Schematic for Troller transformers

[quaddriver]

I have entered a project for the fun of it to map out the old trollers, improve them and fix them. For the people who cannot afford DCC or have the eyes to wire it into 40 year old + locos, we still use DC. I was enamored with the trollers circa 1979/80 and my first was twin momentum 5 which still runs.

Over the years a number of people pointed out quality problems, like 95% of it was due to pizz poor heat sinking of the main transistor and the 'auto pulse circuit is a simple semi-effective design. Using COTS parts from mouser - 99% I have in stock from my stereo repair biz, I have been able to armor these. As I do classes of them I am also using tiny-Cad to create schematics. you can apply this to MRCs in most cases and if you are feeling froggy, transistorize just about any DC ‘rheostat’ pack.

I will attach the schematic for a Transamp 1, which applies to ANY of the transamps even the twins, just ‘times 2’ the guts and go larger on some stuff.

Note, on the older DC locos, even athearns, atlas etc, solder a .1uf 50v or higher MLCC (multi layer ceramic cap) across the motor leads as CLOSE to them as you can get. this gets rid of the ‘sparky spark’ interference they get on bad track or bad pickups which can blow out LED light conversions…

If there is some place this can be stored on the site?

[gregc]

why not post it along with other files on github?

couldn’t trolleys use PWM throttles? This design looks like the early circuits described by Linn Westcott in the 60’s that requires higher power transistors

[quaddriver]

anyone could use a PWM, but the simple models for sure did not. They just simply take a leg of the transformer, use a 220ohm to knock current down, and sum it with the throttle output from the 1600ohm pot. at no or very low settings the voltage does not exceed Vebo so the darlington does nto conduct (typically 1.2-1.4v) as it climbs a very dirty DC signal biases the output trans some of the time (60-120 times per second) and off she goes, as throttle climbs it mostly smooths out, but these run warmer than pure DC. The mod I made with the small cap tries to keep the darlington ON, and with the pulse at low speeds it has low DC and very low DC, but voltage over 1.4v for a longer time.

Since the ‘ripple’ voltage was overwhelmed by later throttle settings (like scale 30mph and up) they got to call it ‘auto’, when in fact its really ‘sheer dumb luck autopulse’

Keep in mind, in THESE instances the power transistor being a darlington has very large gain, the TIP120 is on the order of 1000. The TIP is rated at 5 amps Ic and 65w Pc so sink it well, it should last. Of all the ‘bad’ trollers I have thus far, none are failed open or shorted on the darlington, its been earlier in the momentum circuit.

To understand these, job 1 HAS to be to unsolder the circuit board from the output jacks and divorce it, that way you can at least see it. The parts are either unmarked or marked with proprietary things, so Im essentially reverse engineering the momentum which is really when engaged, a freezing of the capacitor voltage when the switch was thrown, and a time delay circuit to add to or subtract from it. since its ALSO dumb, larger changes in throttle setting per unit time means faster response. (i.e. crank it up from tie crawl to full blast and it will climb voltage faster than just going to half speed) You can ALWAYS gut these down to just simple transistorized DC throttles and add a post box momentum circuit form the back…(do they even make those anymore? I am sure schematics are somewhere)

[gregc]

believe the daigram below is the equivalent of yours.

the transistor is wired in the emiiter follower which means the transistor will pass as much current as necessary to maintain the emitter voltage at 2 diode drops below the base voltage.

This is a common approach with linear regulated power supplies where the pot or base of the transistor is supplied a constant voltage (i.e. zener) in order to maintain a constant output voltage despite varying supply voltage (using a cap)

the circuit without the cap will maintain the output voltage at a fraction of the sinusoidal supply voltage.

as you suggested, the cap makes the supply voltage more constant with ripple. So the pot maintains the output at a fraction of the more constant supply voltage along with it’s ripple

the 220 Ohm resistor provides a 60 Hz pulse to overcome stiction

[quaddriver]

ok I have not seen this one nor the one you referenced ‘from the 60’s’ but there is an error in the 'pulse function on this one…

In THIS drawing the 220 ohm resistor has its 60 hz sine wave as a low current (v/220) with the output of a full wave bridge which shows up as a 120hz period of humps. But this is implied full output and indeed it is at the top of the pot and the collector of Q1 (the darlington) since the voltage is NO MORE than the output of the bridge, VRMS is unchanged (and indeed the effect of any pulse from the AC tap+resistor is smoothed by the cap) and during the ‘off’ period where the tap is negative, it could/would deduct from the bridge output, IF the smoothing car was not there.

