Improving DC performance

[Atlantic Central]

For many years now there has been talk of needing to remove the capacitor or capacitors that are across the motor leads of Bachmann locomotives when installing DCC decoders.

Some recent testing with my Aristo Craft Train Enginner throttles, several conventional power packs, and a number of Bachmann spectrum steam locomotives has yielded some interesting results.

Two Spectrum locos in particular have always suffered from a reputation for less than smooth starts and a slightly high slowest speed. The locos in question are the DC versions of the 2-8-0 and 4-6-0.

While they do perform better on the Train Engineer than on conventional power packs, they still did not equal the performance of many other models, by Bachmann and others.

So I got to thinking about why, since the TE output is a pluse widith modulated signal and makes everything run good, why did these two loocs still seem a little balky?

So I cut out the capacitors, and presto! They now run as slow or slower than any other loco I have and start much smoother. And, the improvement was seen on both the Train Engineer throttle and several different MRC power packs I have.

So all these years these Bachmann locos have been hiding their true potential for DC operation. Well, I'm off to the work shop, I have about 20 more tenders to disassemble and caps to clip.

Sheldon

[Jim Banner]

Very interesting.  Were any of the power packs non pulse?  Anything like an MRC Tech II for example?

Jim

[Atlantic Central]

Jim,

The three conventional packs where:

MRC Tech4 280 Dual pack

MRC TechII-Railpower 2400, with the pluse power "off"

MRC Railpower 1370

Slow speed starts and slow speed operation was excelent on all three of these and on the TE.

There is no doubt the cap/coil circuit is distorting the desirable characteristics of all of these throttles. Removing the cap negates the effect of the circuit and removing or bypassing the related inductence coils does not appear to have any additional benifit.

Both locos started smoothly and ran at 1-3 smph speeds with no problem, in both directions.

Sheldon

[Yampa Bob]

I haven't worked with resonant circuits for 30 years, and only vaguely remember that a positive reactance implies inductive, a negative reactance implies capacative. If I refer to the formulas I get one of PD's headaches. 

When I first learned that Bachmann used noise suppression circuitry on the DCC  board, I just assumed the capacitors and inductors created a resonant filter to cancel out the noise and never gave it further thought.

We often advise others to clip the capacitors if they experience erratic performance, but inquisitive minds might be asking "Why am I doing this?"

[Jim Banner]

In general, an inductor (choke, coil,inductive reactance, L, or other names) in series with a resistive load stops high frequencies from getting through to the load and a capacitor (condenser, capacitive reactance, C, etc.) across the same load absorbs any high frequencies that make it past the inductor.  Together they form a low pass filter, which is what Bachmann uses in their steam locomotives.  Such a low pass filter lets low frequencies, including dc, through but stops high frequencies that could interfere with your TV and maybe your radio from escaping from the motor.  If they did get out of the motor, they would use the track and railroad wiring as an antenna to broadcast the interference through the air and could make your wife or neighbours unhappy.

The output transistors in decoders are very small and cannot handle a great deal of heat.  So when they are used to control a motor, they do it by switching the power to the locomotive's motor on and off.  That is all they can handle, just on and off, nothing in between.  The longer they are on (compared to being off) the faster the motor runs.  The shorter they are on (compared to being off) the slower the motor runs.  This is called pulse control or pulse width control.  You can actually try this out with your ceiling fan if you have one.  Turn it on for 1 second and off for 9, on for 1 second, off for 9, over and over.  The fan will run, but slowly.  After all, it is getting power only 10% of the time.  But its momentum keeps it turning.  Now turn it on for 9 seconds and off for 1 second and keep repeating that.  The fan will turn at almost full speed.  Not surprising as it is getting full power 90% of the time.

The advantage of this method of control is that the decoder output transistors are either on, which means that the power they dissipate is 0 volts across them x full current through them = zero power, or they are off, which means that the power they dissipate is full voltage across them x zero current through them = zero power.  Actually, this is not quite true - the transistors have a tiny voltage across them when they are on so they do dissipate a tiny small amount of power when on.  In addition, the transistors have both voltage across them and current through them during the transition times from off to on and from on to off, so they dissipate some more power during the transition times.  But as long as the transition are very quick, this power is small too.

Now here is the rub.  Remember those capacitors across the motor?  They need to be charged and discharged with every pulse fed to the motor. This charging and discharging takes time.  It radically increases the transition times, and suddenly the output transistors are dissipating a lot more power during the transitions.  This causes them to overheat and the decoder protects itself by going into thermal shutdown until the transistors cool.  It is the cycling between normal operation and thermal shutdown that makes the decoder's operation erratic.

(This problem can also be analyzed in the frequency domain where the high frequency components of the pulses are effectively shorted by the capacitors, leading to the same results.)

Bachmann's solution to this dilemma is to use slower pulses to the motor so that there is less heating of the output transistors.  The problem is that the slower pulses puts them in the frequency range that the human ear can hear.  Some people find this objectionable, particularly in steam locomotives.  The way to get rid of the noise is to use higher frequency pulses as found in "silent" or "ultrasonic" decoders.  But then you annoy your dog, send bats crashing into things, and get poor motor performance unless you remove the capacitors.  Remove the capacitors and then all you have to deal with is a mad dog and battered bats.

Stan Ames and I hashed out this problem over a decade ago and one of the first, or maybe the first, explanation on the web was at the link below, about half way down the page where it starts off with "Poor, erratic DCC speed control..."  For a number of years afterward it was often referred to as "the Saskatoon solution."

http://members.shaw.ca/sask.rail/dcc/2-8-0/index.html

As to why removing the capacitors improves dc performance, I do not know.  I was surprised to hear that it does.  Bob mentioned resonant circuits, and in fact a low pass LC filter does become a series resonant circuit at high enough frequencies.  But with the small coil involved, that would be a very high harmonic of the 120 Hz component in the conventional power pack output, such a high harmonic as to have virtually no energy content.  The capacitor(s) and the coils in the armature would resonate at a lower frequency, but that would be a parallel resonant circuit which would go to a high impedance at resonance, not a short circuit.  I may yet be joining the society of the mad dogs and bashed bats as I bat my head doggedly over this one.

