NREC's SD40-4 with Independent Axle Control

[M636C]

While EMD's SD70ACe-T4 seems to have attracted most attention at the recent trade show, it wasn't the only interesting technical advance on show...

http://www.railwayage.com/index.php/trade-shows/first-iac-locomotive-showed-by-nre.html?channel=35

An SD40-2 (number 4321) modified with a separate chopper control system for each of the six motors is interesting from a wheelslip control perspective. NREC are claiming levels of adhesion equivalent to AC traction (which might be true). Of course the motors are still standard DC motors with time based current limits.

This might be of interest for GE Dash 8 and Dash 9 units coming up for overhaul as a lower cost alternative to AC traction, since the principle isn't limited to EMD units...

M636C

[carnej1]

I was under the impression that the "Choppers" are used in AC traction system (AC to DC to AC traction motor) but aren't used with DC traction (i.e AC alternator to DC traction motor)?

[BOB WITHORN]

Perhaps the SD40 has yet another life to live.

[GDRMCo]

So the traction benefits of AC with all the DC downsides?

[M636C]

carnej1 wrote the following post 5 hours ago:

I was under the impression that the "Choppers" are used in AC traction system (AC to DC to AC traction motor) but aren't used with DC traction (i.e AC alternator to DC traction motor)?

The devices used in converting DC to variable voltage, variable frequency AC are generally called inverters.

The name "chopper" is usually used for a device that converts DC to DC, usually with a variable voltage.

The standard DC series wound traction motor varies its speed according to the voltage applied to it. Thus in conventional DC locomotives motors can be connected in series to reduce the voltage across each motor to give the desired motor speed.

With independent chopper control, the voltage across each motor can be controlled independently. The motors are effectively all in parallel with respect to the alternator but the choppers provide each motor with the desired voltage for the required speed, inluding cutting power to a motor that has been found to be slipping while leaving full power on the others.

Since NREC say that the major components aren't altered, this SD40 will keep its AR10 which provides a DC output to the choppers which in turn feed the D77 motors.

NREC appear to be claiming improved tractive effort which could only be due to taking the motors closer to the onset of wheelslip while relying on the detection of slip and the ability to control the slipping axle while keeping power on the other axles to provide the adhesion improvement.

In comparison, EMD's super series system for DC put all the motors in parallel at all times but used conventional control, which required a higher current capacity in the alternator at low voltages to feed the motors at low speeds.

While the cost of power electronics is falling, a chopper per axle can't be cheap and is the improvement to an SD40-2 worth the cost since it doesn't have the power or tractive effort of a recent AC locomotive?

This is more likely to be of interest to smaller railroads, although it might be picked up by major railroads like NS who are already investing in rebuilding their existing fleets.

If this system can increase the reliability of older locomotives by removing contactors used in transition and other analog aspects of the existing control system and provide quicker repair by exchange of the chopper modules, it might extend the life of older units.

But it might end up being as useful as a hip pocket on a tee shirt...

M636C

[GDRMCo]

Seems this loco has been around for a couple years now, might be now they're looking at a finished product. NRE is claiming a 33-35% adhesion factor which would give a 390,000lbs locomotive a starting tractive effort rating of 128,700-136,500lbs which is a significant improvement over a normal SD40/40-2. They also say this system can be fitted to any DC locomotive (they mentioned SD60, SD75 and Dash 8 and 9).

[Leo_Ames]

Was CSX considering something like this? As I recall, the SD40-3 program initially was going to offer a significant improvement in adhesion before plans were scaled back. 

[M636C]

GDRMCo wrote the following post an hour ago:

 

Seems this loco has been around for a couple years now, might be now they're looking at a finished product. NRE is claiming a 33-35% adhesion factor which would give a 390,000lbs locomotive a starting tractive effort rating of 128,700-136,500lbs which is a significant improvement over a normal SD40/40-2. They also say this system can be fitted to any DC locomotive (they mentioned SD60, SD75 and Dash 8 and 9).

