.

This paper will describe a clever adaptation of an AC freight locomotive into a dual-service passenger/freight locomotive – the SD70MAC-HEP.  It was a project driven (like many projects are) by an aggressive delivery schedule and design requirements.   By taking several existing bits of locomotive technology, a unique and useful locomotive was developed and delivered in nearly record time for The Alaska Railroad.

The SD70MAC-HEP has its lineage in 4 other EMD locomotives:  Figure 1 is the first SD70MAC, delivered in 1994 to the Burlington Northern.  Alaska Railroad purchased 16 units in a winterized configuration in March 2000, as shown in Figure 2.

Figure 1 – Burlington Northern SD70MAC
Figure 2 – Alaska SD70MAC

Figure 3 is the Long Island Railroad DE/DM30AC locomotive.  This locomotive was the first EMD production passenger locomotive with AC traction motors, Siemens GTO inverters and inverter-based Head End Power (HEP) (the EMD F69PH-AC prototype units of 1989 never went into production, although many concepts initially tried out on the F69’s ended up on the DE/DM30’s).

Figure 3 – Long Island DE30AC

Figure 4 is the CSX 4300HP SD70MAC.  This locomotive, designed and built just before the SD70MAC-HEP, put the Tier-1 emissions engine and cooling system onto a 70MAC for the first time.  During its design, the equipment at the long hood end was extensively rearranged so that space could be allocated for the CSX-requested Eco-Trans Auxiliary Power Unit (APU).  The space would later turn out to be key to the success of the SD70MAC-HEP.

These previous designs coalesced into the Alaska Railroad SD70MAC-HEP (Figure 5), a 4300 THP freight / 2400 THP-730kW HEP passenger locomotive.

I had an interesting perspective on this project: I was involved in the locomotive’s original conceptualization, design and build at EMD.  I came to work for the Alaska Railroad soon after they were delivered and then was knee-deep in the post-delivery HEP system commissioning.  Thus, when I need to complain to someone who designed this mess, all I have to do is to look in the mirror!

Figure 4 – CSX 4300-THP SD70MAC locomotive of 2003
Figure 5 – Alaska SD70MAC-HEP 4300 THP  730kW HEP

The Railroad:

The Alaska Railroad connects Anchorage (the state’s largest city) with Fairbanks, 356 rail miles north, along with the seaports of Seward and Whittier 114 and 72 miles to the south, respectively.  Freight service consists of oil trains from the Flint Hills refinery at North Pole, Alaska to Anchorage[1], coal trains from the Usibelli mine near Healy to the port of Seward[2] and interchange cars from our barge operation in Whittier.  Three trainsets of SD70MAC’s and hopper cars move aggregate from quarries in the Palmer-Wasilla area for construction in Anchorage all summer long.  Our passenger season starts in mid May and consists of a mélange of trains:

The Denali Star – Two trains in daily service from Anchorage to Fairbanks, with stops in Wasilla, Talkeetna and Denali (Mt. McKinley).  Eight ARR cars plus 8 to 10 cars owned by the cruise ship companies are pulled every day.

The Coastal Classic – One train to Seward and back, dovetailing with day cruises to Kenai Fjords National Park

The Hurricane Turn, with provides flag-stop service using RDC’s from Talkeetna to Hurricane, where a series of cabins are only accessible by rail.

Trains connecting cruise ship passengers from Whittier and Seward to the airport in Anchorage.

As tourism has increased, more cars are added, which is what started this whole project in 2001.

[1] The Flint Hills refinery shut down in 2014 and AkRR no longer moves jet fuel from Fairbanks to Anchorage.

[2] There is no longer any export coal out of Seward -just coal movements from Usibelli to Fairbanks and Eielson AFB.

HEP for our passenger trains was first handled by 2 paralleling 150-kW generator sets slung underneath the baggage car (see Figure 6).  Later in 2000 and 2001, six GP40’s were rebuilt into HEP-equipped units for passenger service.   3009/10/11 were equipped with 300kW Detroit Diesel generator sets and 3013/14/15 with ex-Amtrak 800kW gear-driven generators.  Each has their disadvantages:

The baggage car gen sets have limited capacity and susceptible to clogging with cottonwood lint.

The 300kW Geeps are noisy and again have limited capacity.

The 800kW Geeps (see Figure 7) are VERY noisy, have very small fuel tanks and are not at all fuel-efficient.

ALL the locomotives (even though recently rebuilt) are getting old.

Figure 6 – Baggage Car w/ undercar generator sets
Figure 7 – GP40-H with gear-driven HEP

As an answer to Alaska Railroad’s passenger power problem, EMD offered F59PH locos, but AkRR[1] considered them very expensive, especially for a unit that would only be needed in the summer.  Robert Stout (AkRR CMO at the time) suggested to EMD, “We like what the SD70MAC does for us – what’s the possibility of putting HEP on a MAC?”  Such a dual-purpose unit would be flexible for both the short passenger season and again useful the rest of the year as a freight locomotive.

When this suggestion wound its way to EMD Engineering, Tim Keck (EMD Manager of Systems Engineering at the time), posed that question to myself and others.  We all realized that one of the two TCCs[2] of a SD70MAC might be modified to provide HEP, leaving the other TCC and truck to propel the unit (and the other truck just coasting).  Consultations with our counterparts at Siemens confirmed that yes, EMD and Siemens could adapt the HEP transformer previously designed for the LIRR DE/DM30AC locomotive with the inverter of the SD70MAC to provide 480V 3-phase HEP.

As the original SD70MAC had no room in its carbody for the additional needed HEP equipment, several different ideas were proposed, including using the locomotive roof and running board.  The only realistic hope was to dramatically shorten the fuel tank and sling the HEP equipment underneath the underframe.  As the LIRR DE30AC HEP transformer was hung similarly under its structural carbody, this didn’t seem too bad of a situation.  The challenge would be designing cabinets for the HEP switchgear for the underslung application.  Design, cost and manpower estimates were made at EMD.  But then, ARR decided not to buy locos for that calendar year – not completely forgetting the idea, just not for that year.

In the fall of 2002, CSX enters into this story.  They came looking for new units – more SD70MACs like they already had – with 3 important differences: 

The engines were hopped up from 4000 to 4300 THP[3].

By EPA regulations, they would be required to meet Tier-1 emissions and thus needed a new split cooling system (first applied to UP SD70DC units but never applied to a SD70MAC)[4].

CSX requested space be designed in the rear of the unit for an Eco-Trans APU[5]

[1] Alaska Railroad’s reporting marks are ARR. When the internet arrived, arr was already taken, so AkRR was used.

[2] SD70MACs have two Siemens-supplied Traction Converter Cabinets (TCCs).  Each TCC is an electronic inverter that takes the DC output of the main generator and chops up the DC into 3-phase variable voltage, variable frequency AC for 3 axle-hung induction motors of one truck.

[3] The 16-710G engine of the original SD70MACs are rated at 4000 Traction Horsepower (THP) at 900 engine rpm.  By spinning the same engine to 950 rpm, 4300 THP can be realized.  Traction horsepower is the power delivered to the main generator. 

[4] The cooling system of a locomotive has to remove heat from water used to cool the diesel engine cylinders and lube oil system (“jacket water”) and the water used to cool the compressed air leaving the turbocharger (“aftercooler water”).  By utilizing two separate radiators with separate cooling fans and water pumps, the aftercooler water can be kept around 110 F while the jacket water can be kept at 180 F, improving emissions and engine performance.

[5] APU – Auxiliary Power Unit – a small diesel to keep the coolant water warm and the batteries charged when the main engine is shut down.

By some creative re-arranging of components, a volume of 3’6”x6’x5’9” was created for the APU by:

Replacing the 4-cylinder WLA electric drive air compressor with a shorter direct-drive-only WLN 3-cylinder compressor.

Rotating the #2 TCC 180 degrees so that its “blind side” faced aft (the Siemens TCC has one face with no access doors for maintenance).  This put the access door for the TCC computer in the same room as the air compressor – less than desirable but acceptable.

Moving the battery box from the conductor’s side to engineer’s side of the long hood.

As always, EMD Drafting worked quickly to deliver the necessary production drawings to the shop for a December 2003 start of construction.  CSX would eventually acquire 130 4300-THP SD70MAC’s.

With the CSX Tier-1 70MAC design complete, EMD realized that the CSX APU space could easily morph into a location for an Alaska HEP system without shortening the fuel tank and without slinging transformers and equipment underneath the loco.  This proposal was offered up to Alaska in the spring of 2003.

Now the problem became one of delivery schedule – Alaska needed locomotives by April/May 2004, so as to avoid leasing locomotives for the 2004 summer passenger season.  However, the lead time for the Siemens-supplied software and equipment for HEP prevented such a schedule (although they could easily deliver freight-only equipment for an April 2004 ship date).  Around this impasse, EMD came up with the idea – deliver the locomotives without HEP capability for April 2004, then install the HEP equipment later.  With that compromise, Alaska signed for 8 SD70MAC-HEP locomotives in June 2003 (up from the 4 units they were originally considering).  This was a very ambitious design/build schedule for Electro-Motive.

After settling details with the railroad in July 2003, it was time for EMD to translate wild-hair, back-of-the-envelope ideas into something that could actually be assembled in London (ONT) – and assembled in time.  The key problem was this: how to shove ten pounds of…stuff into an eight-pound bag!  The design process for the Alaska HEP was fraught with limited time, false starts and changing concepts.  We knew that the transformer could not be delivered until August 2004, but then we realized that the more equipment we could mount in the HEP Equipment Room in London, the less work that would be required to do in Alaska and the better for all parties involved.

The System

The final arrangement for the SD70MAC-HEP ended up as follows:

Siemens GTO TCC cabinet with modified software in Siemens ASG computer (ASG: Antreib Steuer Gerät – German for “Propulsion Control Apparatus”)[1].

Repackaged HEP transformer, based on LIRR transformer design.

Three delta-connected capacitors for AC harmonic filtering – 550 kVAR (capacitive) total.

Switchgear to switch modes between HEP and traction and to disconnect the #1 end HEP receptacles.

HEP blower and filter behind the long-hood headlight.