The purpose of the ‘auto pulse’ is to operate only at low speeds due to the heat they cause the old motors to run rough.

On the troller the AC tap and resistor go to the base of Q1 (pin 1 in your diagram) and the collector gets the DC output - a somewhat smoothed 120hz set of humps in my improved version.

At no to low throttle, the greater signal from the pulse tap is not enough to bias Q1 and nothing moves, as throttle comes up and Q1 starts to conduct at some ‘angle’ of the humps, the sum of the fixed current signals’ voltage drives a little higher that throttle alone would and on the next period it woudl deduct, perhaps to the point of non conducting. at high throttle settings, where a realist would not run, say ‘tyco speeds’, the operating region of Q1 would never be violated, and it would always conduct.

Now since I posted that, this afternoon I opened up a momentum 1, model 34, which is transamp1 horsepower, but with a bell and whistle added. On this the pulse tap is done differently, I dont have it tabbed out yet, but it appears this may actually be a crude PWM where a small npn something (it looks like 2n4014 but I need a glass) is reverse connected to the emitter such that it only conducts on the second period of a 60hz ac tap, but the collector has a diminishing value and throttle voltage climbs, stay tuned. This one appear to be newer than a twin m-5 I have with a fully on throttle and has better marked internals. Troller may have sourced parts based on ‘who is on sale today’, I got a bunch of 8A TIP142s coming from mouser this week.

Otherwise, while this M1 works, I am going to armor it and it WILL be different, using the ‘mistakes’ from the first one to improve the process

[gregc]

@quaddriver

In THIS drawing the 220 ohm resistor has its 60 hz sine wave as a low current (v/220) with the output of a full wave bridge which shows up as a 120hz period of humps

in your drawing, the 220 Ohm resistor is connected to the fuse from the 60Hz supply, not the 120 Hz output of the bridge

i see the mistake in my drawing. The 220 Ohm resistor is connected directly to the base of the transistor.

without the cap, the 120 hz half-wave rectified waveform from the bridge rides on top of the 60 Hz waveform thru R, when the pot center lead moves away from ground, but i think its effect is also reduced when the pot is moved towards its upper limit

with the cap, the filtered half-wave has a 60 Hz component.

[quaddriver]

Update on my momentum circuit findings! This version has markings on the 3 TO92’s on the board, there are 2 2N4401, and one 2N4403, get this…not only are they ALL still available, but they can accept TO92-3 aka to-92Ls. The Ic is 600ma which is VERY salty for a to92 case. I am looking at replacing them with KSC2383 and KSA1013 <—I might have its compliment wrong, all of which are available from onsemi thru mouser. these are 1a Ic devices with a 900mw Pd (Pc), the O suffix should suffice but I stock the Y suffix to keep hfe in the same ballpark and they are triggered of a 60 and 120hz pulse at WORST so the Ft is way more than adequate. There are a lot of 10K-class resistors in there so the current will always be in the mA range. Vebo of the KSC/KSA is 6v and 160v vs 5 and 40ish so any backlash in there should be absorbed. its gonna be a while before I get this tabbed out and all the connections worked out so still stay tuned. From a glance, the TM5 model I have looks a LITTLE different, so there might be un-annotated revs in there that troller did…figures… also the stock devices are EBC and all my replacements are ECB so a bit of leg crossing is called for.

In short, fixing one should not be a problem at all, all the parts are COTS.

No more junk piles for old trollers…

[quaddriver]

This is another reason why we desolder the circuit board from the troller outputs and put it elsewhere…40 years or so later the corrosion of ‘just existing’ can get in the way. On my last unit I used std 1/4" female blades in insulation. you can double tap in the Mom. 1 the connectors for the AC output and run wires back to wherever you put the circuit board (dont be afraid to drill a 1/16 hole in each copper pad) as in this model, since all the gizmos are on the board, 2 wires of AC go directly to the board AND AC accessory terminals and no where else.

[quaddriver]

If this worked, there are 2 pages of schematic for a troller Model 34 aka the ‘Momentum 1’.

If I wanted to be tricky I could squeeze it all to one page. Note that the first ‘half’ of this is just like the model 20, aka Transpack 1, cept in the 34 the Darlington was a BKW23? Its real similar to the TIP120 BUT…it handles 6 amps instead of 5, but it only dissipates 35W vs 65W, so when done, there will be a TIP120 in there. I listed all the parts used, each is COTS available, but I AM up-rating them.

what does a momentum 2.5 or 5 or twin of either look like? dunno yet. I have 2 twin mom 5s, 1 is fine, the other has a throttle that went full on and from the SCM, I would guess that Q201 shorted.

