MRC Ampack upgrade: Attn electronics froods

Folks:

Actually, this is a rather basic question, and a scarcely hoopy frood could probably answer it. Unfortunately, I burned a hole in my towel with a soldering iron and had to throw it out.

I have an old MRC Ampack that works. It has a decent power output, which is nice, and a good-quality rheostat, but poor slow-speed control on many of my locomotives. It was a great $2 deal, but I want to make it better, and here's what I want to do:

formatting link
F1 is some sort of circuit breaker in a glass envelope.

Basically, I want to make it into a very simple voltage-follower transistor throttle, and add a pulse-power switch to cut out one diode. If my calculations are correct, the circuit should work. The TIP-120 is supposed to have a gain of at least 1000, and R3 is there to limit the base current to no greater than 1.2 mA, which seems adequate for controlling 1 A with that gain.

The circuit is pretty similar to what my MRC Railpower 1300 contains, as far as I can tell, and that gives good slow-speed control, even without the option of pulse DC for low speeds. I'm not really worried about heating effects. I don't go in for coreless motors.

I am going to mount Q1 on an aluminum heat sink. I could calculate the size needed, but I'm not going to; I'll just go with "as big as fits" and replace the $1.59 component if the smoke comes out.

Any comments on this circuit before I breadboard it up and try to fry things? It seems simple, but I'm new at this.

Cordially yours: Gerard P. President, a box of track and a gappy table.

Reply to
pawlowsk002
Loading thread data ...

You might want to post this over on sci.electronics repair as well.

snipped-for-privacy@gann> Folks:

Reply to
jJim McLaughlin

J McL: Good idea.

Crossposted upon suggestion...note that this device is a model railroad controller.

Transformer T1 delivers 1A.

snipped-for-privacy@gann> Folks:

Reply to
pawlowsk002

On 10/5/2007 9:43 AM snipped-for-privacy@gannon.edu spake thus:

Plenty good enough. Suggestion: use a simple touch test. Operate the pack with the cover open and feel the heat sink after some decent interval of use. If it isn't too hot to touch, you're OK.

By the bye, the TIP120 is rated at 5A continuous, 8A peak collector current (see

formatting link
for NTE's equivalent part).

Reply to
David Nebenzahl

I'd put a ceramic capacitor across the voltage supply as close to the transistor as possible and a reverse oriented diode across the output (before the reversing switch. Also, if your output voltage doesn't go fully to zero then a diode (or two) in series with the output.

Add a few more components and you can have a momentum controller.

Greg.P. NZ

Reply to
Greg Procter

GP1:

I assume the capacitor would be a filter, or am I wrong about that?

I'm a little uncertain about filtering, because I don't want to smooth out my output current too much. I know this can cause more heating, but that's not something my motors and I have much trouble with. Our problems are more along the lines of "Oh s**te, the Tyco Plymouth at the end of that smoke trail wasn't powered down, it was stalled."

Or is the cap for RFI purposes?

What would the diode be for? I'm trying to use this as a learning experience, as well as a practical controller.

The diode drop is a good plan. I may have to use that with my KF pack, whose selenium cells are getting replaced with silicon diodes, because I don't intend to run BLI K4s . It depends what voltage drop was allowed for; I'll have to stick a meter on it.

The momentum idea is interesting. I have some old MR articles which cover simple transistor throttle circuits (good old LHW and his ability to put basic info, four practical circuits, AND a plug for his True Action Throttle in three pages of dense text and small drawings) and I may use it eventually once I understand the basic principles involved better. I find that a little experimentation helps me do that. Study a little, think a little, try a little, repeat as necessary.

Cordially yours: Gerard P. President, a box of track and a gappy table.

Reply to
pawlowsk002

It's there to cut out high/radio frequency signals, transistors can take exception to those.

That's it.

The diode is there to short out impulses returning from the layout. Transistors can go pop in response to such impulses and a model railway can be an electrically noisy place.

A meter! Good idea! My first transistor controllers were built when I only had old coffee grinder motors that laughed at fractional voltages. My first decently motored loco just ran on and on and on ...

When you're ready :-) After the potentiometer output, add a second resistor and put a (say

100uf) cap from between the resistors to the ground. The loco will now gain and lose speed at a slower than instantanious rate. Add a diode between pot and resistor and the loco will accelerate as before but take forever to brake. Add another pot (with added resistor) between cap and ground and you have a brake lever. Put a switch between the pot and the last resistor and you can switch the momentum on/off. Swap the first resistor for another pot and you have variable momentum rate. Separate the traction current to the transistor from the control to the transistor base and you're back to full wave/half wave. Buy another darlington transistor and parallel it with the existing one, each fed through one diode and you have variable pulse from half to full wave at any given setting. Add momentum to the variable pulse :-)

Hey, you've got a whole new hobby in a plastic box and soldering iron!