 

My ideas are heavily influenced by my experiences in the Pilbara. I remember a BHP Billiton train from 2006 when AC units were still in the minority. This train was two standard rakes and should have had 8000 HP per rake, but the lead pair was a Dash 8 and an SD40R, the trailing pair was two Dash 8s. There was no problem with adhesion but the train was down 1000 HP and climbed the Chichester ranges more slowly than imtended. By Chichester siding, the train had to stop for 45 minutes to allow the DC motors to cool because with the lower power, the time on grade exceeded the time the DC motors were allowed to run at the rated current. This gave me an insight into BHP Billiton's move to AC traction. It was mid winter by the way, not that it wasn't warm...

 

Thirty years earlier I had ridden a train of 144 vehicles hauled by three Alco M636 and C636 units up the same grade when it rained (the only time I can recall apart from cyclones). The locomotives had no sand and we were getting frequent wheelslips on the trailing units. You could tell that a trailing loco had slipped a second or two before the bells rang because you could feel the shock as the whole load transferred to the two remaining units. The driver applied the independent brake on the three units to reduce the wheelslip, and at Chichester siding he asked me to walk back one side and check that none of the loco wheels were cracked, (but under no circumstances to touch them) while he checked the other side. None of them were cracked, and we made it back to Port Hedland. But looking at the motors from other units being replaced in the workshop, the commutators looked pretty worn out and the wheels were worn hollow on the side away from the gear, because the twisting of the axle allowed that non gear wheel to slip more...

 

In the USA the problem of heat won't be as great, but the D77 can only run for so long at a high current, unlike the AC units which can, in theory run all day at continuous rating with no problems. Had there been two AC units on that train in 2006, the DC units could be shut down and the train could have kept moving slowly to the top of the grade and continued to the port without occupying a siding on a (then) single track line.

 

But you can't equate tractive effort with load hauling capability unless the track profile allows the dc motors to keep within their heat limitations.

 

M636C

[GDRMCo]

That's the problem that I see, this really does appear to be nearish AC traction power but with all the downsides that steered RRs away from DC and into AC. There's so many wear parts in the DC motors that putting them under further abuse by allowing them greater ability to lug may burn things up faster. Gains on hauling ability but losses on time between maintenance.

[Wizlish]

M636C

The devices used in converting DC to variable voltage, variable frequency AC are generally called inverters. The name "chopper" is usually used for a device that converts DC to DC, usually with a variable voltage.

Wouldn't a 'chopper' be a device that either performs DC pulse-width modulation (PWM) or approximates 'variable voltage' to DC motors by rectifying AC to 'pulse' DC with a device like a SCR that allows turnoff down to zero crossing and then automatic reset (and then feeding the result through appropriate big LC filters to get rid of the resulting perhaps amazing ripple)...

[M636C]

Wizlish

 

 

M636C

The devices used in converting DC to variable voltage, variable frequency AC are generally called inverters. The name "chopper" is usually used for a device that converts DC to DC, usually with a variable voltage.

 

 

Wouldn't a 'chopper' be a device that either performs DC pulse-width modulation (PWM) or approximates 'variable voltage' to DC motors by rectifying AC to 'pulse' DC with a device like a SCR that allows turnoff down to zero crossing and then automatic reset (and then feeding the result through appropriate big LC filters to get rid of the resulting perhaps amazing ripple)...

 

I've always been under the impression that the device supplied by AC tended to be called a "Thyristor" and one fed by DC was called a "Chopper".

 

Certainly the Swedish electric locomotives of the Rc type using 15kV 16.66 Hz were generally known as Thyristor locomotives while the Italian E444 005 operating under 3000 V DC was called a "Chopper" unit (or in that particular case "Full Chopper").

 

M636C

[Wizlish]

M636C

I've always been under the impression that the device supplied by AC tended to be called a "Thyristor" and one fed by DC was called a "Chopper".