Main HEP contactor (ACC) and sundry HEP contactors, relays and equipment.

HEP switches, pushbuttons and fault lights located in cab on HVC[2] door.

Most of the equipment ended up squeezed into the volume that was created for the CSX APU, which was renamed the HEP Compartment.  Figure 8 is the outside of the HEP Compartment; Figure 9 is the inside and Figure 10 is a one-line diagram of the HEP system.

Figure 8 – Outside View of HEP Compartment
Figure 9 – Inside View of HEP Compartment
Figure 10 – One-Line Diagram of the HEP System

[1] When designing the first production BN SD70MACs in 1993-94, we had duplicate names: TCC for Traction Converter Cabinet and TCC for Traction Control Computer.  To avoid confusion, we decided to use TCC for the whole Siemens inverter and the German abbreviation ASG for the Siemens computer alone.

[2] HVC – High Voltage Compartment, this forms the back wall of a EMD cab.

Siemens TCC

The Siemens TCC (Traction Converter Cabinet) required limited hardware yet extensive software changes to accommodate HEP.  The TCC hardware went through a redesign in 1999 and is identical to TCC’s that were built since that redesign, except for Alaska-specific cold-weather modifications first applied on Alaska’s previous SD70MAC order and reused again on the SD70MAC-HEP[1].

Small changes were required to the ASG hardware.  The chassis required a half-dozen wire wrap mods and one PC board needed a capacitor and resistor so that the unit could monitor the 480V HEP output for closed-loop voltage control.

The largest single change was new software for HEP.  “New software” was not the correct phrase – “extensive additions to existing software” is much more appropriate.  The ASG processor is nearly unchanged from 1992 and its software written in assembly language.  In spite of all the new code, the version software written for the Alaska HEP TCC’s is backwardly compatible with every other Siemens TCC.

When in HEP mode, the TCC is programmed to switch the GTO’s at about 400 Hz to create a stepped 60 Hz (fundamental) 660V output, which is then fed to the HEP transformer.  The TCC can generate the 60 Hz under widely-varying DC Link voltages.  At lower HEP loads, only 1050V (dc) is required, which the diesel engine/main generator can handle at TH2.  Full HEP load can be supported with TH3 and 1250 V(dc).  As the engineer changes his throttle, the TCC adapts to the changing DC link voltage and continues to deliver constant HEP voltage and frequency.

HEP Transformer

The HEP transformer (Figure 11) serves to isolates the 2600V(dc) world of the DC Link from the 480V(ac) of the passenger cars. It is of 939 kVA capacity wound with a delta primary and wye secondary (nominally 660V line-to-line in, 480V line-to-line out).  The neutral of the secondary is connected to a HEP ground relay for ground fault detection.  It was manufactured by Trafomec of Italy to Siemens specifications and wound with a series inductance to form the “L” of an L-C harmonic filter. The core is practically identical to the LIRR transformer except for better insulation to avoid water issues painfully learned on Long Island.

Figure 11 –HEP Transformer on forklift blades

The original Long Island HEP transformer (known as TAPS) consisted of a single underslung package that combined transformer core and harmonic filter capacitors.  It was set up so that its cooling air came from the traction motor plenum, which was pressurized by air that first cooled the LIRR inverter phase modules and exhausted out the sides.

For the SD70MAC-HEP, the transformer core was placed in its own package and capacitors mounted separately in the HEP compartment.  In Figure 11, you can see the first HEP transformer in London, supported by long forklift blades for installation across the walkway at the left rear of the locomotive.  The transformer windings can be seen just above the blades, which is also the warm air exhaust to the transformer.  The sheet metal plenum surrounds the top and sides of the core, but is open on the bottom, forming a sheet metal “skirt” around the base.  This skirt aligns with 2” x 2” foam held in place by channels on the floor of the compartment, forming a tight air seal for the exhaust air. Four holes in the two forklift tubes mate with threaded holes drilled into steel bars precisely aligned onto the underframe.  Guide pins were temporarily threaded into the mounting holes to guide and align the transformer installation.  The back pins and hold-down bolts were reached by removable access plates in the transformer plenum and lower battery box.

The first transformer was air-shipped from Italy to London Ontario in May 2004 so that it could be fitted into the last locomotive while still under construction.  Several minor dimensional problems were discovered in the process, justifying the efforts folks at Siemens and EMD went through to get the transformer delivered in time. The problems found were then corrected on the seven remaining units still in Italy, greatly streamlining their installation in Anchorage. 

As bulky and unwieldy as the 3500-pound transformer appears, the Alaska Railroad mechanics, electricians and forklift operator quickly became adept at the transformer installation.

[1]  Jay was involved in the first order of Alaska SD70MACs back in 1999.  EMD put heaters and insulation all over the locomotive to allow the loco to survive at -40oF below and colder. The Siemens inverter uses a fluorinated hydrocarbon to cools its power electronics that turns to Jello below some low temperature.  There are electric heating elements to keep that from happening in any throttle position.

Harmonic Filter Capacitors:

The three HEP harmonic filter capacitors are mounted behind the locomotive handbrake (Figure 12).  This handbrake was redesigned so that the entire brake was mounted on a large removable vertical steel channel[1].  This way, the brake chain could be detached and the assembly could be lifted out with a crane on the very rare occasions that the capacitors had to be accessed.  The three capacitor cans form the “C” of the L-C harmonic filter.  Each can consists of three 700-microfarad capacitors connected in delta and all three cans are connected in parallel to the transformer busbars.

Figure 12 – Harmonic Filter Capacitor

As soon as the HEP system starts up, 220 amps circulate between each of 9 capacitor terminals and the HEP transformer.  This adds up to 660 amps per phase and 550 kVAR (kilo-Volt- Amperes-Reactive), leading power factor.  Under full load (730kW, 878amps at 1.0 power factor), the inverter HEP system delivers 480V with 2.7% total harmonic distortion (THD).  The resulting current THD is 3.9%.   The worst-case voltage harmonic is the 7th (420Hz) at 1.7% of the fundamental.

An interesting effect of the capacitor bank shows up when summer-type loads (air conditioner compressors and blowers) are powered by the HEP system.   During cooler weather, HEP loads are mostly heaters and cooking equipment and thus nearly unity (1.0) power factor.  In that situation, the TCC has to handle the real current of the load plus the 660 amps of capacitive current of the capacitor bank.  In the summer, the load is much more inductive at ~80% power factor.  The leading capacitor current cancels the lagging inductive current of the compressor and blower motors. This lessens the amperage the TCC has to handle (the TCC being a peak-current limited device) and actually allows for another 110kW of HEP capacity at lagging power factor[2].

[1] The handbrake was portable/removable in the same way a 1960’s portable TV was portable because it had a handle!  When the handbrake was removed for capacitor work (which happened in 2008 on the second order of 4300’s when the capacitor connections overheated due to poor nut torquing), the AkRR mechanic Dave Church placed another unit downhill from the MAC, tied down the brakes and ty-wrapped the couple cut levers.  Anchorage yard has a downhill north-south grade and he was NOT going to let that MAC run away!

[2] AkRR has NEVER approached the HEP capacity of the SD70MAC-HEP with any passenger train it has pulled.


The HEP/TM switchgear (Figure 13) has two functions to perform:

Switch the output of TCC#2 from either the three rear truck traction motors or the HEP transformer.

Disconnect the #1 end HEP receptacles so that they are electrically dead when HEP is not required out the lead end of the locomotive (known as TLD for train line disconnect[1]).

In most HEP locomotives, these functions are handled by two separate motor-operated switchgears.  Roberto Michelassi of Elcon, Inc suggested a method where one switch motor could handle both functions.  A five-module switchgear is mounted on mounting plate bolted to aft end of TCC2.  Two switch modules switch the TCC output from traction motors to transformer.  The other 3 heads handle TLD function. The switch modules are equipped with motor cut out solenoids (just like the switch modules used on an EMD DC locomotive to isolate traction motor fields).  When it is desired to disconnect the #1 end HEP receptacles, the solenoids are energized, and the switchgear rolled back and forth to center the switch fingers on the TLD switch modules.  Figure 14 illustrates the four possibilities that correspond to the four modes of HEP/Traction.

Figure 13 – HeTm Switchgear
Figure 14 – Four modes of HEP/Traction

[1] Somewhere through the history of locomotive Head End Power, someone figured that the front receptacles ought to be electrically dead when not needed, so as not to compound problems if a passenger loco hit an automobile in a grade crossing accident.

HEP Blower and Filter

The HEP blower was no exception to the challenge of trying to get all the HEP equipment into the space available.  At one point of the design cycle, we even considered dispensing with the blower entirely and using dampers to allow traction motor air to cool the transformer.  In the end, EMD found a small 480-V blower that would fit high in the long hood, right behind the headlight (see Figure 15).  This blower draws air from a grill high on the engineer’s side just aft of the radiator hatch and discharges to two 10” flexible ducts that route to the top of the transformer (Figure 16)[1].  The air out exhausts out a labyrinth grill just above the walkway, designed to prevent wash water entering the transformer compartment (Figure 17).  The blower and motor are mounted to a removable roof hatch bolted to the end of the long hood.

Figure 15 – HEP Blower Air Intake
Figure 16 – Flexible Ducts to top of HEP Transformer
Figure 17 – Transformer Exhaust Grill

One interesting, unexpected concern that didn’t arise until the locomotives arrived in Alaska was cottonwood!  Up and down the Railbelt in June and July, cottonwood lint wafts through the air.  The lint can be so bad that during the height of the season, undercar gen sets (the cruise train companies have their own generators) need their filters changed twice from Anchorage to Fairbanks.  This problem was pointed out by ARR senior electrician Gary Odens and a solution was quickly found.  The HEP blower air intake is nearly the same size as the 25” x 16” x 2” carbody filter used on our MP15’s.  ARR boilermaker Jim Blakely quickly came up with an easy-change filter holder that would bolt on to the air intake (Figure 18).

Figure 18 – HEP Blower Paper Filter Installation

Fortunately, the conservatism in the HEP blower design paid off here.  As EMD was uncertain of the pressure vs volume characteristics of the repackaged HEP transformer, the air flow engineer over-designed the HEP blower.  Thus, there was plenty of static pressure for the air filter and still allow for sufficient cooling air thru the transformer. 