Greg.P. (tongue in cheek)

Reply to
Greg Procter

snipped-for-privacy@gannon.edu wrote:

That circuit will work, and the emitter follower TIP120 will maintain a constant voltage at the emitter, no matter how much juice the motor draws. The theory of emitter followers is simple. Given a constant base voltage, the emitter voltage will be 0.7 volts (one diode drop) less than that base. The transistor will flow whatever current is necessary down from the collector to keep the emitter voltage one diode drop below the base voltage. Given a constant voltage out of the emitter, the motor will maintain a constant speed even if the mechanical load varies from sticky side rods, wheels out of round, grades, whatever. It's a good improvement and will make loco's run smoother than a plain rheostat will. If you have 1 amp going thru the transistor, base current might be as bad as 2% of that, I.e. a gain of 50. Large signal gain of power transistors is lower than the small signal gain. To stabilize the circuit you want current going thru the pot to be large compared to the base current, so that the base current draw doesn't alter the voltage at the pot wiper. I'd make the pot more like 500 ohms than 5K, that's only a 1/4 watt of power in the pot, a 1/2 watt pot will run cool enough and the standing current in the pot will be 24 mA, large compared to the estimated 2 mA base current. Then I'd make the base resistor about 100 ohms. The only purpose of the base resistor is to surpress oscillation in the emitter follower, and 100 ohms is enough to do that job. With only 100 ohms, the 2 mA base current will only drop 0.2 volts across the base resistor. With a 5K base resistor, 2 mA base current drops 10 volts, nearly your entire supply voltage, across the base resistor. Build this circuit and get it running before getting into momentum throttles. A momentum throttle needs a lot more gain then you get with a single transistor. The fundamental theory of a momentum throttle is a large capacitor which changes of throttle setting charge or discharge slowly. For this to work, the power amplifier cannot draw much current from the capacitor, which means a high gain amplifier, either several transistor stages or an opamp. It's a more complex circuit which has a lot more ways of going wrong. I'd recommend the fairly straight forward circuit you have for a first try. Do the fancier stuff after you get something simple working.

David Starr

Reply to
David Starr

Reply to
pawlowsk002

GP1:

The diode idea seems good. I presume it can be fairly small, since it doesn't handle full output current? I will also consider the RFI filter. I suppose the disc cap should be small, so as to shunt high frequencies without smoothing off my pulses too much...this is going to take some figuring out.

DS:

I may have drawn my sketch in a somewhat misleading way, because Q1 isn't a single transistor, but a Darlingon arrangement, so I probably should have used some sort of IC symbol. I looked up the specs to get the gain of 1000, which is at 3v collector-emitter and 3A current, but I'm going to test the circuit first. I think I'll hook up some wirewound resistors across the output and check voltage under load, and then perhaps try it with a motor, and if the voltage seems to drop too much I'll reduce the resistor values and try again.

One thing I'm confused about - isn't the base resistor also there to limit base current? 100 ohms would do that, of course. I used the large resistor because

The circuit worked on the breadboard, with no load. This was a very pleasant surprise.

All the advice is very much welcome. At this point I know enough to get into trouble. I'm trying to learn enough to get out before proceeding farther in -- already, I'm doing better than last time I tried something like this, which was an attempt to build Peter Thorne's similar circuit in an old MR. It didn't work, but I couldn't figure out why, so I'm trying to avoid blind copying this time.

Cordially yours: Gerard P. President, a box of track and a gappy table.

Reply to
pawlowsk002

On 10/6/2007 8:43 AM snipped-for-privacy@gannon.edu spake thus:

Yep; consider the possible sources of reverse-polarity voltage coming

*from* the layout (noise, reverse EMF from loco motors, etc.); there's not going to be much current coming from them.
0.1 uf. Don't overthink this; it will have *zero* effect on performance from your point of view.
Reply to
David Nebenzahl

A signal diode will do, or an IN400x which is the cheapest I know of.

Just solder a little one in from the junk box. :-)

I knew what you meant. the normal Darlington represertation has two interconnected internal transistor bits inside the circle.

The base resistor is there to limit base current. Advance the pot to full voltage without the base resistor in circuit and a short circuit load and there's full voltage (= infinite amps) between base and emitter, not a good look! Ditto if you put the momentum capacitor in there without a base resistor.

Life's like that, it sets you up to expect success and then it runs you over with a steam roller! ;-)

The components aren't expensive, so as long as you don't set the house on fire or rub your nose with the soldering iron it's all gain.

Which one was that? I built most of them way back and learned enough to figure out when it was my fault. ;-)

Greg.P.

