Proto HO PA Load Analysis Testing - December 1998 update.

HOME > TECH BULLETIN > ARCHIVE > LOCOMOTIVES >

Bruce, Tony,

I read your item at the TTE-dcc website about the Lifelike PA motor / DCC problem. Your efforts are greatly appreciated. Thanks!

You are correct in that the low motor resistance is causing most of the problem. A low resistance motor will draw more current when rapidly started and if plug reversed. This will happen with most transistor type controllers, DC and DCC, when the voltage setting is suddenly changed. The effect is masked if an old fashioned rheostat control is used. But another ingredient for peak pulse currents in the motor is the armature inductance (L). The L/R calculated value is the electrical time constant. Motors with a small time constant will draw high peak currents on PWM type power sources, like DCC decoders. LifeLike has undoubtedly selected a more powerful motor for smooth operation while cranking on the big flywheel. The back side is that many high performance motors have low values of L as well as R. As a result they will have high peak motor currents on pulse power throttles and DCC decoders. The PWM pulse width and repetition rate (frequency) is a designer's trade-off within the decoder.

Your test data confirms that all decoders are not alike. Those with low operating pulse frequency or wide pulse widths will have high values of peak current when powering the PA motor. The Lenz LH130 apparently does have peak current limiting. [This may be accomplished by any one of several current measurement schemes in the power H-Bridge circuit and cutting off the FET/BJT devices on a pulse to pulse basis.] This makes the decoder self protective and more robust, exactly what is needed in this case.

Suggestions

1. Replace the PA motor with one from the Proto 2K E8/9 units. These motors have a more conventional resistance characteristic. However replace the motor and flywheel as the bearings may not be able to support the original PA flywheel.

2. Add a resistor in series with the PA motor. A 2 ohm, 3 watt resistor will limit peak currents to about 1.9 amps. The resistor won't get very hot under normal operation as it will dissipate about (0.6 x 0.6 x 2 =) 0.7 watts. This will also resolve the Auto-Reverser problem.

3. Adjust the decoder frequency to be as high as possible. This will minimize the low time constant effects.

4. Ensure by programming that the accel & decel rates of the decoder don't allow rapid changes in motor voltage.

5. Pop for the Lenz LH130 decoder. This seems to be the easiest choice. Let the electronics do the work. If it does have peak current limiting as your data suggests, this will also solve the Auto-Reverser problem. A call to Lenz should verify that fact.

Please keep up the good work. The decoder manufacturers don't tell us enough about their products. It sure would be nice to know which ones work better & why before spending $$. We need a "Consumer Reports" type evaluation here. Are you up to it?

Don Vollrath
Dvollrath@magnetek.com

Proto HO PA Load Analysis Testing.

Tony's on behalf of all DCC folks thanks Bruce for his great effort (used with permission).

There have been numerous reports of these engines causing problems when operated under DCC conditions. Problems reported have been booster shutdowns, decoder-engine combinations causing poor slow speed operations, and decoder failures.

Three Life-Like HO PA engines were tested. One belongs to a friend, and was purchased locally. The other two were supplied by Life-Like for this testing.

The test equipment is listed below.[1]

DC Tests

Motor Resistance

The first test was to measure the DC motor resistance. Each of the three engines measured a low of 4.2 Ohms to a high of 7.6 Ohms. The resistance varied with the position of the armature. These values are consistent with observations by others.

Since the first engine was borrowed from a friend and the three appeared nearly identical, the remainder of these tests were done with the Life-Like supplied engines.

Steady State Current[2]

Running on a level grade with 12VDC applied the engines draw a bit less than 200mA. This value varies considerably with grade, load, etc., as would be expected.

Slipping Current

The engine was prevented from moving, so the wheels slipped. Running on 12VDC, one engine measured 600mA, the other 700mA.[3]

Stalled current

Not measured. Calculated to be approx. 2.9A @ 12VDC.[4]

Startup Current

A 12VDC pulse was applied to a stopped engine. As the engine accelerated from a stop the current decreased from a time(0) high to the steady state value. The initial t0 current for each engine was 2.6A. This value decreased to 1A after approx. 600mS and was under 0.4A after 2 Seconds.

Forward to Reverse Current

The engine was run forward at 12VDC for approx. 6 feet. Then the voltage polarity was reversed, and the current measured. At the moment of reversal the current draw was 4.5A. After 500mS the current was approx. 2.1A. The current did not fall to under 1A until after 900mS. At the end a 2 Second test the current was approx. 0.5A.

Coast Test

This engine has a huge flywheel effect. Turned turtle on the bench and spun up with 12VDC, the engine continued to run for over 11 Seconds after power was removed. On the test track the coast time/distance could not be measured; the engine was still coasting at the end of 6' test track. (>12 Sec.)