The blower is controlled by temperature sensors inside the transformer.  Three 100-ohm RTD (resistance temperature detectors) sensors are installed in the windings.  These are read by the TCC#2 computer and fed to the EMD computer, which turns on the blower when any sensor is hotter that 115C and then turns off the blower when all sensors are below 75C.   

[1] We dubbed these flex ducts “Snuffy Trunks”, after Mr. Snuffleupagus of Sesame Street.

Remainder of the Equipment:

The rest of the electrical equipment was wedged into the space available. A small cabinet was set into the long hood aft of the radiator hatch and was dubbed the Small HEP Cabinet (appropriately enough).  Figure 19 shows the cabinet.  It contains the contactor and circuit breaker for the HEP blower, the HEP Ground Relay, a Train Line Voltage relay (prevents the main contactor from closing or the HeTm switchgear from moving if the external HEP trainlines are energized) and a pilot relay for the big ACC main contactor.  

Figure 19 also shows the rest of the ancillary equipment.  The Potential Transformers (PT’s) are 100:1 transformers that provide voltage feedback of two line-to-line HEP voltages to the Siemens ASG.   An “old-fashioned” Under-Over-Voltage relay serves as a hardware backup to open the ACC if the voltage control of the TCC fails.  The main ACC contactor is partially obscured by the HEP air ducts.  This is rated at 1200 amps and is equipped with CT’s and a thermal overload element.  Both are somewhat redundant as the TCC quickly acts to cease HEP whenever there is an overload or short circuit in the passenger cars.

Figure 19 – Small HEP Cabinet

HEP Controls:

The HEP controls for the SD70MAC-HEP were, in some ways, a step forward into the past.  The LIRR DE/DM30 uses several Display screen menus to control its HEP system, but very early in the design process, Dennis Melas (EMD Software Systems Manager) told me, “You know, I can pay for a lot of switches and pushbuttons before I can justify spending man-hours to write code for more menus—especially for just 8 locomotives!!!” 

So instead, we used the old EMD “eggcrate” lights for status and fault information (controlled by the computer), pushbuttons for start, stop and fault reset (read by the computer) and a multi-deck 4-position rotary switch for the HEP/Traction mode select (this also handles the convoluted Train Line Complete logic).  Instead of digital or analog meters for voltage and current, a Display default meter screen is just one FIRE screen button press away.  Thus, we made a control panel that’s a combination of the look and feel of our older GP40-H locos.  Moreover, we also duplicated the pushbutton/light sequence of the GP40-H as well.  The montage in Figure 20 illustrates the back wall panel and engineer’s station display.

Figure 20 – Montage of HEP cab controls


The results of this project are eight versatile passenger/freight locomotives – ARR 4317 through 4324.  When the HEP system is off, the unit is a 4300-THP locomotive.  With HEP on, it is a 2400-THP, 730-kW HEP locomotive.   Moreover, it can deliver that HEP load in TH3 (490 rpm) and up to 270kW HEP in TH2 (370 rpm).  An F40PH or GP40-H in contrast would run at a constant 900 rpm to deliver the same HEP.

The system has turned out to be very reliable.  If the 710 engine starts, then HEP is available. There is no pony engine to service and thus no pony engine cooling system, fuel system or lube system to deal with.  The only filter to change is the HEP blower filter and so far, the cottonwood lint at 15 feet off the rail has been very manageable.  The only failures have been one HEP blower contactor and one broken switch module. 

Presently, we have four 4300’s on our Anchorage-Fairbanks daily service.  Two units leave Anchorage and two leave Fairbanks every morning at 8:15.  One unit in each consist provide HEP and 2400 traction horses, the other provides full 4300 THP.

Last summer (2005), one 4300 handled the Seward Coastal Classic train by itself, but this summer the train gained two cars, putting it outside our comfort zone of a single 4300 making HEP and pulling the 3.0 percent grade approaching Grandview.  So, we put our older P30 power car (a 4-axle ex-E9B) and a 4300 making full traction.  Our cruise train service from Seward, Anchorage airport and Whittier is handled by another 4300 and our P31 cab/HEP power car. 

The MAC’s making HEP do have a unique sound – the switching of the inverters produces a distinct “EEEEEEEEEE” pitch at about 400 Hz, very noticeable when standing right next to the TCC#2.  In terms of sound levels, the whine of the MAC is 82 dBA at 20 paces versus 85 dBA of a GP40 in HEP mode.  But a single-tone whine is much less objectionable than the roar of a 900rpm 645 engine!

The HEP system efficiency turned out to be a bit of a surprise when finally tested.  Efficiency was never a contract requirement, but upon testing we found that full-load efficiency was only about 88%, rising to 92% at part loads.  This represents nearly 100kW of losses at 700kW load.  Consultations with Siemens revealed that part of the poor efficiency was due to the limitations of the TCC computer.  The LIRR computer has a more modern processor so that more efficient switching pulse patterns could be selected.  These could not be realized with the SD70MAC TCC computer.

But in many ways, the “poor” efficiency is of little consequence.  Seldom will we see large 700kW HEP loads.  To improve that efficiency would require more copper and more steel in the transformer.  That in turn would cost more money and demand more space in an already crowded locomotive – even if there was money to do a redesign!  Thus, the poor efficiency really is just an acceptable result of a logical engineering trade-off.

The SD70MAC-HEP units arrived in Alaska in April and May 2004 and immediately started pulling freight and passenger trains (with HEP provided by other means).  HEP transformers were installed in September-December 2004.  EMD and Siemens engineers arrived in October 2004 to test and commission the HEP software.  Limited runs with the HEP system running were made on our weekend Aurora trains starting in January 2005.  After correcting one extremely annoying software bug in the spring, the units entered day-in/ day-out on May 15th of the same year.  I see no reason to expect them not to be running 20 years from now (Figure 21).

Figure 21 – SD70MAC-HEP on first day of 2005 passenger season


All photos and diagrams are by the author, except for Figures 1, 2, 3 & 4, which are EMD photos from the collection of Jay Boggess.

The Alaska HEP project had many participants; all who deserve recognition and all who without their help this project would not have succeeded: Ulrich Foesel, Hartmut Wagner and Horst Nowy of Siemens: Ulrich was my counterpart at Siemens, Hartmut was the software engineer on TCC’s and Horst was the long-time service engineer in Alliance, NE, who did the ASG hardware mods.

The following Electro-Motive folks: Dennis Melas (software manager), Curtis Montgomery (software engineer) and Margaret Foltz (software testing).  These three had to translate, write and test the code for 70MAC HEP.

Forrest Green (systems engineer): He and I worked together (along with many design/drafters) to get the 10 pounds of stuff into the HEP compartment of the 70MAC.

Todd Lail (systems engineer).  He got to pick up the pieces after I left EMD for Alaska.

Tony Bladek (lead engineer for LIRR DE/DM30) and Craig Prudian (systems engineer); Both whom I bounced many ideas off of, especially in the fields of passenger locomotives and Head End Power systems.

Plus, dozens of others at EMD in LaGrange and London who pushed pencils, swung wrenches, found wayward parts and translated barely-dry drawings into a completed locomotive.

Roberto Michalessi and Frank Garrone of Elcon, Inc (Minooka, IL):  We worked together on the 2002 incarnation of Alaska HEP when EMD thought we’d mount HEP equipment beneath the underframe.  Elcon didn’t get to build the cabinets for the final version but did build some subassemblies.

The electricians, machinists and boilermakers at the Alaska Railroad who installed the HEP transformers in Anchorage and helped commission the HEP system. 

Finally, Tim Keck and Dave McColl of EMD and Robert Stout, formerly of the Alaska Railroad (now with Colorado Rail Car); the idea of the SD70MAC-HEP first germinated in their minds. I and everyone else just watered the seedling and let it bloom.

Postscript November 2021 – Alaska RR bought 4 more SD70MAC-HEP locos in 2006, which were delivered to Alaska in 2007. ARR 4325 – 4328. Combined with the first 8 HEP MACs and the original 16 4000THP SD70MACs means that AkRR has 28 6-axle AC locomotives.

Jay Boggess left AkRR in 2010 to work on hydroelectric dams for the U.S. government.

General Motors sold the Electro-Motive Division in 2005.  The new owners renamed it Electro Motive Diesel.  Caterpillar through its subsidiary Progress Rail purchased EMD in 2010 and Progress Rail eventually dropped the EMD name.

Many souls Jay worked with at the time have now retired and some have since passed away.

Thanks as always to my cohort in Vintage Diesel Design Jay Boggess for sharing and updating this fantastic look at these locomotives. While no, I guess a 17 year old locomotive is not technically “vintage”, it is of course an important part of EMD’s vast history.

A Chain Driven Crankshaft.

Yeah, when I first head about that, I scratched my head also.

We are all familiar with the Fairbanks Morse 38D opposed piston engine. The engine has its roots back to 1933 with a 6-cylinder design. It used a very boxy, cast iron block, with a small 5″ bore and 6″ stroke, hence the 38A5 model designation. The odd thing about this engine, is that it used a chain driven upper crankshaft!

Check out the original patent for this engine, filed by Heinrich Schneider and Percy Brooks on behalf of Fairbanks Morse in September of 1933.

Fairbanks Morse 38A5 Patent

By 1938, a new welded frame was introduced, and gave the OP the appearance we all know of today, however, that upper crankshaft was still chain driven.

The engine, with a 5″ bore and 6″ stroke was first used in a doodlebug railcar for the Milwaukee Road. At the same time, a larger 8″ bore engine was introduced. The engines were a success, and would catch the eye of the US Navy.

I was recently able to purchase a manual, which best I can tell is from 1937, for the Fairbanks Morse 38C5 engine, the slightly newer version of that prototype mentioned above. While the manual is extremely primitive, it does illustrate that chain driven upper crankshaft assembly. A second chain was used for the camshafts and timing.