Reply to
Greg Procter

snipped-for-privacy@gannon.edu wrote:

Testing with a resistor load is a good thing to do. A 12 ohm load resistor will draw one amp at 12 volts. It will also get hot, dissapating 12 watts at 12 volts. Motors, so long as they are turning, don't draw much current unless they are driving a real mechanical load. For the home workshop it's hard to arrange a suitable mechanical load, but easy to find a resistor that will take enough juice to stress the throttle-under-test.

Not really. The pot should be holding the base at a constant voltage. The base emitter junction will be forward biased, which causes conduction from collector to emitter. Enough current will flow thru the emitter to bring the emitter up to a diode drop or two from the base. That's why it's called an emitter follower, the emitter follows the base voltage. If you have significant resistance in the base circuit, then the voltage drop across the base resistor lowers the base voltage, which lowers the emitter voltage. This isn't the end of the world, but the purpose of the circuit is to keep the emitter voltage steady against fluxuations in emitter (load) current, so the locomotive motor can draw all the current it wants to keep turning at a steady speed. The motor sees variations in mechanical load from bumps in the track, trains wheels with sticky spots, or stickiness from the siderods, worm gear, valve gear etc. Nothing is perfect, and the load the motor sees changes as the shaft goes round. If the motor can draw more current when the extra load tries to slow the shaft, the motor will run more smoothly. If the extra current draw lowers the voltage going to the motor it will slow down. In short, you don't want to limit base current in an emitter follower. A darlington pair acts just like a single transistor with the benefit of much increased gain. The gain of a darlington pair is egual to the product of the two individual transistor gains. If the bigger power device has a gain of 10 and the smaller piggy backed darlington transistor has a gain of 100, the combined darlington pair has a gain of

1000. The downside of the Darlington is doubled base emitter voltage, but that doesn't matter much in this circuit.

David Starr

>
Reply to
David Starr

Ceramic capacitors have a low enough capaitance that it would have no noticeable affect on the sinusoidal pulses, but it would greatly reduce (filter) any high frequency components. I assume something in the 0.01 uf to 0.1 uf range, voltage rating greater than the output measured in normal mode X 1.414. Actually, I can't imaging any high frequency stuff coming back from the layout, but it couldn't hurt, possibly to surpress any oscillations from the high gain darlington. On the other hand, I wouldn't make a special trip to buy one. It probably isn't going to be needed.

You could have used just a regular power transistor, with lower value R2 and R3. It would have reduced the heat in the transistor from about 1 watt per amp to about 0.3 watts per amp.

I have a power pack with pulse mode but I never looked at the output on a scope. I thought they used some kind of pulse width modulation, where the pulse amplitude was always at max voltage, and they went from sliver thin at the almost zero setting and got wider and wider as the knob was turned up, until they disappeared at the 100% full voltage point.

With your proposed circuit arrangement, you can run it in the normal full wave rectifier mode with 120 sinusoidal pulses per second, or the half wave rectifier 60 sinusoidal pulses per second mode, which I expect will result in a slower top end loco velocity, but if the loco is running and you switch from 60 PPS mode to 120 PPS normal mode, I predict there will be a noticeable sudden increase in speed, if that's objectionable to you.

Also, since the transistor is electrically "hot", you'll need to use an insulator under it, or somehow insulate the heat sink from ground. And don't forget the silicone thermal grease under the darlington.

Let us know how it works!

Reply to
dgw

Gerald, with the proposed circuit, the output current will start to flow when the poteniometer setting is about 1.4 volt about ground (zero volts), because the two emitter junctions in series in the darlington each have a

0.7 volt drop to be overcome before they conduct any current. You can overcome this "dead" range near zero on the pot by adding two silicon diodes (any small kind) in series, pointing down, between the bottom of the pot and ground.
Reply to
dgw

David: while you have some good points in your analysis, I disagree on a couple of things.

The output voltage will still drop with increased load due to a dropping collector voltage, due to voltage sag caused by the transformer impedance. If a fairly rock solid stable output voltage is desired then the voltage at the top of the pot has to be regulated. The simplest way is with a zener diode and resistor. The drawback it that the top end output voltage is then a lot lower. Plus, even if the voltage were rock steady, the loco that is binding or going up hill is still going to slow down because of increased load on the motor.

My other comment is that using a 100 ohm base resistor will blow out the darlington when the pot was cranked all the way up. I don't know the specs on the transformer, but a 5 K base resistor may save the darlington in the event of a short out on the tracks.

Anyway, the purpose of the mod is to make the engine crawl better as a result of the quasi pulse mode.

Reply to
dgw

Gerald and Greg - Greg, you talked about adding a diode in series with the output IF the voltage doesn't go all the way to zero when the pot is turned all the way down. I agree with you.....the diode would fix or reduce a problem like that.