DCC Tests[5]

Each engine was tested with a variety of DCC decoders. These were (sequence alpha)

NCE D102 (Ver21)
Lenz LE130 (Ver51)
Digitrax DH140 (Ver94)
Digitrax DH83 (Ver?? Pre FX)

All tests were done with default conditions, default speed tables and 28 SS mode to maintain compatibility across the lines. Current was measured using the internal current sense port in the booster. Peak current was measured using an oscilloscope, average measured using a DVM.

D102

This Class B decoder runs this engine well. With start voltage (CV2) set to zero, the engine develops a good roll even at speed step 1. This engine is an excellent candidate for custom speed tables.

Initial current spikes at SS1 were about 1.6A steady state, average was just over 10mA. The peak value measured was between SS4 & SS8. Peak current was 2A, average at SS8 was 200mA. At SS28, peak current decreased to under 0.5A, average was just over 250mA. [6]

LE130

This is a back-EMF decoder. Watching this decoder output on a scope is very strange. In order to maintain constant motor RPM this decoder measures the voltage generated by the motor during the pulse-off periods.[8] Speed is maintained with a combination of variable pulse timing and pulse width.

At the low SS ranges, this decoder limited motor current to below 1.5 A. Typical peak current was under 1A, and it varied a lot. Average values were consistent with what was previously measured. It would appear there is some kind of current limiting effect here, but it was unclear how this is done, or how the decoder survives it.

DH140

This is a full featured Class B Digitrax decoder. As with the NCE decoder, SS1 caused the engine to take on a considerable roll. By SS12 the engine was rolling at near full speed. What was surprising was the peak currents. At SS1 there were 2.9A peaks. As the SS increased, the peak dropped off. At SS16 they were under 1A. Average values were consistent with the other decoders tested. Scope observations indicate the higher peak currents are due to a higher rise time in the decoder outputs.

Stalled Test

For this test a D102 decoder was used. With the motor stalled, peak current at SS6 was 2.5A. At lower SS's the value was less, due to the rise/fall times. The highest SS tested was SS14, the peak remained the same, and the decoder started to get VERY hot.

Forward-Reverse Test

As with the DC test, the engine was ran forward at SS28, then reversed direction.

For the NCE D102 the peak current was measured at 3.5A (!). The Lenz LE130 was 1.8A peak. The Digitrax decoders will be tested soon.

Each of these decoders handles a power reverse differently. The Digitrax decoder seems to do a near instant full-full voltage reverse. The NCE does a slow spin down and then reverses, in about 2-3 Seconds. The Lenz takes about 10 Seconds to do the same. In an engine of this type the built in time delays will significantly reduce the peak reverse currents, and this was observed. Conversely, instantly reversing the motor voltage will significantly increase the peak current load through the decoders.

"Torture Test"

To see what a decoder would take, a D102 decoder was installed. The engine was ran back and forth quick cycled continuously at SS 28 (This test not recommended, DO NOT TRY THIS AT HOME!!!!! Still its fun!)

After about 2 minutes the decoder began to give off the characteristic smell of hot electronics. After over 3 continuous minutes of this the engine start to run "funny", burping, coughing, stuttering. After powering down the decoder was smoking...burning fiberglass and epoxy, stunk up the whole place. Tests showed the decoder was really fried. Still, it took a lot to destroy the decoder, much more than was expected.

Layout Running Tests.

DC Running

Even with a little MRC 1500 (12VA), these engines ran around the layout just fine. As expected, excellent performers, with no observed problems.

Since the average currents measured were well within normal operation ranges no ill effects would be expected when running with a decent DC power supply. Problems may show up however with some pure pulse transistor output power packs which cannot handle the high peak currents under starting conditions...be careful!

DCC Running

With just one "killer" exception, the NCE Dual 5 Amp booster ran two of these engines just fine. No problems were seen under normal conditions.

With three engines MU'ed, a full speed reverse did cause the booster to shut down once and a while, maybe 1 in 10. Since this would not be considered "normal" operation it should not consider a problem.

With the exceptions of obviously needing custom speed tables both the NCE and Digitrax decoders performed well. The Lenz decoder has a very different speed table and gave superb low and midrange performance. The back-EMF feature worked very well.

The big, REALLY BIG, problem here are booster gap sections. Tests were run with two engines MU'ed. When the consist crosses a gap (each side powered by a separate booster section), AND one booster was set to "LOOP" for autoreversing, sparks started to fly...literally!!! The problem is at low speeds. The peak currents from the motors trip the autoreverse circuit, even though the gaps are aligned. This obviously causes a short, and another reverse...and the process continues. The booster did not shut down, but continued this...it makes a racket, and blows 15 Amp fuses very quickly. Current was not measured, but to blow a fuse this fast requires a 50-100% overload, somewhere over 20A! Additionally, this short current load flows directly thru the loco wiring...not good at all.