According to C.H, Wendel in his 100 Years of FM Engine Technology book, the chain would be replaced in 1937 with a vertical drive shaft operating off of bevel gears on each crankshaft, which is still used to this day, however that second timing chain is still in place. The engine would go through a multitude of improvements, in which the engine reached its two main production sizes, the smaller 5 1/4″ x 7 1/4″ introduced in 1939 and the 8 1/8″ x 10″ in 1938. The larger of the two still being produced, with the smaller being discontinued in the early 1970’s. A full post on the smaller engine is planned.

While Diesel engines were in their infancy with design work changing daily, I still look a this and go “What were they thinking!?”

Amazingly enough, a pair of those original FM 38A5 engines still exist, from the USS Enterprise (CV-6). The engines were saved by the Rock River Thresheree, north of Beloit, WI.

A Turbocharged Failure – The Story of the Cleveland 498, Part IV

Part IV, the final part of this series will be a photo documentation of the tug Idaho.

Be sure to view the previous parts:
A Turbocharged Failure – The Story of the Cleveland 498, Part I – Introduction
A Turbocharged Failure – The Story of the Cleveland 498, Part II – Engine Design
A Turbocharged Failure – The Story of the Cleveland 498, Part III – Engine Installations

By the early 1990’s, the Great Lakes Towing Company (GLTC) would have the only running Cleveland 498 engines left in the US (See note on the bottom). The Towing Company as they are known has a rich history dating back to its formation in 1899, consolidating several smaller tugboat companies on the Great Lakes. GLTC currently serves numerous ports across the Great Lakes, and is the largest user of Cleveland 278 (A and non A engines) left in the country.

Starting in 1907, the company began to build their own tugs in house, in their own shipyard. The yard, originally in Chicago, and moving to Cleveland is still churning out all new tugs for the company today, as well as doing outside work.

In 1931, the yard constructed Hull # 67, and named her the Idaho. GLTC had two sizes of tugs, the smaller, “Type I”, which were named after cities, and larger “Type II”, named after states. The Idaho would be the last new tug built until 2008.

The Idaho was originally powered by a single cylinder, 26″ x 28″ steam engine. The tug was 84′ 4″ long, 20′ beam and a 12’6″ depth. The tug was one of three that would receive a raised height wheelhouse for doing lake towing.

The Idaho after receiving her raised wheelhouse. Please note that this photo is not in my collection, simply one from my files. If anyone knows the photographer or archive please send me a message so it can properly be credited.
Ironically, one of the very first pieces of Cleveland Diesel ephemera I would add to my collection would be one depicting the new 498 powered tugs of the Great Lakes Towing Co.

In 1956, the Idaho was on the block to be converted to Diesel propulsion. The engine chosen was the new 498 from Cleveland Diesel, as outlined in previous posts. Cleveland Diesel Order #1640 was placed in early 1956, for a pair of left hand rotation, 1400HP, 8-cylinder 498 engines to convert the tugs Montana and Idaho (Montana was an identical sister, Hull #60 of 1929). The engine for Idaho, #46002 was shipped from the factory on 12/13/1956, having to only go a few miles up to the companies shipyard. The tugs would receive Diesel-Electric propulsion packages, utilizing WWII surplus Destroyer-Escort main generators and propulsion motors. Disaster struck the Idaho shortly after being rebuilt on 10/21/1960. The tug was assisting the lake ship C.H. McCullough, Jr. in Chicago, when the tug was sunk. She would be raised, dried out and put back in service. A photo of her being raised appears in Alexander Meakins “The Story of the Great Lakes Towing Co.”

The 498 powered tugs would never stray too far from the main yard in Cleveland, typically working the ports of Cleveland, Ashtabula, Toledo or Detroit. The porthole aft of the wheelhouse is the tugs small bathroom. Photo by Isaac Pennock.

Great Lakes Towing Company would ultimately have a quartet of 498 powered tugs. The Diesel-Electric Montana and Idaho, and the Clutch tugs Tennessee and Pennsylvania which were converted in 1960 from Steam. Montana would be retired in 2006, Tennessee in 2012 and the Pennsylvania in 2019. Ironically, the Pennsylvania would wind up receiving a replacement engine at some point in her life, originally out of the towboat Leila C. Shearer. This too was replaced with an EMD 12-645, however the conversion was never finished.

Sister tug Montana received the first 498 engine to be sold, seen here being lowered into the tug at the Cleveland yard. From Cleveland Diesel’s “More Power For You” brochure.

Noting that the last surviving 498 was likely nearing the end of her life, we reached out to the company to see about the possibility of documenting the engine and tug, and maybe see about preservation options. Unfortunately, we would be a touch too late. While the tug was still around, it was sitting laid up having suffered a catastrophic engine failure in 2016, however we were welcome to document her anyway.

The tug was laid up in Detroit for a few years, and was being used for parts for the other tugs in the GLTC fleet (while the engines were different, the tugs still share many parts between them).

The heart of the Idaho is her Cleveland 498 engine. Note the exhaust jumpers are rusty, having no water jacket around them, and by this point, no insulation either.
The tug had a WWII surplus, Allis-Chalmers 525V DC propulsion motor, rated for 1090kW at 720RPM. On top is a 120V DC shaft generator.
The power package installed in the Idaho. From Cleveland Diesel’s “More Power For You” brochure.
The propulsion switchboard. At left is a pair of excitation generators.
Also from the Destroyer-Escort is the propulsion motor. This was built by Westinghouse, and rated for 1225HP.
Farrel-Birmingham reduction gear, with a 4.233:1 ratio. The tug has a 102″ x 87″ stainless 3 blade propeller.
The portside, aft end of the engine room has the steering gear pump, as well as a motor-generator set. The fuel tanks are located behind the aft bulkhead.
Portside of the engine, the air starter is mounted on the floor level. On top are the various gauges and governor.
Detroit Diesel 3-71 with a 30kW generator. In front is the tugs oil fired steam boiler for heating.
Air compressors.
The heart of the 498 is the De Laval turbocharger. The air intake filter is seen in the middle, with the compressor on top. A discharge tube feeds air into the intercooler on the bottom.
From the intercooler, the air fed into the roots blower (at left) from the bottom end. It was mentioned these engines sounded like helicopters.
Looking aft, the large cast cover is over the camshaft balancer, proudly displaying the maker of the engine.
Taking on ode from the 248 engine, the 498 used a two piece top cover. On the bottom right is the blowdown/safety valve. Former engineers for this boat mentioned heads and head gaskets were a big failure point being addressed often.
The reason the Idaho was retired. Shortly after startup, the #4 piston locked up, thus the connecting rod snapped, in turn swinging around and slamming into the airbox and both liners.
Unfortunately, parts for the 498 were long since unavailable, and a failure like this is typically a death sentence anyway, especially in an 80+ year old hull.
The hydraulic power pack and head tightening tool.
G tugs still have a tiller handle for rudder control, along with the Lakeshore throttle stands.
Wheelhouses were rather spartan, with a simple bench, small chart table and propulsion gauges. These tugs were only intended to do day work, with no real provisions of any kind.
These tugs were built for one purpose, docking ships, thus the low profile deckhouse. The stairs in front lead down into the forecastle.
A few basic bunks, lockers and a simple table in the bow.
The tugs official number, gouged into the steel 80+ years ago.
Great Lakes Towing exclusively used towing bitts manufactured by the Montague Iron Works in Northwest Michigan.
“G” tugs as they are known in the lakes, got their name from the large stack insignia.
The Idaho returning from a job in her last year in service on the Detroit River. Photo by Isaac Pennock.
Removal of the stack insignias traditionally mean the end is near. This was likely the last photo of the Idaho in one piece. Bill Kloss Photo.

Unfortunately, all things must come to an end. In January of 2019 the tug was towed back to Cleveland, and with the last few usable parts removed, the tug was scrapped. We can’t thank the Great Lakes Towing Company enough for allowing us to photo-document the tug.

With only 58 engines built, and being that virtually all of the engines stateside were replaced long ago, it is highly unlikely any of the foreign sold engines remain. We heard a rumor of one driving a water pump in Egypt, but again, this would have had to have been a relocated engine, and is highly unlikely it exists. Somebody please prove us wrong!

That wraps up our four part series on the Cleveland Diesel 498 engine. Please be sure to view the previous posts on this engine, linked on the top of this page. I will say it again, if anyone has any 498 manuals, brochures, stories, parts, anything, please get in touch with us. Should anything new arise, we will make another follow up down the road.

2022 Update – At some point this year we will post another part in this series, with some additional information we found.

A Turbocharged Failure – The Story of the Cleveland 498, Part III

Part III of A Turbocharged Failure will be fairly straightforward – a listing of ALL 498 engines built.

Be sure to view the previous two parts:
A Turbocharged Failure, Part I – Engine History
A Turbocharged Failure, Part II – Engine Specifications

Production of the Cleveland 498 commenced with the first engine shipped in May of 1956. Most production would take place in the fall of 1956 (16 engines built), and the summer of 1957 (17 engines built). 1958 saw only a pair of engines, a trio in 1959, and the last 4 were built in 1960. A total of 29 8-cylinder, 9 12-cylinder, 17 16-cylinder and 3 test engines (one 8, and two unknown) were built over the course of production, for a grand total of 58 engines.

A brochure for the engine issued not long after being announced at Powerama. Click for larger.

1) Tug Montana – Great Lakes Towing Company, Cleveland, Ohio
Engine 46001, Shipped 5/2/1956, Order #1640, 8-498, 1400HP/850RPM

2) Tug Idaho – Great Lakes Towing Company, Cleveland, Ohio
Engine 46001, Shipped 12/13/1956, Order #1640, 8-498, 1400HP/850RPM

Great Lakes Towing Company needs no introduction here, they are the largest tug company performing shipdocking on the Great Lakes, using “G Tugs”.   We will do a more detailed feature on these down the road.   Great Lakes put in the very first order for 498 engines, with the first one going into the tug Montana.   Montana was built in 1929, with a single cylinder steam engine.   Idaho followed a few months later.  Idaho was the last “new” tug built, in 1931.   Both tugs were identical and built-in house, receiving electric drive propulsion packages using surplus Destroyer-Escort generators and propulsion motors***.   The Montana was retired and scrapped in 2006, and the Idaho was scrapped in 2019.  The 4th and final part will be dedicated to the Idaho

Tug Idaho shortly after being converted to Diesel power. VDD Collection.