However, the voltage will go all the way down when the pot is turned all the way down. In fact, the voltage will go to zero when the wiper of the pot gets down to about 1.4 volts, so, there is no need for a diode in series with the output. If you use one to stop some phantom reverse voltage coming from the track, the diode must be speced to handle the full current plus some. a 1N400X is not beefy enough. Also, if you add a diode on the output, the output voltage will now drop to zero when the pot wiper gets down to 2.1 volts. So if you want the output voltage to get to zero when the wiper voltage gets to the end of its travel, you will need to add three diodes in series between the bottom of the pot and ground, instead of the two I suggested in a precious post.

Reply to
dgw

It's a good theory, and is almost certainly true with straight emitter/follower circuits but can go wrong once you start adding momentum etc.

Sure, but I've been doing it for ... umm ... forty years and never had that diode fail.

Fair enough. I've often played with series and parallel resistors and diodes to give the pot non-lineal characteristics. Giving the 0-6 volt range to about 60-65% of the pot arc and 6-12 volts

40-35% arc works for me.

Reply to
Greg Procter

dgw wrote:

Quite true. No real world power supply can furnish infinite amounts of current. In fact the original rheostat only power pack relied upon transformer winding resistance or a circuit breaker to limit current into a short on the layout.

Protection is a problem, every railroad has derailments which can short the rails together. Plus a world of other faults. However, I would not rely upon a base resistor to limit transistor current. Gain of transistors is not well controlled. Typically it varies over a range of

3:1 for any single type. The Fairchild data sheet for the TIP120 gives a gain of 1000, that means all devices will have a gain of at least 1000, and plenty of them will have a gain of 3000. This means a resistor that protects a low gain transistor won't protect a high gain device. The resistor that protects a high gain device will lower the performance of a low gain device. One of the rules of transistor design is don't design circuits that depend upon the magnitude of transistor gain. If you do, you find that some of them work and some of them don't, depending upon the luck of the draw in transistor gain. The first line of defense is the biggest possible heatsink. Let's guess a good heatsink could hold the transistor case temp down to 75 C, allowing the device to survive 36 watts. Assume part throttle where we drop 6 volts across the TIP120 and supply 6 volts to the track. At this setting, the transistor can handle 6 amps before going poof. I doubt the transformer/rectifier can furnish that much current. In short, we protect against shorts by relying upon a stout well cooled transistor being tough enough to withstand them. The best heatsink is the case of the power pack, assuming the case is metal. For a plastic case, the best heat sink might be a piece of aluminum the size of the power pack muonted underneath it. Probably the case has four rubber feet secured by screws to the bottom of the case. Pull those screws, drill the heat sink to match the case screw holes, get some 1 inch screws, and some 1/2" standoffs. Mount the heatsink on the standoffs, the feet on the bottom of the heat sink, and the transisor on the top of the heatsink. This gives air circulation on both sides of the heatsink and protects the transistor from mechanical damage. Incidently, your idea of putting two diodes in series at the bottom of the pot is a good one. Without them, track power goes off when the pot wiper is at 1.4 volts, about 10% of pot rotation. The effect is a dead spot at the slow end of pot setting, nothing happens for rotating the pot the first 10%. The two diodes are an elegant solution yielding nicer control action.

David Starr

Reply to
David Starr

dgw:

My sources say that Fyffe packs put out these kind of variable-width pulses. I have no idea what happened to Fyffe, but I suspect it had something to do with suspected heat damage to motors. Whether such damage commonly occurred is a good question, but the model RR control universe has always revolved around a massive black hole of hype -- it's easy to dig up rumors, but real information is harder to come by, and that probably goes equally for DC and DCC.

I know what you mean about the half-wave pulses - the RMS current would indeed be halved if I cut out one diode. I thought about wiring a DPDT switch to use the whole 24V transformer with a half-wave rectifier, but usually when I've seen pulse-power switches like this they are wired the way I have it. It would only be for starting and very low speeds. I'll see what happens.

If your pack is an MRC Throttlepack or Tech II 1400 or 2400, it's probably equipped with the same kind of half-wave pulse power as shown here, though I can't verify that since I've never had one apart. I actually have some reason to suspect that the Tech IIs mentioned above used something much like my circuit, or Peter Thorne's which mine is probably ripped off from -- but then, how many ways are there to make a voltage follower?

This was Thorne's simple 6-component transistor throttle, Greg P., and the one I tried to build before but couldn't. In retrospect, I probably mixed up the transistor leads. Oh, and of /course/ I own a meter, which I use for electrical troubleshooting on appliances and suchlike - I'm an electronics newbie, but I'm not totally brain dead. :)

Anyway, this thing is going together, bit by bit, in odd moments, and I should be able to report on its operation in a few days. I'm still using the 5k pot and 10k resistor, because as it turned out the Rat Shack didn't have a pot that was any smaller value. Yes, I do have a Jameco catalog, and my brother has one from Digi-Key, but let's go with what we can buy in town for now.

Cordially yours: Gerard P. President, a box of track and a gappy table.

Reply to
pawlowsk002

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.