Diode Matrix - Frustrating Problem

[drmditch]

This is an appeal for help.

 

I have been struggling with an electrical/electronic problem for two months now, and within the limitations of my skills I seem incapable of finding a resolution.

 

I would be very grateful if some of the nice, knowledgeable, and highly skilled members of RMWeb could help me.

 

My railway is being built/re-built on three levels. The lower level storage sidings have been constructed and are switched using a diode-matrix which although not the most elegant of devices appears to work well. (I had a problem with the specification of diodes which several people on here which people on RMWeb were able to advise me about.)

 

It is the simple four road mid-level sidings that I am having the problem with.

 

Here is the logic diagram for the diode-matrix.

 

Post_01

 

As designed it should allow one contact to be made for each of the four roads.

(For the purposes of this post I have left of the switching for turnouts G, H and J since they are not currently a problem.)

 

Turnouts A,B,C E, and F all operate as designed.

 

The problem I have is Turnout D.

 

It does not operate at all when ‘fired’ through the ‘input’ side of the matrix.

 

The system is intended to work on a probe/stud contact basis from the control panel (not shown because that does not seem to be a [resent problem. It uses AC power from a Gaugemaster CDU providing 1.6v AC. I have used this system for many years without significant problems.

 

A,B and C are operated by (quite elderly) Peco PL-10 (and other variants) solenoid point motors surface mounted as seen here:

 

Post_02

 

 

D, E and F are operated by the ‘new’ Peco PL-11 Side Mounted motor

 

Post_03

 

 

The matrix was constructed as seen below, using IN5401 1844 diodes (which are nice and ‘fat’).

Please excuse my primitive soldering and wiring skills. In mitigation I would say that my railways have usually worked and if problems occur I can trace and correct faults. The wiring will be tidied up when it all works!

 

Post_04

 

 

 

Post_05

 

 

 

Traction DC Power wiring is all red and black, Turnout Control AC wiring is all white (return) and yellow and green.)

 

 

I can test the matrix function using a probe at various points.

 

As stated above, it is only D that does not operate as designed with power from the input side (LH terminal block with blue wires). All the others work correctly.

 

So:-

 

It does work when :-

1.    fired directly from the output side (RH terminal block)

2.    fired from the output side of the diodes.

3.    When (as a test) I cut the input side of the diodes (to remove the connection to the input block) it also fired from the input side. (you can see the soldered repair in  Post_04

 

It does not work when:-

4.    One of the other motors is disconnected (in case six motors at once were too ‘power hungry’.)

5.    Fired directly from the input side of the diodes connected as normal.

I have checked for all the obvious conditions I think of. My only conclusion is that somehow the motor is receiving current from both directions at once. I have checked for any obvious short-circuits on the board especially in positions marked X and Y.

 

Any shorts on the ‘input’ side of the matrix would appear to be contra-indicated by all the other five motors firing correctly. This is true whatever sequence is chosen (ie 1 to 2, 1 to 3, 2 to 4 etc)

 

This mid-level problem needs to be overcome so I can lower down the top level baseboards (seen propped up in Post_02) and get on with the functional railway.

(The top level is/will be all mechanical turnout control – I know where I am with that!.)

 

 

I would be grateful for help.

 

 

Caroline

[steve W]

1 hour ago, drmditch said:

 

 

 

 

Post_05

 

 

As stated above, it is only D that does not operate as designed with power from the input side (LH terminal block with blue wires). All the others work correctly.

So:-

It does work when :-

1.    fired directly from the output side (RH terminal block)

2.    fired from the output side of the diodes.

3.   (you can see the soldered repair in  Post_04

 

It does not work when:-

4.    One of the other motors is disconnected (in case six motors at once were too ‘power hungry’.)

5.    Fired directly from the input side of the diodes connected as normal.

I would be grateful for help.

