Model train controller

[Crackle]

I am thinking of embarking on a model train hobby. My son seems keen and has come up with some layout plans,
the wife has also shown interested in the idea.

I am using a 3 circuit train controller made by my brother in law back in the early 90's and I discovered recently a problem with reverse voltage by the wrong operation of points damaging the home made 3 circuit speed controller, which is based on LM317 modules. (I think they are the 1.5A LM317K types)
I think the main problem was caused by the tantalum capacitors, which were originally used when the controller was made back in the 90's. They liked the reverse voltage less than the LM317's and they blew their tops off and took one of the bridge rectifiers with them. The LM317 was OK.
Can you please comment on the suitability of using PPTCs, namely the 0.5A 32 volt type see here. https://www.littelfuse.com/~/media/elect...et.pdf.pdf .

Where is the best place in the LM317 circuit to fit the poly fuses, before the rectifiers, before the IC's or before the fwd/rev switch. Will the added cold resistance effect the motor torque much.
Can anyone suggest a simple alternative to prevent reverse voltage problems?

Thanks
Mike

[ppppenguin]

The usual way to protect a regualtor against the output going higher than the input is a diode between output and input. In the direction that will conduct if the output goes high. The reservoir cap will then absorb the impulse. The bridge shouldn't mind.

BTW, you marked the reservoir caps 100n. While it's a good idea to put a 100n cap close to the regualtor input to stop it oscillating you'l want at least 1000u reservoir too.

[Crackle]

Hi Jeffrey
I think the fault that blew the tants and rectifier was connection to another of the regulator circuits, either in phase or reverse connection, I am not sure which.
The circuit was assembled by someone else not me, and I agree a larger reservoir/smoothing cap would be beneficial.

So a 5 amp diode in circuit with each rectifier should protect the regulator, in the event of a high positive voltage being connected to the output.
Something like this;
   
But what about the case of a short circuit over current, how can the rectifier be protected.

What are efuses, they are discussed on the web but there does not seem to be anywhere that supplies them.
What one needs is a small self contained efuse IC which can trip accurately in the event of a fault and then be reset by removing the supply. Is there such a thing?

MIke

[Refugee]

All train sets including full size ones have an electromechanical circuit breaker of some kind.
You can actually see several sets of three phase circuit breakers on google earth in a compound between the shuttle terminal and the tunnel entrance at Folkston. They are just big ones like the little one on a 1960s model transformer.

[Alistair D]

Some interesting reading here;

https://www.instructables.com/id/Simple-...M-inertia/

One version of the controller includes a diode and some other components to help prevent reverse breakdown.

The article also covers pulse width speed control and the advantages of using it.

Al

[ppppenguin]

So a 5 amp diode in circuit with each rectifier should protect the regulator, in the event of a high positive voltage being connected to the output.
Something like this;

But what about the case of a short circuit over current, how can the rectifier be protected.

The rectifier should be able to look after itself. Just uprate it a bit if necessary. An inverse diode across the output of each regulator will protect from reverse voltages.

Are you trying to protect against transients or a sustained fault? The diodes should protect against transients without any trouble at all. A sustained fault needs a fuse for protectin. Either conventional or polyswitch.

[Refugee]

I would go for something that can be reset so that you are not forever replacing fuses after derailments.
When I last thought of model railways I wondered if anyone had thought of fitting a speedometer.
Across the current limiting resistor sense with AC coupling the motor ripple and using data from the full size loco work out how many ripples the motor produces to travel its own length and convert it to speed.
From that using a constant length pulse circuit to create a drive signal for an on panel speedometer.
No need to modify the loco.
You could even add a speed limiter for models of modern locos that have them fitted.

[Crackle]

Hi Al
Some interesting reading and Ideas in that article, thanks.

Jeffrey
As I have said I think, but not sure, the fault which caused the problem was a reverse current from one of the controller circuits going to another controller. This would in effect put a short to ground on both of the controllers involved. Because it was originally built by my Brother in law with tant caps, I suspect the rectifier may have been taken out by one or more going short.
The symptoms were a "pop" and eventually stinky smoke, probably from the rectifier encapsulation.
The reverse diode protection is a good idea I could easily try.

But going back to my original question about poly fuses, what would be the best location in the circuit to fit one.

thanks

Mike

[ppppenguin]

I reckon Polyfuse in series with the output.

The regs have inherent overcurrent protection. The 2 diodes will protect against reverse and overvoltage applied to the output. The bridges really ought to be rated sufficiently not to blow up if abused slightly. Perhaps 200V 4A or so. Very cheap, readily available and way beyond anything they're going to encounter in real life.

I have this dim memory that model trains were happier with a rather rough DC than with the nice smooth output from a regulator. For example this H&M controller had a switch to give you halfwave rectification. https://www.yesterdaystoys.co.uk/Hammant...ATJVW.aspx
Or is this an obsolete idea. Or an urban myth.

[Terry]

I have this dim memory that model trains were happier with a rather rough DC than with the nice smooth output from a regulator.  For example this H&M controller had a switch to give you halfwave rectification.
Or is this an obsolete idea. Or an urban myth.

The rough DC almost certainly was because that was what controllers had always supplied going back before solid state days - my first one used a copper oxide (might have been selenium) bridge rectifier. They were cheap and cheerful, so nobody bothered to make anything better.

The half wave rectification idea is not an urban myth - it gives better slow speed control for heavy freight trains, etc. However, the motors seem to chuck out a lot of noise - a more realistic way, I think,  would be to smooth the DC and use PWM at a higher frequency.

The auto shutdown shown in step 3 of Alistair's link gives you the protection wanted without the need for fuses.