Simple power amp for model train DCC

[twohills]

Hi folks

Model trains have come a long way since we were kids. They have onboard chips driven by a digitial squarewave signal on the track power supply.

The onboard units - "decoders" - for indoor model railways are US$15-25 each and deliver 12V at 1 amp to the locomotive motor. the decoders for outdoor garden railways have to deal with 24V at 5amps and cost $50-80. But it is the same signal on the rails (known as DCC).

The decoders nearly all deliver a pulse-wdth-modulated square wave to the DC motor, some at 800HZ and some at 16kHz. It seems to me it would be simple to build a crude amplifier driver stage to connect between the output of a smaller indoor decoder and the motor of a garden rail locomotive. It does not have to have good switching response, in fact some smoothing of the wave to the motor would be a good thing (the motors sometimes buzz and heat up). I am trying to find a circuit for a cheap driver: a crude 5amp switching circuit - I don't want to build something that ends up costing as much as a commercial large-scale decoder.

What would be the cheapest semiconductor device I could connect to a decoder PWM square-wave signal to switch 5 or even 10 amps 24V at 16kHz with no real concern about signal degradation? A SCR?

 

[mneary]

It won't be just one device. I assume you have a garden system with 24V. So first you will need to create 12V for the 12V DCC receiver. Then you need to interpret the 12V PWM and generate driver signals, and send them to a 24V 5A H-bridge.

Much cheaper to just get a 24V DCC unit.

 

[petrv]

I'd like to add more details:
These decoders rectify the track voltage to get power but also connect to the track to decode DCC. The motor is usually controlled with a full H-bridge (4 MOSFETs) using PWM.

You have 3 problems:

1. You will possibly exceed the maximum voltage allowed for the small decoder with the track voltage on your garden railway.
2. The rectifier in the decoder is not powerful enough (as it has to power both the decoder and the motor) exceeding its max. current rating
3. Same problem as 2. for the H-bridge controlling the motor

I agree with the previous post - to fix all those three problems would be complicated and probably not worth the effort.

 

[twohills]

Perhaps I'm not explaining adequately

Current to the motor will not flow through the decoder at all. it will flow through a thyristor (SCR) or similar device, which gives halfwave rectification, or alternatively I rectify it through a bridge first. The motor is connected in series with the thyristor directly to track power, not to the output of the decoder. All the decoder has to do is drive the thyristor gate with its PWM signal, virtually no load.

yes the voltage needs to be dropped for the decoder - probably. As far as I am aware they can all handle up to 16V, some will undoubtedly use 30V components. I would put a small zener and resistor on the track input to the decoder to drop the voltage - it will be drawing only the current to any function outputs, there is virtually no current to the motor.

Zener $1. SCR $3. bridge rectifier $5. HO decoder $15. if it works that's half the cost of a G scale decoder. Even with additional connector hardware, perfboard etc it's still not much over half. if the HO decoder is already lying around, it is very attractive

 

[mneary]

Your "PWM" must trigger the thyristor at the appropriate times, synchronous with the 50/60 Hz mains frequency. Once a thyristor is triggered, it stays on for the remainder of the cycle until the current draw ceases.

If your DCC controller is truly phase-controlled as required for thyristor triggering then it would not work on DC track supplies.

 

[twohills]

No you're right it won't. So rectify the track power to DC and switch it with a device driven by the decoder output signal then. What would be the lowest cost device for simply turning 24V DC power on and off at up to 16kHZ, slow switch times acceptable - doesn't need to be a sharp square wave output, in fact better if it isn't.

(it is square wave AC input - I'm assuming it rectifies to the same DC voltage)

 

[mneary]

If your 24V locomotive doesn't ever need to reverse, I suppose you could take the 12V from the original DCC controller and drive a MOSFET gate.

This is over simplified and doesn't deal with how you get the 12V DCC to be properly powered yet still receive the control signals.

 

[twohills]

You are right, I need something that will reverse the motor. Possibly if I use exactly what the decoder has only bigger, i.e. a 4-Mosfet H-bridge, then I can power it from the rails via a rectifier and control it from the DC PWM output of the decoder and sense with diodes when the decoder reverses the polarity. So you guys are right it starts getting marginal on cost and complexity. Sigh. Fun though.
Thanks for the input.

 

[petrv]

Yes, if you need reversal you will need full H-bridge and that means 4 MOSFETs. Now driving the high-side MOSFETs can be tricky, especially if you do not intend to use a gate-driver IC (I'd recommend using such an IC). Your small decoder can possibly contain such a driver but not necessarily suitable for your bigger and higher voltage H-bridge.

 

[tommyheadleycox]

What if.......

I'd like to add more details:
These decoders rectify the track voltage to get power but also connect to the track to decode DCC. The motor is usually controlled with a full H-bridge (4 MOSFETs) using PWM.

You have 3 problems:

1. You will possibly exceed the maximum voltage allowed for the small decoder with the track voltage on your garden railway.
2. The rectifier in the decoder is not powerful enough (as it has to power both the decoder and the motor) exceeding its max. current rating
3. Same problem as 2. for the H-bridge controlling the motor

I agree with the previous post - to fix all those three problems would be complicated and probably not worth the effort.

What if all you wanted to do was switch something on , then off. For example, a string of LED's inside your passenger cars, with their own battery power? Would it then become a much simpler matter to build something that would do this? J

I'd like to pose the same question for a sound module. Suppose you already had a decoder in your g scale engine, but you wanted to add one or two sounds, let's say a bell and a whistle. Could you build a sound-on-a-chip circuiit, (Hallmark greeting card,) power it with batteries, then trigger it on and off with the decoder?