Adding Filtering to a Power Supply

[Gristle McThornbody]

I have a 100 watt model train transformer that I need to upgrade. It is a simple linear PS with an output of about 20 vdc at a max of about 6 amps. There are some simple train movement controls based on timed decay of the output voltage to simulate braking, etc., but these controls are not an issue. It looks like the variable output voltage is controlled directly from a pot to a single 3055 clone, so there's no regulation to speak of.

Unfortunately, this PS has no filtering at all and therefore lots of AC ripple and this causes problems with some train electronics. I want to install a filter cap on the rectifier output and maybe one on the overall output as well to decrease the ripple. However, placing a suitably sized cap on the front end causes the rectifier outlet voltage to increase beyond what the fan and other control circuitry may be able to handle. Regulation is poor to non-existent so the output voltage increases also.

If I place a decently sized cap on the output, it triggers overcurrent protection and shuts down the power supply. So I can't do it from either end. Is there another way to go about reducing the ripple voltage? I don't need it super low, just under a half volt or so.

 

[Harald Kapp]

You're facing different issues.

1) Placing the filter cap right after the rectifier will charge the cap to the peak voltage which is approximately 1.4 times the average output voltage of the transformer-rectifier combination. This explains the high voltage you see.
2) Placing the capacitor at the output results in the regulator "seeing" a short circuit at turn-on. Therefore the current limiter or overcurrent protection sets in.
3) It is probably not a good idea at all to smooth the voltage. A model train's DC motor operates better especially at low RPMs when the voltage is not pure DC but has a pulsed component with higher peak amplitude than the average voltage (the voltage peaks lead to higher torque which allows the motor to overcome any braking forces even at low speeds).

Can you provide the trains and the electronics with different supplies? YOu could, for example, tap the rectifier output of the existing transformer. add a series diode to decouple it and place the smoothing capacitor (rule of thumb: 1000µF per 1A used) after the decoupling diode. The output voltage will still be too high (*1.4) but you can now add a simple linear regulator (e.g. LM317, good for 1-1.5 Amps) to supply the electronics with a well regulated voltage.

You still have the pulsed DC for the motors.


Harald

 

[Gristle McThornbody]

Thanks for the comments. I pretty much understood why these things were happening, I just wondered if there was a simple way around it.

Your thought that the large ripple voltage might act as a way to control the motors at low speed is a good one. This may have been a "poor man's" way of accomplishing that without the trouble and expense of designing in PWM. Thanks. I'll just leave it the way it is.