I will be providing my subs as I go on.

The point being, if you have a broke troller, and are feeling sporty, here is your road map.

If you want to know what that circuit does just ask, ps, I reserve the right to be somewhat wrong on pin to pin wiring, I was on my second rev of a sheet of paper and found a goof I THINK I fixed, but I gotta look again, I am like 75% done up-rating my model 34.

I will NOT be buying these to fix and sell, no margin in it, to do it right they consume parts and people want $25 or more, plus shipping to get them in a broken state. Id have to sell em for over $100 to feel good about it. But I will fix them now for people. Problem is you gotta get em to me. Good if you live in Pitt, bad elsewise.

Last word: when I started this project I could find nothing on the web and AI only found superstition. Most of what I did find in terms of writeups or youtube videos was horsewash. The weak link of these was too small of a heat sink with too loose of a fastening and too little venting. Fix that and these should go ferever. and would it have killed them to add a power light to ALL models???

[quaddriver]

Ok I pushed the files so far to a new github. apparently I already had one from when I GPL’d some database code years back, could not remember the PW so the new uname is THE-Quaddriver

[quaddriver]

What can I say? I added a power light. I have a crap ton of soft red and green 5mm leds from my HIGH SCHOOL days…you drive these with 1K ohm becuz they dont enjoy current AT ALL. And you note, they are ‘soft’ in color.

Bonus: I discovered a way to make the momentum adjustable by replacing 1 part. Aint done it yet cuz I either have to order said part from Mouser or find it in my junk boxes…

BLUF: (bottom line up front), having gotten the SCM for the MRC Tech 2500, these are NO WORSE, once you correct the heat problem from the Q1 mounting. In fact, you can claim they are copies, my cap mod for the model 20 Transamp1 puts it on the same footing as the base Tech 2500, the only instrinsic diff is the 14.something v Zener on the MRC to limit the base current to some max. Given the speeds we want, Id say a 1 amp 12v zener will do as well. we dont use 2/3 of the throttle on DC packs - maker be darned - if you want real speeds.

[quaddriver]

Next up to bat, the ‘Twin Momentum 5’ only cuz I own 2 and CAB2 on the ‘new’ one I got went full blast, based on what I now know, I can guess the transistor for the base current pump shorted CE.

Here iz a pic

Oh, and stop sawing off the control knobs for the TM5’s! DESPITE what the endless youtube videos say, they are NOT glued on!!! (recall what I said about all the ‘horsewash’ online) There is a clip in there, not at all unlike the one that retains the headlight knob and wiper knob on a Ford OBS F series…you use a pick to get at it and off she slides.

Im pretty sure Troller, nor no one else is making repro knobs. To remove: throttle to zero, enter from the ‘bottom’ at just BEFORE ‘stop’ and use the pick or small flat screw driver to push the clip ‘up’ as you pull off the knob.

Back to the pic, the TM5 seems to be ‘the same’ as the M1 for the most part. It has unmarked silicon, but if my map shows its the same thing then we know what to sub. The cap ‘glued’ to the case is electrically equivalent to the 470uf on the board on the M1 and my 47uf in the TA1 across the bridge…all the wires go the same place. This uses a fault light that is just an incy across the ‘fusible tube’ thingy, since all Inc lights will likely be fully outlawed in 2029, Im going to spec in an LED, also it looks like CAB1 powers both the DC outlets and the AC outlets - thats so unfair, when I move the PCB I will likely move DC over to CAB2 to share the burden. and lastly for now, they use a CT transformer (likely 24v, gotta test) and tie the CT to one leg, it makes it a 12v trans, but doubles the current - I think, I gotta white board this one.

[fwright]

Thank you so much for the schematics and explanations. You make it easy to build a decent new DC throttle. I’m working on getting my tiny HOn3 locomotives to run well, it’ll be a long time before I get around to DCC.

Fred W

[Engine_1988]

@quaddriver

Im pretty sure Troller, nor no one else is making repro knobs.

I can make 'em, if need be!

[quaddriver]

I located an interesting issue with troller momentum units and if MRC copied too much of the troller design, it might be in mrc units and when you do align the stars to hit it, its insidious.

So a gentle ask, does anyone have a momentum MRC unit? if so, can you tell me if a fast recovery diode is soldered reverse bias across the emitter-collector on the output trans? it is a something I am adding to my ‘re-design’ of trollers…

This diode acts as a flyback diode (oft used in relay drivers and across the outputs sensitive stereo amps) and gobbles the current created if the throttle is dropped hard or the pack loses power while running a train.

Please advise…