 

 

Caroline

 

Ok, not a fix but possibly narrow it down a wee bit:

 

  5.  When (as a test) I cut the input side of the diodes (to remove the connection to the input block) it also fired from the input side.

Taking this as the case and assuming all other outputs are connected, it would seem probable that the problem is in the area between the input wire terminal block and the diode input, assuming the circuit from the stud probe contact to terminal block is OK. 

-Check both wires on the screw terminal block that the insulation is not trapped under the screw and therefore possibly not making a connection.

- Check track 3 of the vereo board where the short blue wire and diodes connect.  In the picture of the underside of the board there is a connection on this track, second hole up from the bottom of the group, that looks as though it might be shorting to the 2nd track (might just be the angle of the picture of course). Reflow the solder joints just to make sure all is well.

Are you using lead free solder? This can be notoriously difficult to get a good joint.

I very seldom use these point motors but would have thought that six solenoids working together would require a very large capacitor to work, but it should go with all the others temporarliy disconnected.

As an aside, I note you have cut the copper tracks where needed with a very thin cut. It may not be your problem here, but tiny bits of copper are easily left behind and can form almost invisible short circuits. Better to use a drill bit ~4mm dia hand held or in a cork or wood handle and twist it in the hole for a better and more visible gap.

 

Best of luck

 

Steve W.

 

 

 

[Hroth]

As mentioned above, the soldered joint might be "dry", especially if using lead-free solder.  You could try unsoldering it, cleaning up the diode wire and stripboard, then resoldering.

 

This might be a possibility you have tried already, but another possibility is that one of the diodes is duff.  Disconnect one end of the diodes you have indicated, and test them with your multimeter on the resistance setting.  One way should conduct, the other should not. if either are non-conducting, then try replacing them.

 

 

[drmditch]

1 hour ago, steve W said:

 

Ok, not a fix but possibly narrow it down a wee bit:

 

  5.  When (as a test) I cut the input side of the diodes (to remove the connection to the input block) it also fired from the input side.

Taking this as the case and assuming all other outputs are connected, it would seem probable that the problem is in the area between the input wire terminal block and the diode input, assuming the circuit from the stud probe contact to terminal block is OK. 

-Check both wires on the screw terminal block that the insulation is not trapped under the screw and therefore possibly not making a connection.

- Check track 3 of the vereo board where the short blue wire and diodes connect.  In the picture of the underside of the board there is a connection on this track, second hole up from the bottom of the group, that looks as though it might be shorting to the 2nd track (might just be the angle of the picture of course). Reflow the solder joints just to make sure all is well.

Are you using lead free solder? This can be notoriously difficult to get a good joint.

I very seldom use these point motors but would have thought that six solenoids working together would require a very large capacitor to work, but it should go with all the others temporarliy disconnected.

As an aside, I note you have cut the copper tracks where needed with a very thin cut. It may not be your problem here, but tiny bits of copper are easily left behind and can form almost invisible short circuits. Better to use a drill bit ~4mm dia hand held or in a cork or wood handle and twist it in the hole for a better and more visible gap.

 

Best of luck

 

Steve W.

 

 

 

 

Thank you. I have tidied up the part you mention, and unfortunately it had no effect. I am going to make sure that are no other trace connections though, by using a wider drill bit to ensure tracks are appropriately isolated.

[drmditch]

15 minutes ago, Hroth said:

As mentioned above, the soldered joint might be "dry", especially if using lead-free solder.  You could try unsoldering it, cleaning up the diode wire and stripboard, then resoldering.

 

This might be a possibility you have tried already, but another possibility is that one of the diodes is duff.  Disconnect one end of the diodes you have indicated, and test them with your multimeter on the resistance setting.  One way should conduct, the other should not. if either are non-conducting, then try replacing them.

 

 

Thank you. I have tested the two diodes involved (and all the others!) and they appear to function normally. 