A thyristor is just the name for a solid-state version of a thyratron tube; I don't think there is much if any difference between it and a gate-controlled SCR.  However, I think there is an important difference in how the gating physically works between the tube and the SCR: in the thyratron, I believe the gate signal is used to induce avalanche conduction through the tube down to zero crossing of the AC, whereas in many SCRs the gate turns off conduction until the device 'resets' at zero transmission current.  The net effect in both cases, though, will be to "chop" the AC waveform effectively twice per cycle, reducing the average (RMS) voltage (without necessarily affecting the peak voltage if the gating occurs anywhere in the falling part of each half-cycle).

The chopper is simply a glorified switch, turning the DC current on and off a certain  number of times per second and reducing the effective voltage seen across a connected load. Note that because this is DC there is no 'automatic reset' and the switch has to be physically clocked on and off, not just gated by reference to a falling voltage threshold as in the AC signal.  This implies a separate clock circuit, etc.-- but that is trivial to provide with 'modern technology'.

This is normally done at very high frequency, perhaps thousands or more pulses per second (which reduces some of the need for large LC "filters" to smooth out the resultant DC 'ripple', voltage with capacitance and current with inductance).  Occasionally this will make motors 'sing' in resonance or harmonic with the pulse frequency.

Note also that a chopper can produce pulse-width (the same effective thing as pulse-duration in this context) modulation, if the switched on-time is not the same as the switched-off time or ifeither of the two is not constant over time.

[Paul Milenkovic]

Wikipedia on "chopper" https://en.wikipedia.org/wiki/Chopper_(electronics)

A chopper is described as any of a number of electronic circuits using any of a number of types of on/off electric switch.  The purpose is to effect a pulse-width modulation of a DC voltage, which after averaging, supplies a different DC voltage.

A thyristor (or silicon-controlled rectifier (SCR) or triac) is a type of electronic switch that can be turned on with an electrical pulse and needs the supply voltage to return to zero to turn off.  "Back in the day" of electric railroads contemporary with the 1960's Japanese Bullet Train (Shinkansen), electric power came from an AC source that need to be modulated or controlled in some way to regulate the speed of the traction motors. 

The prior way of doing this, especially with DC electrification such as interrurban railroads and transit lines was the "cam controller" (so described in Railway Age during that era), essentially selecting the traction motors in series or parallel, with different "taps" on the motor fields are the introduction of resistors for "field weakening", effecting a variable-speed motor control in "steps" or "throttle notches."

What was then called a "chopper controller" took advantage that the DC waveform after rectification of AC was not a constant voltage but still had ups-and-downs in it that brought the voltage to zero to allow turn-off of a thyristor switch.  These choppers could accomplish turn-on over a variable interval of the AC-cycle present after rectification.  The sales advantage of chopper control is that you didn't have the lurches of the throttle notches in the conventional, cam controller that made discrete selections of motor connections and winding taps.

Some while ago I built something similar for my HO model trains using a light dimmer.  The light dimmer uses a triac switch to accomplish turn-on during a variable interval of the AC cycle, where turn-off happens when the AC waveform crosses zero.  I followed the light dimmer with a door bell transformer for step-down to safe model train voltage followed by a bridge rectifier to operate a DC model train.  It was a little buzzy in the model train motors, and owing to the properties of how triacs are triggered, it didn't reach down to barely-creeping train speed, but it was an improvement over the rheostat "throttles" that I could afford in the 1970's when I started with model trains.

More recently, I joined a model train club specializing in Flyer 3-rail O and 2-rail S with AC power, and my colleagues had come up with the light-dimmer-to-control model trains idea independently, putting a light dimmer on the 120 VAC side of the variable transformer used to run the trains.  They called this device a "reducer" as it limited the trains speed to a safe value to stay on the track when the transformer was operated by the public in our Participation Layout.

We gave up on this light-dimmer reducers because it seems that our AC motors were overheating -- instead, we put a tape mark on the transformer and rely on parental guidance of our participation audience (along with a Plexiglas border to the layout) to prevent trains from flying of the rails and onto the floor.