3) Tug Hoboken – Delaware Lackawanna & Western Railroad – NY, NY
Engine 46003, Shipped 10/31/1956, Order #1807, 8-498, 1400HP/850RPM

4) Tug Buffalo – Delaware Lackawanna & Western Railroad – NY, NY
Engine 46004, Shipped 11/30/1956, Order #1807, 8-498, 1400HP/850RPM

5) Tug Syracuse – Delaware Lackawanna & Western Railroad – NY, NY
Engine 46005, Shipped 12/28/1956, Order #1807, 8-498, 1400HP/850RPM

6) Tug Utica – Delaware Lackawanna & Western Railroad – NY, NY
Engine 46006, Shipped 1/14/1957, Order #1807, 8-498, 1400HP/850RPM

7) Tug Nazareth – Delaware Lackawanna & Western Railroad – NY, NY
Engine 46007, Shipped 1/21/1956, Order #1807, 8-498, 1400HP/850RPM

Delaware Lackawanna & Western placed an order for 5 Diesel-Electric tugs with Bethlehem Steel of NY, built to General Managers Association (GMA) design for moving carfloats in NY Harbor.   Erie Lackawanna started to sell off the tugs in the early 1970’s, these were the first to go, and every one of them was repowered not long after being sold (all being repowered by the early 1980’s).  Two would go on to get GE engines, two would get Alcos, and the last an EMD.  The Utica, the last survivor, is now working in Panama.   These tugs will be covered extensively in my upcoming book on Railroad Tugs, coming out later this year.  

Diesel Times/J. Boggess Collection


8/9) Towboat Lelia C. Shearer – O.F. Shearer & Sons, – Winchester, KY
Engines 46008, 46009, Shipped 10/19/1956, Order # 1883/1884, 8-498, 1230HP/750RPM

Hillman Barge & Construction both designed and built this 2700HP diesel-clutch twin screw towboat for the O.F. Shearer & Sons company.   She was repowered in 1964 with a pair of EMD 16-567C engines.  The towboat kept her name through several companies and was finally scrapped in 2014.    This was the first 498 powered towboat. 

Diesel Times/J. Boggess Collection


10/11) Tug M.P. Anderson – Brown & Root, Inc.
Engines 46010, 46011, Shipped 7/30/1956, 731/1956, Order # 1974, 8-498, 1400HP/850RPM

M.P. Anderson was designed by Brown & Root and built by Gulfport Shipbuilding.  This 123-foot, twin screw, Diesel-Electric tug worked in the Gulf for most of her life and was also repowered with a pair of EMD 16-567C engines, with reverse-reduction gears in place of the electric drive.  She is now working in Baltimore as the Austin Krause (and has one of the largest tug engine rooms I have ever been in).

The M.P. Anderson was covered in the June 1959 issue of Diesel Times. J. Boggess Collection


12) Tug William C. Gaynor – Great Lakes Dredge & Dock Co.
Engine 46012, Shipped 9/11/1956, Order # 1956, 8-498, 1400HP/850RPM

This 94’ tug was designed by Joe Hack under Cleveland Diesel for Great Lakes Dredge & Dock.  The tug was built by DeFoe shipbuilding and spent her entire life in the Great Lakes doing dredge work.   Today she is working (under her original name) for Sarter Marine in Sturgeon Bay, WI.   The tug was repowered with an EMD 12-567C in 1990.


13) Test Engine
Engine 46013, 8-498


14/15) Towboat Gulf Inlander – Gulf Oil Corp.
Engines 46014, 46015, Shipped 10/26/1956, 11/30/1956, Order 1923/1924, 8-498, 1400HP/850RPM

Gulf Inlander was a twin-screw towboat built by St. Louis Shipbuilding for Gulf Oil.   Now known as the Mary Lynn, she was repowered and now has a pair of EMD 16-645 engines. 

Diesel Times/J. Boggess Collection


16-26) All engines 8-498 Non-Magnetic, 1400HP/850RPM

MSO-521 Assurance, Engine 46016, Shipped 10/28/1957, Order # 62562
MSO-519 Ability, Engine 46017, Shipped 7/29/1957, Order # 62562
MSO-520 Alacrity, Engine 46018, Shipped 8/5/1957, Order 62562
MSO-519 Ability, Engine 46019, Shipped 6/22/1957, Order 62563
MSO-520 Alacrity, Engine 46020, Shipped 8/7/1957, Order 62563
MSO-521 Assurance, Engine 46021, Shipped 9/10/1957, Order 62563
Naval Supply Depot (spare engine?), Engine 46022, Shipped 11/30/1960, Order 62672
MSO-519 Ability, Engine 46023, Shipped 7/31/1957, Order 62572
MSO-520 Alacrity, Engine 46024, Shipped 8/27/1957, Order 62572
MSO-521 Assurance, Engine 46025, Shipped 11/6/1957, Order 62572
Naval Supply Depot (spare engine?), Engine 46026, Shipped 11/30/1960, Order 62675

USS Ability MSO 519 stricken 1971
USS Alacrity MSO 520 sold for scrap Dec 1979
USS Assurance MSO 521 scrapped in Dec 1979

All we know about these three minesweepers with non-magnetic 498s is what we can find in Wikipedia & Navsource. We have no idea how long the 498s lasted or how well they did – it is likely the reason these ships were retired was because of the 498’s.  Since these three ships were scrapped over 40 years ago, we suspect that information is lost to the ages.  BUT, if there are any ex-Navy sailors out there, drop us a line.    

Diesel Times/J. Boggess Collection


27/28) Towboat Eleanor Gordon – Two engine order, shipped 4/24/1957, Order 2039/2040, 8-948, 1400HP/850RPM. 

Designed and built by Nashville Bridge Co. for Mid America Transportation Company.   This 149’ towboat was powered by the pair of 498 engines with Falk reverse reduction gears.  Apparently Mid-America was so displeased with these engines that the towboat was repowered within 18 months. 

Diesel Times/J. Boggess Collection

The engines were sent back to Cleveland, who rebuilt them and reshipped them under a new order to Great Lakes Towing Company, who installed them into a pair of tugs, the Pennsylvania and Tennessee

 Pennsylvania would be one of the tugs assigned to work all the way down in Florida on a Navy contract in the 1990’s.   Tennessee was scrapped in 2012, with the Pennsylvania being scrapped in 2019.  The Pennsylvania was repowered with an EMD 12-645, however the repower was never completed before GLT decided to scrap her (?). 

Tennessee was an identical sister to the Pennsylvania, and also worked in Florida.  Both of these tugs were the only “G” tugs to have fixed Kort nozzles, with 102” wheels. 

Tug Pennsylvania
Engine 46027, Shipped 11/30/1959, Order 3936

Tug Tennessee
Engine 46028, Shipped 11/30/1959, Order 3937

The Pennsylvania and Tennessee on the job in the early 1970’s. VDD Collection.


29) Tug Alexander Wiley Robinson Bay, St. Lawrence Seaway Development Corp.
Engine 46029, Shipped 11/15/1957, Order 2573, 8-498, 1400HP/850RPM

Robinson Bay is a 103’ Diesel-Electric ice breaking tug designed by Merritt Demarest for use in the St. Laurence Seaway.   The tug was repowered by Great Lakes Towing in 1991, who kept the engine as a spare parts source.  The tug is now powered by a Cat 3606 with a 1750HP GE 581 propulsion motor. 

The Robinson Bay at work in Northern New York. Will Van Dorp Photo.


30) Development Engine
Engine 51001. 12-498S


31/32) Towboat Cypress, Chotin Transportation
Engines 51002, 51003, Shipped 8/31/1956, 12-498, Order 1653/1654, 2100HP/850RPM

Cypress was a 140’ towboat for the Chotin Transportation Company designed and built by J&S Shipbuilding.   The towboat has been out of documentation for some time and repowering/disposition is unknown.

Diesel Times/J. Boggess Collection


33) Tug Ralph E. Matton, John E. Matton & Sons, Cohoes, NY
Engine 51004, Shipped 7/31/1957, 12-498, Order 1726, 2100HP/850RPM

Ralph E. Matton was a New York Canal tug, designed and built by Matton.  The tug was repowered with an EMD 16-567C, and later became the Mary Turecamo, and Albany.  It was scrapped about 15 years ago. 

Courtesy of Dave Boone


34) Tug Spartan, James McWilliams Blue Line, NY, NY
Engine 51005, Shipped 9/14/1956, 12-498, Order 1893, 2100HP/850RPM

Spartan was a NY Canal tug, designed by Cleveland Diesel (Joe Hack) and built by Calumet Shipyard.   The tug became part of the Ira Bushey & Hess family of companies and was reefed in 1986. 

VDD Collection


35) Tug Matton #25, John E. Matton & Sons, Cohoes, NY
Engine 51006, Shipped 10/20/1956, 12-498, Order 1939, 2100HP/850RPM

Matton 25 was a New York Canal tug, designed and built by Matton.  The tug was repowered with an EMD 16-645, and later became the Joan Turecamo, and Everglades of Seabulk Towing.  It was reefed in 2017.


36) Tug Matton, John E. Matton & Sons, Cohoes, NY
Engine 51007, Shipped 4/29/1957, 12-498, Order 2210, 2100HP/850RPM

Matton was a New York Canal tug, designed and built by Matton.  The tug was repowered and later became the Kathleen Turecamo, and Troy.  It was reefed in 1990.

Courtesy of Dave Boone


37) Test Engine
Engine 51008, Order 3133


38) Gen-Set, Bell Telephone Co., Philadelphia, PA
Engine # 51009, Shipped 7/17/1957, Order 2118, 12-498, 1840HP/720RPM


39/40) Towboat Oliver C. Shearer, O.F. Shearer & Sons, Cedar Grove, WV
Engines 51010, 51011, Shipped 7/14/1960, Order 5058/5059, 7/21/1960, 12-948, 2100HP/800RPM

Shearer returned for another set of engines for a second towboat, the Oliver C. Shearer.  She was designed by Friede & Goldman Inc. and built by Marietta Manufacturing.   The towboat was repowered in 1965 with EMD 16-567C’s and has since been repowered several times with EMDs.  The towboat is still in service under her original name. 