Jim

[pdlethbridge]

oh my head

[Yampa Bob]

I'm looking at the circuit board in a Bachmann medium Vanderbilt tender. The two inductors are readily visible, but no obvious capacitors that can be clipped. There are 4 tiny flat caps on the board, it would require micro-surgery to remove any of them, and would probably destroy the traces in the process.

When I tried replacing the "on board" Bachmann 44915 decoder with a Digitrax DH123, the loco would not move, no lights, nothing. I reinstalled the Bachmann decoder and said "to heck with it".

My next frustation was with the USRA Medium tenders. I ordered 4 to replace the Vandy tenders on my Connies. Neither Bachmann or Digitrax decoders would work. Apparently I received some really old versions. According to one source, all these tenders had the 8 pin socket wired wrong.

Bachmann replaced the tenders, although the board appears to be the same, the new ones will at least work with Bachmann decoders. I notice Bachmann has discontinued the USRA medium tenders.

Frankly I'm a bit weary of modifying boards so I can have a choice of decoders.

[Atlantic Central]

Bob and Jim,

I know why, I just don't know the values involed.

The Train Engineer works the same way a docoder works, a transistor is switched on and off to power the motor. Depending on it size in Farads, the cap "softens" the pulses into more of a steady signal. The problem there is two fold. That rasies the effective voltage and kills the torgue producing leading edge of the pulse causing a higher starting speed for the motor.

As for the "conventional" DC packs, they must have some sort of motor speed control circuitry that is similar and being effected the same way.

In fact, the TE has a "linear" setting in addition to its pulse width control setting. I suspect the linear setting just puts some big cap across the outputs.

Bob, All my Bachmann tenders have a yellow disc cap (sometimes two in series) right where the motor wires leave the circuit board. It is marked "104" or "105"?

Sheldon

[Atlantic Central]

Bob,

You even called out the two caps on your photo in your post "pictures and diagrams".

Sheldon

[Yampa Bob]

Sheldon,

The capacitors I noted in the first post of that thread are in the USRA Medium tender. If you refer to the 6th post, there is a picture of the Vandy tender, with a note "no capacitors above board".

http://www.bachmanntrains.com/home-usa/board/index.php/topic,7439.0.html

Here is a repeat of a post I made later in the same thread.

I found an "Important Notice" packed with one of my Spectrum 2-8-0. The notice has not been included with any of my other Connies. Thinking that others may not have seen the notice, I shall include it here: (quoted exactly from the notice).

IMPORTANT NOTICE

"In accordance with strict international electronic emission regulations, Bachmann locomotives are equipped with suppression capacitors and inductors (RF chokes) to reduce interference with radio and television signals.

NMRA conformant DCC decoders,including Bachmann E-Z Command decoders, will work properly with these capacitors and inductors, but these components may cause a malfunction of decoders that do not conform to NMRA standards.

To prevent damage to locomotives equipped with decoders that do not meet NMRA standards or are not part of Bachmann's E-Z Command line the suppression capacitors and inductors (RF chokes) on this locomotive must be removed from the circuit board before non-NMRA decoders are installed."   (end quote)

(I added emphasis in last paragraph)

[Atlantic Central]

Bob,

I see that now, well my vandy tenders have them, not sure why yours are different. Anyway, I have found them in every Bachmann tender I have, long vandy, medium vandy, hicken, short tall tender from 4-6-0's, USRA medium and the regular tender on the 2-8-0.

They will all be gone by tonight!

Found a similar situation in an IHC Pacific this morning - fixed that too. Now it runs much better as well.

Sheldon

[Yampa Bob]

Sheldon,

I added to my last post, you might have missed it in the interim.

[Atlantic Central]

Bob, that's interesting, I've never seen that notice in any of my locos.

But, since I remove decoders rather than install them, I might not have read to closely anything that looked like DCC instructions.

I will post any further results as I check the rest of the fleet.

Happier than ever with my wireless DC control system.

Sheldon

[Atlantic Central]

Sheldon,

The capacitors I noted in the first post of that thread are in the USRA Medium tender. If you refer to the 6th post, there is a picture of the Vandy tender, with a note "no capacitors above board".

http://www.bachmanntrains.com/home-usa/board/index.php/topic,7439.0.html

(I added emphasis in last paragraph)

Bob, now that I have checked most of my fleet I have found the reason your vandy tender has a different circuit board. With the introduction of factory supplied decoders, the 2-8-0 was changed from 12 volt headlights to LED's.

This required changing the circiut board. It does have the capacitors, but they are small PC board types that look like just like PC diodes. I removed them from my three connies that came with decoders (and I removed the decoders as well) and achieved same improvement in slow speed.

They are marked c3 & c4 right near the motor wires. Removing them will likely make your other decoder work as well.

Sheldon

[Yampa Bob]

Sheldon
Thanks for the info. While tracing the pinouts for my thread, I assumed C3 and C4 were the culprits. There is also C2, but I think that is a "keep alive" for the headight.

Even with a strong mag visor, I have difficulty seeing the traces. Did you clip the caps from the top, or unsolder from the traces on the bottom?   I have a small flush cut end nipper that might quickly do the job. 

All my Vandys have Bachmann decoders installed, as such they run very well. If I decide to remove the caps, think I will also bypass the inductors. If I mess up the board, I can always hard wire a decoder.