So this is what a "thyristor" chopper used-to-be in prototype and model electric trains.  The AR-10 alternator supplies rectified DC that can supply such a chopper.  But what "these guys" are doing could be anything because there has been in recent years a revolution in power electronics, where you can get high current switches that can turn off as well as turn on, and you can "chop" a smooth DC instead of a rectified output of an alternator.

But I am guessing that if what these guys are offering is some affordable retrofit for an SD-40, they are using a 60's style electric railway thyristor chopper on each axle, fed by the full-wave rectified DC waveform that returns to zero twice each AC cycle of the AR10 alternator.  I doubt it has the technical sophistication of "filtering" the alternator output to make it smooth and then chopping it with high-current MOS-FETs or other on/off capable switches at high frequencies.

[Paul Milenkovic]

This page http://www.railway-technical.com/tract-02.shtml explains what I have been saying about a thyristor chopper being suited for rectified DC supplied either from an AC overhead wire or a Diesel-engine alternator.  It goes on to explain the electrical trick of a resonant circuit to permit use of a thyristor chopper with a DC electric supply.

The article goes on to explain about the newer high-current switches that can be turned off without the supply voltage having to go to zero as when the supply is AC or rectified DC -- there are graphs illustrating what I mean.  But since they are just regulating a voltage -- per axle -- from the rectified output of the AR-10, these guys could easily use the older-type thyristor chopper.

[erikem]

M636C

While the cost of power electronics is falling, a chopper per axle can't be cheap and is the improvement to an SD40-2 worth the cost since it doesn't have the power or tractive effort of a recent AC locomotive?

I suspect that the modules can be made for a fairly low cost. The guts of the chopper would be one switch, most likely an IGBT , and a free-wheeling diode (e.g. PWRX CM1200E4C-34N IGBT/Diode for maybe $1500). There may be a requirement for some sort of low pass filter to minimize problems due to the D-77's not being designed for chopper service (which would probably cost as much as the IGBT/Diode module, but cheaper than a new motor). My guess is that the BOM would be between $5k and $10k per motor.

 - Erik

[daveklepper]

I'd like to bring some EE's understanding to this discussion.   Chopper control, of itself, does not allow the dc motors to produce more tractive effort, over all.  In some conditions the rapid on-off application of torque may in fact REDUCE tractive effort (before wheel-slip).  What this new system does do, however, is match the overall average current and average voltage at any one time to the needs of the particular axle and motor, and this is very difficult and expensive to do with conventional AC-generator - DC motor technology.  And this independent control of each axle is what makes for greater lugging power, greater tractive effort. Also, it permits the AC genrator *(alternator) or even a DC generator, to rotate at its most efficient speeds for the diesel engine to operate, thus saving considerable fuel.  This cannot be done with conventional dc-motor technology.   With chopper control, transition should not be necesary, the motors can be connected in parallel even at low speed.  Not requiring transition is a feature of ac technology (and so is optimum diesel and generator rotational speed) arried into chopper dc technology.  This smoothes operation, from my own observations.

Chopper control was originally developed for transit cars, from streetcars to commuter cars, to eliminate the wasting of energy in grid resistors for speec control.  Here, also, it has been largely replaced with ac-motor technology.

[D.Carleton]

Obviously, the question on my mind: does this provide any advantage for passenger locomotives? Could this aid acceleration or not make any difference?

[daveklepper]

When you are discussing new locomotives, the expense of going all-AC woiuld not be that much greater.  To me, the dc-motor road diesel is obsolete, and I think ac-motor diesels have enough advantages to make that technology the winner except possibly in light industrial service, but definitely includng long distance and commuter passenger service.

Retrofitting an existing dc-motor diesels makes sense for commuter and most passenger service and costs a lot less than rebuilding for ac motors.  Smoother operation, fuel savings, and less  maintenance.  Not so many pneumatic and/or solenoid high-current contactors/relays.