Diesel Times/J. Boggess Collection


41) Development Engine
Engine 57001, Order 4150, 16-498S


42/43) Towboat Mark Eastin, West Kentucky Coal Co., Madisonville, KY
Engines 57002/57003, Order 1775/1776, Shipped 12/14/1956, 11/30/1956, 16-498, 2800HP/850RPM

The 177’ Towboat was at the time, the most powerful twin screw towboat on Inland Rivers.  Repowered in 1969 with EMD 16-645 engines.   In service today as the Kevin Michael

Diesel Times/J. Boggess Collection


44-53) Gen-Sets, Cia Cubana de Electricidad, Havana, Cuba
All engines are 16-498, 2850HP/720RPM, Order 2361

Engine 57004, Shipped 7/13/1957
Engine 57005, Shipped 7/19/1957
Engine 57006, Shipped 7/31/1957
Engine 57007, Shipped 8/6/1957
Engine 57008, Shipped 8/30/1957
Engine 57009, Shipped 8/31/1957
Engine 57010, Shipped 9/14/1957
Engine 57011, Shipped 9/21/1957
Engine 57012, Shipped 9/26/1957
Engine 57013, Shipped 9/30/1957

The largest order of 498 engines were these stationary 2000kW engines for a Cuban powerplant.   It is unknown how long, or if they still exist.   Anybody in Cuba want to go exploring for us? 

Diesel Times/S. Zelinka Collection


54/55) Gen-Sets, Saudi-National, LTD, PEX-583007
Engines 57014, 57015, Shipped 12/3/1958, 12/12/1958, Order 2982, 16-498, 2800HP/720RPM


56-58) Dredge Alaska, Great Lakes Dredge & Dock

Pump Engines
Engines 57016, 57017, Shipped 9/24/1959, Order 3756, 3757, 16-498, 2770HP/800RPM

Engine 57018, Shipped 12/28/1959, Order 3760, 16-498, 2800HP/800RPM

Hydraulic Dredge Alaska used a trio of 498 engines.  Two engines drove the main pump drive unit, with the 3rd driving three generators, a 1250kW, 500kW and a 200kW, all on a common frame.   The Alaska is still in service, but of course was repowered, and currently has EMD 710 engines. 

Diesel Times/J. Boggess Collection

While most of the above users of the 498 were featured in a dedicated issue of Cleveland Diesel’s newsletter Diesel Times, the 9/1957 issue showcased the current maritime users of the engine. Click for larger.

Coming up in the final part of A Turbocharged Failure will be a post dedicated to the Great Lakes Towing tugboat Idaho, the last known 498 engine to be in use.

Thanks to my Cleveland Research Partner J. Boggess for proofing and sharing the above issues of Diesel Times.

A Turbocharged Failure – The Story of the Cleveland 498, Part II

In this second part of A Turbocharged Failure, we will go through some design features of the engine.  What better way to do this then to simply go through the engine manual and show a few key areas of the engine design.  Numerous additional photos of the 498 will appear in Part IV. 

Be sure to view the other parts of this series:
A Turbocharged Failure, Part I – Engine History
A Turbocharged Failure, Part III – Uses and Installations

The initial “catalog photo” of a production 12-498, with a Falk MB series reverse-reduction gearbox.

I have two versions of the 498 manual – both of which are titled as “preliminary” manuals.   The older version, which is undated and likely from around 1957, and lists 4 models of the 498; a 6-, an 8-, a 12- and 16-cylinder.   Like all manuals, an end view diagram is included, however this is a rather primitive, hand- drawn sketch. 

Engine cross section drawings. A colorized version would never be done. Click for larger

By the time the second edition was printed, dated for July of 1960, an all-new diagram was made including outlining various parts of the engine.   Along with this, several additional diagrams appear in the manual, as well as some more photos of various engine parts and repair techniques. At some point the 6-cylinder version was dropped from documentation, and would ultimately never be built.

Engine data and ratings – Click for larger

Engine Operation
Like all Cleveland engines, a simple lever is attached to the injector linkage.  A small thumb latch allows the lever to control the engine with no governor input.  When unlatched, the governor takes over all control of engine operation, used in conjunction with whatever remote propulsion control system is used.   On the 498 (and some Cleveland 567’s) equipped with reverse reduction gears, a second lever was added to control the air clutch, so that propulsion speed and direction could be controlled right at the engine.  498 engines were air-started- using an air motor, unlike the 278A engines which had direct air start into the cylinders.  

Engine operating levers on the 498. Click for larger

Like the previous Cleveland models, the 498 used an all-welded crankcase of various forged parts and steel plate.  A balanced, alloy steel crankshaft is used, interestingly enough the 12-cylinder crankshaft did not have any counterweights.  The crankshaft is drilled for oiling the connecting rod main bearings and wrist pins.  A vibration damper and balancer are mounted on the front end of the engine. 

Pistons, Connecting Rods & Liners
One of the biggest sets of improvements to the 498 engines from the previous 278A – are actually a few concepts borrowed from sister division EMD and the 567C engine. 

While the 278A and all previous models used a semi-water deck style liner (like the EMD 567 though the “B” block) – the 498 used a sealed liner which was attached to a water manifold in the airbox by a jumper (much like the 567C). 

A look in through the crankcase inspection cover at the connecting rods, showing the “pee pipe” attached to the lower end of the liner for piston cooling.

Again, borrowing from EMD, the pistons are a two-piece, trunk style floating piston (introduced on the LST 12-567 in WWII), whereas the 278A used a more traditional one-piece piston and a wrist pin.  On the floating piston, the piston itself sits on a thrust washer, which in turn sits on the piston carrier attached to the connecting rod.  Again, departing from the 278, Cleveland adopted the “pee pipe” piston cooling scheme EMD used since the first 567 of 1938, as opposed to the drilled connecting rod of the 278 & 268, which directed cooling oil from the crankshaft to the bottom of the piston. In the 498, the drilled connecting rod is only used for oiling the bearings and wrist pin.

Piston cooling on the 498 (top) used a drilled connecting rod to lubricate the bearings and wrist pin, however used a jet of oil which sprayed into an orifice directing oil into the cavity below the piston crown. The 278 (bottom) simply used the drilled connecting rod to lubricate everything, and used a spring check plate to retain oil in the crown.

The connecting rod of the 498 used a strap style of “cap” to contain the bearing and connect to the crankpin.  Like all Cleveland’s, each connecting rod used its own bearing on the crankpin, unlike the EMD engines which use a shared crankpin and bearing set by use of the fork and blade connecting rods. This allowed the EMD to be a slightly shorter length overall, as the cylinders were directly opposite each other, versus slightly offset on the Cleveland.

Cylinder Heads & Exhaust
Again, sharing with previous Cleveland models, the 498 used individual external heads, however these had some upgrades.  One of the big downfalls with the 278A style head, is there is a half-circle seal against the back of the head, which seals the cam pocket to the head.  Unfortunately, this is a major source of oil leaks.  The Navy devised a tool in the 1950’s to help combat this problem – a bracket clips into the injector control pocket on the block and a set screw presses the head back into the pocket, thus compressing the seal before the head is torqued down.  The 498 head had a specific tab on them (visible in the above photo of the operating levers), in which a bolted clip catches, allowing one to compress the head back into the seal.   The head itself was also torqued down in an interesting way. The head used stretch bolts, in which a special hydraulic tool was attached to pick up on the bolt before it was tightened down. The 498 returned to a two-piece valve cover design like the Winton 248 used.  The fuel lines were also moved inside now. 

Looking down on the cylinder head, which is a bit more cramped then those on the 278 family. Both fuel lines are now inside the engine, connected to main fuel lines under the exhaust manifold. The 498 used a two piece cover, like the original Winton 248 engine (I wonder if they are the same castings..). The new head design uses a combination safety and test valve, which were separate valves on the 278. EMD did not utilize these valves, which open should any excessive pressure build while the engine is running, preventing a bent connecting rod or worse. Note that the exhaust jumper has some sort of spray on insulation.

The hydraulic tool for tightening the head bolts was a rather simple process. The tension shaft is threaded onto the top portion of the stud, the tool is slipped over the tension shaft, and a nut on top secures it together. The tool is pumped up to 5,000PSI, and the actual nut holding the head down is tightened with the socket handle inside of the tool. The later 1960 manual indicates that these could also be manually torqued down to 1,030 ft. lbs. Click for larger.

A slight revision on the exhaust jumpers as well was devised.  Previous engines used a completely water- jacketed jumper (the older manual incorrectly stating that the 498 had this as well), however with the 498, it was preferable to keep the exhaust gases as hot as possible entering the turbocharger – thus no water jacketing on the exhaust jumper. A small pipe exits the head and carries water to the main exhaust manifold, which was still water jacketed.   The main exhaust manifold itself used diffuser sections to carry the exhaust gas to the turbocharger. 

Camshaft, Accessory and Governor Drive
Nothing all that special going on here.  The water pumps and blower are driven off the accessory drive on the front of the engine.  The camshafts are driven from the rear end of the engine by the crankshaft through a set of gears.   6- and 8-cylinder engines use one-piece camshafts, with 12- and 16-cylinder engines having two-piece camshafts.  On the forward end of the camshafts, the left side has a vibration damper (not used on the 12-cylinder engine) and counterweight, with the right side having the fuel pump mounted. 

The governor drive is also driven through bevel gears on the camshaft drive.   The engine uses a Marquette hydraulic governor for operation.  Driven from the back of the camshaft, the overspeed governor is a bit of a complex mechanical/hydraulic device devised by Cleveland, rather than using an additional off-the-shelf Marquette governor like the previous models. When the engine overspeeds, a centrifugal flyweight arrangement closes off oil flow to the small oil pump in the governor, forcing it to build pressure which discharges to a small piston on top.  The piston acts against a spring and controls a set of linkage going to each injector rocker arm.  When the overspeed is tripped, these arms engage onto the rockers, and hold the injector down in the no fuel position. 