This problem would probably be seen with any autoreverse circuit which uses peak current to operate the polarity change. This would be any "in booster" autoreverse, and many of the add on modules as well.[10]

The autoreverse circuit by design, must make this change quickly. The peak current loads of the two engines together are more than enough to trip the 5A sense circuit, so it switches. This is just what it is designed to do. Other than significantly increasing the trip currents, there would appear no way around this problem. Other power boosters would be expected to trip in this manner as well.

Observations and conclusions.

Under DC conditions these engines run really nice, probably the best running "plastic" engine out of the box ever seen.

The large flywheels contribute to this smooth running, but the tendency of these engines to coast does take some getting used to. Running these engines in MU with other types could result in bucking etc. if voltages are not changed slowly.

Under DCC Running the engines run very well using default decoder settings. Fine tuning the speed tables and pulse frequency would result in even better operation.

Under normal operation, most 1A DCC decoders should be adequate to operate this engine safely, even with the wheels slipping. If operations are anticipated where full speed reverse or stalled conditions may occur, then a higher current decoder may be appropriate.

There is more than enough room for any HO 1 or 2 Amp decoder under the hood. A "N" scale decoder would not be recommended in this engine due to the smaller size and lower heat dissipation capabilities.

1-2A BJT output decoders would not be recommended. These transistors just cannot stand the kind of overloads FET's can, and would probably fail quickly.[11]

The autoreverse issue is bad. Two obvious solutions are to run only a single powered engine or avoid running across autoreverse gaps. Neither of these solutions would be adequate for many users. The culprit here goes back to the original issue, the motor. The motors in these engines just draw too much current for operations under these conditions. This issue also points out some serious flaws in the basic autoreverse concept.[12]

Previous tests have shown the Life-Like Proto2000 PA's are not be compatible with in-booster autoreverse circuits. These tests show these engines are not compatible with the external MRC Auto- Reverser either. Running a single PA would normally not cause problems, however running a single PA with any increased current load (MU'ed engines, passenger car lighting, etc.) would likely cause problems with this Auto-Reverser.

Overall Conclusions

These engines are great runners, and fill a big void in the hobby market. Sadly, under DCC conditions, the motors draw too much current, and severely limit their operation.

Special thanks to

Life-Like Products, makers of Proto2000
NCE Corporation, Jim Scorse
Tried and True Trains, Lenz USA Agency, Debbie Ames
Tony's Train Exchange, Tony Parisi
The members and users of the NMRA DCC-SIG

Notes

[1]The Test Equipment

Test Track: 6' HO, Atlas C100/NS Flex, nailed to board.

The Engines

2 Life-Like HO PA
- For DC test used with OEM lamp circuits.
- For DCC test, circuit board removed, decoder hard wired thru 4 pin connectors.
1 Life-Like HO PA DCC test only.

DVM: Fluke Model 77
Analog Meter: Simpson 260
Oscope: Philips PM3267 Triggered Dual Trace, 100MHz.

DC Tests

Power Supply: BUFF, DC Supply 0-30VDC @ 40+ A
Series Resistor (I Sense): 0.5 Ohms, 100W, 3%

DCC Tests

Command Station: NCE Master Series
Booster: NCE Dual 5A
Power Source: Stancor P8687, 18VCT @ 8A+8A (Parallel)

[2]All tests done with engine traveling Forward (Left to Right)

[3]The difference due to one engine "hopping".

[4]Stall current was not measured for three reasons. First, with the low resistance values, it would be very hard on the motor and internal wiring. Second, with the variable resistance measured, it would be mostly meaningless. Third, it can be easily figured using Ohm's law, 12VDC / 4.2Ohms = 2.9A Peak.

[5]DCC Tests:

DCC Decoders configured with the following common CV's:

	1	3	Addr
	2	0	Start V
	3	0	Accel
	4	0	Decel
	29	02	Config, 28 SS enabled
	Default speed tables
	Lenz LE130 run with back-EMF enabled
	All other CV's in Mfgr default conditions

DCC tests done with all functions OFF.

[6] High peak currents at low SS are not unexpected here. This is a low duty cycle pulse, and the motor back-EMF is still low due to low RPM. As the engine increased speed, the back-EMF increases, so the peak currents go down.

As with the prototype, lower "running currents" are expected and observed as the motor spins up and develops back-EMF.

[7] All NCE and Wangrow System 1 decoders are reported to share common output circuits. NCE states these results can be extrapolated across these entire lines. The exception to this are the NCE-Kit decoders, which use BJT outputs.

[8] Connecting the Lenz LE-130 output to either a resistor or a capacitor validates this observation.

[10] A MRC standalone reverse module has recently been received for testing from Tony's.

[11] NCE Corporation specifically recommends AGAINST using the NCE-Kit decoders in these engines.

[12] This issue has been pretty well hashed out. What this engine shows, and the point of this comment is even though the average loads are well within the boosters capabilities, the capabilities of the autoreverse feature are exceeded.

We welcome comments or suggestions from readers; please write or call.