I also (using the terminal blocks) removed both connections and bypassed them by using two new diodes. This provided a direct link between the LH terminal block 3 and 4 and the block for DL and DR. The same problem seemed to exist. This suggests to me that there is a short rather than a loose connection - but I will try re-making the joints anyway!

 

(The metal and coating on the Chinese manufactured diodes needs to be cleaned thoroughly to get a good solder joint!)

Edited March 20, 2020 by drmditch

[Donw]

When wiring a friends layout up I found the gaugmaster CDU adequate to throw up to six peco point motors. Any that were correctly set to  start with will draw less current as once the solenoid has operated the core will then be inside the coil and the impeadance much greater. If needed a second dab will have the power more concentrated on any that are sticking.

 

That said if D is not working onboth coils i.e. DL and DR it suggests to me the problem would more likely then be in the return. If there is a bit more resistance in the return than the others the D coils would get a lesser share of the power through the matrix but if tested by applying current direct the CDU would be supplying that one alone and the extra resistance in the return would not matter.

I would check that first.

 

Don

[Junctionmad]

If D is working when triggered from the output side of the diodes as you state . Then debug 

 

1. test the input side of the diode ( note DL doesn’t need a diode anyway as it’s operation is not shared with another route ) 

2. based on 1 , work back into the input wiring , if 1 passed , or forward into the point motor wiring of 1 failed 

[petrovich]

Good morning all

 

Sorry to interrupt this conversation but which model of diode are you using as this is showing all the symptoms of a low level of power.

 

I can explain more if required.

 

Regards Peter

[petrovich]

Apologies did not pick up on the earlier description mentioning you are using IN5401 1844 diodes.  Can I ask are you switching Point D in a sequence and if so how many Points are operated before Point D is operated? 

What happens if the other points are disconnected and you operate Point D on its own?

 

None of the diodes you would suspect have failed they are simply a one way current blocking device therefore you would assume there has to be something with Point D that the CDU cannot cope with. 

Out of interest have you disconnected everything including the CDU, removed all power sources and tested the diodes on the board to ensure there is no leakage across them your multimeter will do this with either it set to diode or resistance?

 

As above I apologise if I am repeating things you have tried or give the expression of teaching to suck eggs.

 

Regards Peter

[drmditch]

Thank you to all who have replied.

Over the weekend I have made progress.

I did take all the measures suggested to eliminate shorts.

Situation as of 'before first coffee' this morning:

.    With everything connected the problem remains for the 'D' turnout only.

.    However, if a PL10 motor is directly connected to the output block of the matrix, everything 'fires' when it should.

     The same is true using a PL 11 motor.

.    since the two ends of the track loops are some 7' apart (as the cable runs go) there is another terminal block close to the D,E, and F motors.       

     Removing all the connections to D, and substituting the same PL10 test motor as above, then firing from the input side of the matrix works  

     all the motors as designed.

 

I should point out that I test all motors before installation; strip and solder all wire ends before inserting them into terminal blocks ; where wire joins are needed (very few) they are soldered, hooked together, and re-soldered.

But somehow, somewhere...….

 

Time for 'second coffee' !

Edited March 23, 2020 by drmditch

[steve W]

I'm rushing around at the moment, so only a quick read. Is there a very long wire to this point motor eg more than one/two foot from the CDU?    If this is a thin wire , say 7.0,2 or even 16/,o2 the resistance in the wire may be limiting the ability of the point motor to work. Try running an extra one or two wires on both the coil and return wires and retry.

 

Cheers

Steve W

 

[kevinlms]

1 hour ago, steve W said:

I'm rushing around at the moment, so only a quick read. Is there a very long wire to this point motor eg more than one/two foot from the CDU?    If this is a thin wire , say 7.0,2 or even 16/,o2 the resistance in the wire may be limiting the ability of the point motor to work. Try running an extra one or two wires on both the coil and return wires and retry.