Addition 9/2021 –
J. Boggess made the note that I did not catch, in that the 498 engine uses the same style of hydraulic overspeed governor that the 268A family of engine used. The 16-278A engine overspeed has a dedicated flyball thingy that when overspeed RPM is hit, it locks out all injectors until you press a reset button on the overspeed governor.  The 498 and 8-268A overspeeds are self-resetting;  Engine overspeeds, the overspeed governor locks out injectors until the speed gets to “normal”, then it releases the injectors, thus going up and down if the cause of the overspeed is not fixed.

Overspeed operation. Click for larger.

Oil & Cooling system
The 498 uses 3 lube oil pumps, a scavenging pump and a two pressure pumps (one for main oiling, one for piston cooling).  Diesel-Electric engines had an additional scavenging pump installed for the support bearings on the generators.  An additional small oil strainer is mounted on the feed line for the turbocharger bearings. 

The cooling system for the 498 is virtually unchanged from the 278A except for the lack of water- jacketed exhaust elbows mentioned above.   The 498 uses a raw water-cooled intercooler mounted between the turbocharger and the blower. 

Intake & Exhaust
What sets the 498 apart from her sisters is of course the use of the turbocharger. In addition to the turbocharger, the Roots blower is also used.  See description below.  Since the Roots blower is not doing all of the work providing scavenging air, it was found a much smaller lobe length would be required, although they did spin at a slightly higher RPM then those on the 278A.   

The turbocharger for the 498 was furnished by De Laval Steam Turbine Corp. and was a basic “gas turbine driven compressor”.  The Model A turbocharger was supported in its own service manual supplied by De Laval.  The unit used a “monorotor” construction with both sets of blades mounted on a common central hub.  The housing between the turbine and compressor is water cooled from the engine freshwater cooling system.  The engine also supplies lube oil for the bearings, with an optional self-contained oil pump if so required.

The turbocharger on the 498 was mounted to the front end of the engine, with the air intake sandwiched between the turbocharger housing and the intercooler. A duct ran from the intercooler to the bottom intake side of the Roots blower.

An interesting note on the turbochargers.  On most engines, the turbocharger was mounted vertically, as seen in the photo above.  On the batch of engines sold to Cuba (more on this in Part III), the turbocharger was mounted horizontally.  It is unknown why this was done, be it for clearance issues in the building, or some other unknown reason likely lost to history.

Another interesting note, the tug Robinson Bay (again, more on her in the next part), used an 8-498 engine.  However, it appears this engine did not use a De Laval turbocharger, but it looks to be an Elliot-Bucchi design! More questions we likely will never know the reason why to.

The 16-498 engines built for Cuba used a horizontal De Laval turbocharger. The tug Robinson Bay used what looks to be a much smaller Elliot (or so it appears) style turbocharger, but the engine was still rated at 1400HP. (1959 Diesel Engine Catalog Left, 6/1958 Diesel Times Right, J. Boggess Collection). Click for larger.

498 engine plumbing for a Diesel-Electric tugboat (click for larger)

In Part III we will go through every 498-engine built (it was only 58!)

Sidebar: My co-conspirator & former EMDer Jay Boggess & I have concluded that we really started this project about 10-15 years too late!  Too many souls have moved to the Great Beyond – souls that could answer the questions our research has uncovered.  We do not have clear reasons why the 498 didn’t make it (more on this in Part III and IV), only guesses and suppositions and the little bit we have been able to gather talking to guys who worked on these engines in the last few years.  But then, 15 years ago, we didn’t have the internet to bring folks from across the country together, sharing common interests and information. And besides, 15 years ago, I was in junior high living 900 miles away!

Special thanks for this part go to Preston Cook, who sent me a Xeroxed scan of a 498 manual several years ago. I have since been able to acquire several versions of the original manual and service newsletters thanks to Great Lakes Towing Co., who was gracious enough to send a few surplus copies to me when we started this research project. I would love to find a service parts book, and an De Laval turbocharger manual (we only have a photo scan of it) for the Cleveland 498, and would happily pay a good price for them! My contact is in the upper Right of this page.

A Turbocharged Failure – The Story of the Cleveland 498, Part I

June is a the two year anniversary of this blog, and with that I am kicking off a series dedicated to the Cleveland 498 engine. The 498 engine has been shrouded in mystery over the years, and was one of the main driving forces of creating this page. I wanted to do a writeup on the engine, but had no place to put it! Just to put this right on top – if anybody has any stories, recollections, information, photos or documentation on these engines, PLEASE send me a message! I am trying to document these engines as best as I possibly can.

In the days after WWII, medium speed, 2-stroke diesel engines essentially hit a horsepower wall, around 1600HP or so.   A common way or obtaining a higher horsepower rating, was simply to add more engines!  Unfortunately, adding more engines, means more space is being taken up.  So, the solution is to try and get more horsepower out of what you already have. 

Enter turbocharging.  

Now, turbocharging was not a new concept by any means. Many diesel engines benefited by use of turbocharging, but these were almost all 4-stroke engines.  Cleveland Diesel had a single turbocharged 4-stroke engine design during WWII, the 258S (originally a Winton engine) which was a 2000HP direct reversing engine built for subchasers.  Even several WWII aircraft, including the B17 Bomber were turbocharged. Turbocharging a 2-stroke engine was an entirely new concept.  As it is, a 2-stroke requires some form of positive displacement blower for scavenging.  The issue with adding a lone exhaust-driven turbocharger, is in periods of startup, lower idle and acceleration, the engine gets starved for air, as it is not providing enough exhaust to spin the compressor.  Kind of a catch 22 situation.  More on this later. 

The basic operation of a turbocharger from a Garrett-AiResearch manual.

Throughout WWII, General Motors Diesel (Cleveland Diesel, Detroit Diesel & Electro-Motive) was the leading Diesel engine supplier to the war effort.  Cleveland Diesel would supply over 13,600 engines (from 7-1939 thru Dec 31, 1945), be sure to read our history about Cleveland Diesel here: Cleveland Diesel Engine Division – GM’s war hero turned ugly stepsister

Cleveland Diesel WWII Production:
16-278A: 1,930
12-278A: 771
8-278A: 243
6-278A: 554
Total 278A Production – 3,498 engines
16-278: 20
12-278: 352
8-278: 55
6-278: 72
Total 278 Production – 499 Engines
268/268A (all models) – 9,136 Engines
248 (all models) – 268 Engines
Miscellaneous – 239 Engines
Total – 13,640 Engines

The Cleveland 16-278A engine was one of the most widely used engines during the war and peaked at about 1800HP, which was about on par with the EMD 16-567C, which was introduced in 1953.  Alco was already there with their 12-251B, also making 1800HP, however this was a 4-stroke, with a turbocharger already.  Fairbanks-Morse cracked the magic 2000HP barrier in a medium speed engine with the 10-cylinder 38D OP engine by 1950, using only a Roots style blower.  

With General Motors (and Cleveland Diesel) still working closely with the Navy, an experimental test was devised by the Navy’s Engineering Experiment Laboratory in Annapolis, Maryland in 1947 to start testing turbochargers.  A proof of concept test was launched, using a Detroit Diesel 1-71 (yes, GM turbocharging has its roots in the diminutive, little 1-71 engine!).   With the proof of concept done, more testing was devised in the early 1950’s at the Engineering Lab using a bone stock 16-278A engine.   A test was devised in which a mock “turbocharger” (another Roots blower) was installed on the test floor, operated by an electric motor, to feed the engine in a simulated and controllable environment.  A goal was set to maintain a cylinder firing pressure in the area of 1300 PSI (compared to the stock 850-1050 PSI) and make 3,000HP. Numerous tests were conducted with various configurations of inner and after coolers, blower sizes, injectors, controlling exhaust timing and use of snorkels for Submarine use. A similar test was conducted using an 8-268A engine as well.   Unfortunately, I have yet to come across any photos of these tests. 

The winner – Using the stock configured 16-278A engine, with turbocharger feeding the Roots blower with an aftercooler made an impressive 2,990HP at its rated 750RPM.  With controlling the exhaust timing, the engine made 3,130HP.  Amazing numbers for a stock engine! Not to mention, a true testament to the engineering of this engine, and its ability to take such punishment.

Performance ratings for the test engine, from Turbo-charged engines for the Navy, by L. Wechsler and T.W. Shipp, Internal Combustion Engines Branch, Bureau of Ships

After the tests, three turbocharger manufacturers would begin working with the Navy to spec out an appropriate design, and how to supply the air to it, be it via individual ducts from each cylinder (common on 4-strokes), divided manifolds or a single manifold using a venturi system.   The results of the testing were concluded in a presentation at the SAE National Diesel Engine Meeting on October 27th, 1954.  The report, High Supercharging, Development of a GM 16-278A 2-Stroke-Cycle Diesel Engine, was presented by Warren G. Payne and Wolfgang S. Lang of the US Naval Engineering Experiment Station. 

Unfortunately, not all the testing was complete at the time of this paper, so it is unknown just how well the testing progressed when the turbochargers were installed on the engine.  What is known, is that testing further proceeded at the Engineering lab on the 16-278A, The Lanova Corporation handled the 6-71 testing in New York, and De Laval Steam Turbine further tested the 8-268A at their own lab. 

The test 8-268A test engine at the De Laval test lab, used a model B-8 turbocharger. From a 1955 De Laval advertisement.

After the presentation, a discussion panel ensued, which is also part of the transcript of the report, in which comments were heard from other engine builders and engineers.  One such stands out:  Rudolph Birman of the De Laval Steam Turbine Co., who essentially picked apart the findings. Mr. Birman states several things, such as:

“Water cooling of the exhaust manifold cannot be tolerated in a turbocharged 2-stroke engine.”

 “All starting, idling and high exhaust back pressure problems are eliminated, however, if the positive displacement blower is retained and the turbocharger arranged to operate in series therewith.”

“There is a similar disagreement between the findings of the authors and those of De Laval with regard to the location of the intercooler in a turbocharger-positive-displacement-blower in series arrangement.”

I do not know if De Laval were working behind the scenes with GM/Cleveland Diesel already (given the time frame, they must have been), however, Mr. Birman’s commentary would essentially be the entire basis for what would become the 498 engines in just a few short years. 

The concept drawing of the Cleveland 498 first appeared in the August 1955 issue of Diesel Times, along with some basic specifications and features.

