Diode failure and Peco PL-11 side mounted motor

[drmditch]

I am attempting to construct a diode-matrix to control an eight siding storage yard. Since it is/will be on the lowest level of the layout I am trying to make the controls as simple as possible.

Each siding needs between two and four points to be switched.

Not having done this before, I carefully constructed a logic diagram, and then transferred this to vero-board(?) with soldered connections.

Testing with my multi-meter before installation did not reveal any problems, (although I am now unsure that I was using the meter properly.)

 

After installation behind the control panel and connection to four of the eventual seven points it became apparent that I was encountering 'apparently random' problems with points switching when they should not.

There is a CDU in the circuit, and the point motors all work correctly when activated directly from the terminal wires connected to the matrix.

It is only when using the matrix that the problems occur.

 

After some extensive research on-line and double checking every soldered connection and every wire run I concluded that the problems had to indicate a failure in some of the diodes.

I am afraid I cannot remember the specification, as I have had them in stock for some years. They were sold as suitable for model railway use and I have no reason to distrust the supplier.

 

After testing and removal of some of the diodes in place (annoyinly tricky but I only soldered my hand once) I removed the matrix to my worktable and tested each diode again.

Out of 26 diodes, 7 have failed and appear to conduct current in both directions.

It may or may not be significant that all of the 7 failures  were those connected to a Peco PL-11 side mounted point motor.

(The 8th diode connected to that motor appears to function correctly)

 

Electronics is not my customary skill set, and I have been very annoyed with myself for not analysing this problem more quickly.

 

I would be grateful for advice on whether:-

 

Diodes degrade in storage?

Diodes can be damaged by clumsy installation or overheating with soldered connections?

Anybody else has encountered a problem with this particular point motor?

Is there likely to be any problem with the point motor that I can test for?

How often are diodes likely to fail in ordinary service, and should I re-design my installation so as to make them more easily removeable?

 

(I would supply a picture, but at the moment the whole board is a bit of a mess!)

 

 

 

 

 

 

  

[uax6]

What diodes have you used? I would say that you have overloaded their current rating, as they are fairly impossible to kill any other way!

 

Andy G

[Butler Henderson]

I have come across people thinking 1N400x series diodes are 4 amp ones , presumably because of the 4 in the number when they are in fact 1 amp. For a point motor diode matrix the 1N540x series diodes should be used which are rated at 3amps.  The final digit (x) relates to the working voltage and basically any diode in the series can be used as the lowest working voltage is 50V.

Edited June 30, 2019 by Butler Henderson

[uax6]

I agree that you want at least 3 amp diodes (*), there's a lot of current generated in the back EMFs when the coil flux breaks down..

 

Andy

(* we used to call 1N4xxx diodes thin diodes and 1N54xx diodes fat diodes. The fatter the better!)

[drmditch]

Many thanks for your help. I will order some 'fatter' diodes to 1N540x

 

I hope I can salvage the board, although I might just start again.

[DCB]

I use 1N400 series diodes in my matrixes and have never had a failure or faulty diode in over 30 years. However I don't solder them.  I just slip insulation stripped from 5 amp mains cable over the bare wires and clamp them between two Choc Block strips.  Where I don't need a diode I use a piece of wire. Simples.   I shift up to 6 big Hammant and Morgan point motors at a time with a big fat 25000 omh 25 volt capacitor without even a CDU so they couldn't have a much harder life.

My guess is you are destroying them with excess heat while soldering.   With choc block if you make a mistake you can just unscrew the diode and move it.

[Junctionmad]

Just a small point , it’s likely you used IN4001 with a 50V peak absolute rating and it’s this spec you are exceeding due to the use of a CDU and hence a much higher voltage in the back EMF 

 

4001 diodes actually have considerable surge capacity approaching 30A , so its very unlikely the current is the issue

 

you need to up the voltage  rating ie IN4002 or greater imho 

 

dave 

Edited June 29, 2019 by Junctionmad

[Tony Teague]

I have about 6 diode matrices in regular and long term use. I have never had a diode fail in operation, so once correctly set up they should run and run.

 

I agree that the smaller diodes are not likely to be up to the job - my units fire up to 5 Peco solenoids and 2 relays simultaneously.

 

 

The matrices were built for me by Giles Walburn and the only issue we had when installing them was that the unit furthest from the CDU and from the Control Panel was not firing everything reliably (~30ft away).

 

After investigation we realised that the voltage drop between the CDU / Panel and the matrix was more significant than we had expected, and so we upped all of the trigger wires plus the return to / from that unit to 24/02 cable from 16/02 - problem solved.

 

My personal view is that a well soldered joint will always be more reliable than a choc-block connector and with so many joints being within a matrix I would not use a choc-block if I could avoid it.

 

Tony

[Butler Henderson]

IME its pretty difficult to destroy a diode through soldering - if in doubt  clamp a suitable heat shunt (could be a pair of pliers)  on the lead that is being soldered between the solder point and the body of the diode. To make absolutely certain use a multimeter that has a short circuit buzzer to check the diode conducts in one direction only and fundamentally its the right way round.

[Suzie]

It is quite likely that the back EMF is killing the diodes. Just put a diode across each of the motor coils to shunt the back EMF and stop it building up high voltages greater than the 50V that your diodes are probably rated at.

 

If replacing the matrix diodes use 1N5401, and for back EMF protection diodes us 1N4148 or whatever you have to hand.

[drmditch]

On 30/06/2019 at 17:45, Suzie said:

It is quite likely that the back EMF is killing the diodes. Just put a diode across each of the motor coils to shunt the back EMF and stop it building up high voltages greater than the 50V that your diodes are probably rated at.

 

If replacing the matrix diodes use 1N5401, and for back EMF protection diodes us 1N4148 or whatever you have to hand.

Many thanks. I am ordering some of each. I'm sorry to be so ignorant, but could you describe to me 'how to put a diode across each of the motor coils'

I do want to learn new skills as I grow older, but I am also finding that there si so much to learn!

 

[Butler Henderson]

The back EMF diode goes the wrong way across the terminals in terms of polarity.

[Suzie]

Yes, the cathode (stripe end) of the 1N4148 diodes goes to the positive side of the coil, and the other end to the negative side.