Power for CDU - how to stop cutting out?

[Playing trains with kids]

Help please!

I'm using a home made CDU aiming to throw 2-3 Peco point motors at once. (old ones mounted under the point directly to the point)

The circuit is:

(apologies for the quality)

 

I take the +12v/-12v from an old ATX power supply rated at 0.8A for the -12v. The 470R is 5W.

This works fine for 1 point, but for 2 points the power supply cuts out when you throw the points (it does throw the points!). Same happens with only one capacitor 2200uF

My understanding is that using the capacitors should reduce the load on the power supply as it only charges the capacitors, and the sudden flow of charge comes from the capacitors.

 

I have 2 questions:

1. Why might the supply supply be cutting out? Is there something wrong with the circuit (note there probably should be lots more diodes - I have tried a reverse diode across the output, to no avail. Trying to keep construction simple as I want to build a few of these locally for different parts of the layout). Am I right in thinking this circuit shouldn't draw more than around 600mA (read this somewhere......)? How could it be modified to prevent the cutting out?

 

2. I'm really interested in the idea of using SPDT switches with one capacitor per point - but these require a power supply capable of 3-4Amps (briefly!). What do these power supplies actually look like - and could you use a CDU arrangement to provide the necessary power bursts from a less strong supply (i.e. an ATX!)?

 

Does there need to be a small resistor between the emitter and the capacitors (30R by my calcs?) to limit the max charging current to 0.8A? If this is the case, is there any value to the transitor other than to stop the coils burning out if the switch shorts? What would this do to the charging time?

 

 I really like using the ATX as it provides 12v for the tracks, 5v for signals, accessories and servos, potentially 24v for the points if this works, and the whole lot is in one box, turned off by one switch and runs from one plug socket.

Note - I have another older ATX which is less sensitive to current spikes and quite happy with throwing three points - but of the three ATXs I've tried two have this issue.

 

Many thanks

[Playing trains with kids]

Actually, thinking about it, it can't be the capacitor charging - because the circuit works fine to throw one motor.......

[kevinlms]

Take a look at this circuit. You are missing a diode that contributes to back EMF blocking.

 

https://www.talkingelectronics.com/projects/CDU-2/CDU-2.html

 

Or else you're tripping the overload on your ATX supply, the way to fix that is to slightly increase the 470R value.

[Playing trains with kids]

Brilliant - the extra diode answers Question 1 - circuit works fine now. Actually it works without any of the diodes except the initial one from the supply to the collector. Some sites had dismissed the idea of back EMF blocking - turns out it's needed.

Thank you

 

Actually, looking around a bit more I found you could use as low as 220R (Brian Lambert's design) - I added a second 470R in parallel and it still works and gives quicker recharge.

 

 

[kevinlms]

13 hours ago, Playing trains with kids said:

Brilliant - the extra diode answers Question 1 - circuit works fine now. Actually it works without any of the diodes except the initial one from the supply to the collector. Some sites had dismissed the idea of back EMF blocking - turns out it's needed.

Thank you

 

Actually, looking around a bit more I found you could use as low as 220R (Brian Lambert's design) - I added a second 470R in parallel and it still works and gives quicker recharge.

 

 

Beware that reducing the resistor too much, MAY cause the minimum current to be in excess of the point motor ability to take a continuous flow of power. Nothing worse than cooked coils. This is assuming a switch fails and gets stuck on.