Another set of comments worth noting, was from A.K. Antonsen and E.L. Dahlund of Fairbanks, Morse & Co. FM was working in-house on their own turbocharger design, starting in 1945 on a basic 10-cylinder 38D 8 1/8th OP engine used in submarines, as well as a smaller 3-cylinder 5¼” OP engine.  Full production of a turbocharged OP engine was not offered commercially until sometime in the late 1950’s (Anybody have a specific year?). The Turbo OP would be a very popular stationary power engine, and would peak at over 4,400HP for the 12-cylinder engine.

FM’s Turbocharged OP engine is still produced today, producing astronomical amounts of horsepower mainly for standby power generation. Note that like the Clevelands, it retains the Roots blower. FM Brochure

As mentioned above, one of the shortcomings of the turbocharger on a 2-stroke is the lack of enough scavenging air.  The issue was addressed by simply retaining the Roots blower, but it was found a smaller one would work (we will get to this more in Part II).   With the testing on the 268A engine, in place of the blower a small hydraulic motor was tested mounted to the turbocharger.  In periods of low RPM, the hydraulic motor would turn the compressor, essentially making artificial air pressure with the turbo.   The pump for the hydraulic motor was driven by the engine.  

An early Detroit Diesel 6-71T engine used for an industrial application.

With Cleveland Diesel now working on a whole new turbocharged engine – GM sister division Electro-Motive was doing the same.  EMD started their own program in January of 1955 to turbocharge their current 567C engine, unlike Cleveland, they did not start by redesigning the entire engine from the ground up. Like the Navy tests, EMD used an electrically-driven Roots blower in a mock test using a 12-567C engine for development purposes, but EMD would design their own turbocharger for the 567C engine.  Instead of using the combination Roots blower and turbo in series, EMD designed their own all new turbocharger, which would be mechanically-driven from the camshaft through a geartrain during starting, low speed, low power and accelerations, providing scavenging air. The turbocharger is connected to the geartrain through an overrunning clutch.  At certain power levels (approximately Throttle Position 6 on a locomotive), there is enough energy in the exhaust so that the turbo runs faster than the geartrain, the overrunning clutch disengages, providing “free” turbo-supercharging.  This would go on to become a very successful design and used throughout the 710 line (with several refinements of course).   EMD’s first production turbocharged locomotive, the 2400HP SD24, was introduced in 1958. We may do an article specific to EMD turbocharger history down the road, but for now we will stick to the CDED 498. 

The prototype turbocharged EMD 16-567C engine from “Performance of a Turbocharged 567C engine” by A.N. Addie/EMD. Production turbochargers would be used only on 16 cylinder engines, and were given the “D” model. Turbochargers would not be used on 12-cylinder engines until the 645 line.

Union Pacific Railroad was doing their own separate development with adding turbochargers to the 567C used in GP9 locomotives starting in 1955.  Working with Garrett-AiResearch – (later makers of the turbocharger used on the 6-71), a manifold was devised, and four small turbochargers were added feeding into the stock Roots blowers through an intercooler.  UP would also test engines with turbochargers made by Elliot, but using only two slightly larger ones then the Garrett installation.  These tests were successful, and several engines were converted.  UP would send GP9’s to EMD in 1959, which were upgraded with new EMD turbochargers for further testing. Ultimately these test engines were converted to EMD turbochargers, or had them removed.  I urge everyone to read Don Strack’s Utah Rails page on the Omaha GP20’s for much further information on this test program. Please be sure to visit the links below.

Omaha GP20’s, Union Pacific’s GP9 turbocharging program
Omaha GP20’s Diesel Era V7 #6, 11/12 1996

The quad Garrett turbochargers installed on the 567C. Note the complex plumbing for the exhaust and charge air going to the blowers. Union Pacific Photo, Don Strack Collection.

The Elliot installation was a little more simplistic, with a single exhaust manifold feeding a pair of slightly larger turbochargers, with each one feeding one of the blowers. Union Pacific Photo, Don Strack Collection.

The Cleveland 498 made its public debut at the General Motors Powerama Festival.  Powerama was held August 31st-September 25th of 1955 in Chicago, Illinois.  The event, “A Worlds Fair of Power”, would be a giant showcase of products from General Motors, including Cleveland Diesel, Electro-Motive, Detroit Diesel, Euclid, Allison, GMC Truck & Coach, Fabricast and Frigidaire.  On display were numerous engines, pieces of heavy equipment, locomotives, and even the Great Lakes Towing tugboat Laurence C. Turner, and the Fleet Submarine Tautug SS-199.

The first Cleveland 498 displayed at Powerama. I have my doubts that this was a full production engine, as it just does not look “right”, especially the exhaust jumpers and manifold. I think this was more of a mock up model for display. Note the differences just in the cutaway model on the left. The first production engine used commercially was still several months out. Unknown photographer, VDD collection.

Stay tuned for Part II, where we will discuss the 498’s design features and specifications.

This will be a four part series, the links of which will appear for each post as they are added.
Part II – Engine design features & specs
Part III – Uses and installations
Part IV – The last one, Tug Idaho

Note: A complete set of bibliography and notes will appear in Part 4

When Bad Things Happen I

As the saying goes, Ship Happens. Sometimes, worse then others. In todays case, this is a piston and rod on display at the Hoosier Valley Railroad Museum from Erie Lackawanna 310, an Alco S1. The engine in case is an Alco/McIntosh & Seymour 539, a 4 stroke engine with a 12.5″ bore and a 13″ stroke. A valve dropped while the engine was running, and decided to do a little dance in the cylinder.

A Roots Blown Experimental Alco?

Last year, I picked up several rolls of Navy Microfilm full of engine goodies, two boxes of which were marked as “Alco 16-251A Experimental Submarine Engine”. I pulled them out when I got them, but did not go very far, as it is literally every blueprint sheet to build this engine. Thinking it was just another 251, I put them back in the box.

Last night I dove into them a bit deeper, and naturally the LAST frames on the reels (It looked like a roll of film exploded in the living room) had an elevation drawing. Cool! I devised a way to scan these, although frame by frame on my flatbed. It is a project, but it works. I need to draw up and 3D print some holders to do it more efficiently.

Click the following for larger versions.

After studying the drawing for a second, I noticed the exhaust was not connected to the intake side at all. Wait, 251’s are 4 stroke, and have a turbo…where is the turbo? There is none! That’s a roots blower on the front!

Front mounted on the engine is a blower. The discharge from the blower feeds into a raw water cooled aftercooler before going into the intake side of the engine block.

So, naturally, this raises plenty of questions. I can not find a lick of information about this engine in my usual places, so if anyone has anymore clues as to its history, shoot me a message. I don’t know if this was meant as an emergency generator engine, or propulsion. If anyone wants to build one… I have 600+ plans!

Loss of a Museum Tug – Pegasus

It was sad to hear that this past week, the tug Pegasus made her last trip to the great shipyard in the sky. Figure I would throw together a little post about a cool old vintage tug that would meet an unfortunate end this week.

The Pegasus was built in 1907 by Skinner Shipbuilding in Baltimore, for Standard Oil Company, as the S.O. Co. 16. The tug would later be renamed the Socony 16, and eventually wound up as the Esso Tug #1 after several rounds of company reorganizations. McAllister Towing of New York would purchase the steam powered tug, and rebuild her. Converted to Diesel propulsion, an EMD 567 was installed in place of the large engine and boiler. Now renamed the John E. McAllister, she would join the companies massive fleet doing shipdocking and other harbor work. McAllister would also purchase sister tug Esso Tug #2, and rebuild her the same way, now renamed as the Roderick McAllister. Another Socony sister tug – the Socony #14, would find a new home with Philadelphia’s Independent Pier Company, and was renamed the Jupiter. She also is a museum tug in Philadelphia.

Unknown photographer, Courtesy of Dave Boone
Ernie Arroyo Photo, Courtesy of Dave Boone

By the 1980’s, towing companies were selling off the last of the older, converted steam tugs. Numerous smaller companies would benefit from this, and would give many of these older tugs a new life. In 1987, the John E. McAllister was purchased by Hepburn Marine Towing of New York, where she was renamed as the Pegasus.

Photo by Jay Bendersky
Photo by Jay Bendersky

Hepburn Marine would do various work throughout the city, including spending several years towing carfloats for the New York Cross Harbor Railroad. Hepburn would ultimatly charter the tug James E. Witte from Donjon, the former Central Railroad of New Jersey tug Liberty for doing this work – a tug much better suited. Pegasus would be retired in 1997.

The Tug Pegasus Preservation Project was formed, and spent many years actively restoring the tug from the hull up. Volunteers spent several years actively restoring various parts of the tug, and the Pegasus would tow the Lehigh Valley Barge #79 (The Waterfront Museum – see link below) numerous times around the city. I was only ever inside the Pegasus once, a few photos are below.

Pegasus at the 2009 New York Tugboat Races
Inside the deckhouse.

McAllister would repower the tug with a WWII surplus LST package – a 900HP EMD 12-567ATLP, with a Falk (Falk designed, however several contractors during the war built them, including Esco and Lufkin) reverse-reduction gear. This was one of the most common tug repower packages used after WWII, and I am slowly working on a large post about them.

The engine in the Pegasus was originally installed in Landing Ship Tank (LST) #121, shipped by EMD 6/16/1943. LST 121 was launched August 16, 1943 by Jefferson Boat & Machine. 121 would spend her career on the Pacific front and was present at the Marshall Islands, Iwo Jima, The Marianas, Western Caroline Islands and the Tinian Capture, earning 5 battle stars. She would be sold for scrap in 1946.

The Pegasus project fell dormant, and was looking for new caretakers and leadership for several years. Unfortunately, nothing would come to fruition. The museum ship world is one of the hardest aspects of preservation out there, and it gets harder every year as these boats get older. We have lost numerous preserved tugs just in the last few years. Times are tough, but be sure to help support your favorite museum ship. Every one of these groups needs all the help they can get.

Links of interest:
John E. McAllister/Pegasu
s –
LST 121
(Former) Website of the Preservation Project
Tug Jupiter, Socony #14
Tugster posts on the Pegasus
Waterfront Museum (LV #79)