 

Cheers

Steve W

 

Speaking of wiring. OP what sort of wire are you using? If it's single stranded (bell wire), is it broken inside the insulation. A common cause and is the reason why such wire should never be used, on a model railway as simply speaking it can break when moved.

[Donw]

From what you say it sounds very much as though there is a problem with motor D assuming the connections at the terminal block were good I would suspect a dodgy connection at the point motor giving a slightly higher resistance so it will only work with the full wack of the CDU but not when shared with others.  If the wires to D were much longer than to the others or finer wire that might have a similar effect.

Would it not be simpler to replace PM D with another and see if that does the job?

 

Don

 

 

[drmditch]

Thank you both.

I never use single strand wire.

I do use multi-strand 'layout wire' (7x0.2mm) , with thicker (core thickness 1mm - approx!) cable for some elements.

I have identified the problem as being mechanical force needed to move the spring on the turnout. 

The turnout when removed from the turnout functions correctly.

Th resistance on the original motor is the same as a new straight from the box motor. 

 

Since it does work the point when 'fired' from either the 'output side' of the matrix, or the 'local terminal block', (the difference being some seven feet of wire, I shall try 'steve W's suggestion of doubling the wire on the 'coil' links. The return link is already using the thicker wire.

 

 

[St Enodoc]

17 hours ago, drmditch said:

Thank you both.

I never use single strand wire.

I do use multi-strand 'layout wire' (7x0.2mm) , with thicker (core thickness 1mm - approx!) cable for some elements.

I have identified the problem as being mechanical force needed to move the spring on the turnout. 

The turnout when removed from the turnout functions correctly.

Th resistance on the original motor is the same as a new straight from the box motor. 

 

Since it does work the point when 'fired' from either the 'output side' of the matrix, or the 'local terminal block', (the difference being some seven feet of wire, I shall try 'steve W's suggestion of doubling the wire on the 'coil' links. The return link is already using the thicker wire.

 

 

I think that's a good idea. With five motors to be fired simultaneously for routes 5 and 6 you might be getting to the limit for the size of capacitor that you have (if it's the standard Gaugemaster unit it should be 4400 microfarad) and the highest resistance path might not be getting enough current to throw the motor reliably. If it's still a problem when you've doubled the wire, I'd look at a bigger capacitor. I read once that a good rule of thumb is 1500 microfarad per motor firing simultaneously but you might not need to go that far. Another alternative would be to increase the voltage, possibly by powering the CDU from a computer power supply at 20V or so instead of 16V.

 

Good luck.

[RobjUK]

On 23/03/2020 at 09:30, drmditch said:

I should point out that I test all motors before installation; strip and solder all wire ends before inserting them into terminal blocks ; where wire joins are needed (very few) they are soldered, hooked together, and re-soldered.

 

Probably not directly relevant to your existing problem, but to prevent future ones:

 

It's not a good idea to tin wires that will be used in screw terminals of any type.

Solder is relatively soft and will flow over time, allowing the joint to become loose.

 

That's professionally been considered bad practice since the 60s and it is a specific item in electrical safety regulations as with power circuits, a loose connection over time can cause overheating or start a fire.

The regs do not apply to low voltage hobby wiring, but the reliability principles do.

 

Another downside is that you get a stress point at the end of the tinned area and the individual strands are more likely to fail there over time with movement or vibration.

 

A detailed explanation here:

https://cdn.thomasnet.com/ccp/00142951/263810.pdf

 

If you want to prevent the possibility of stray wire strands, the preferred was is to use wiring ferrules, aka Bootlace ferrules. They both hold the strands and act as a strain relief, giving some support to the wire insulation.

 

Hopefully this may save another faultfinding session in a few years time!

 

[drmditch]

Thank you RobjUK, I will take note.

Since this is a permanent layout, but installed and tested the wiring doesn't get moved much.

But, as you point out solder can flow over time.

 

Doubling the wire didn't work!

(or at least not on first testing.)

 

I shall have a light lunch, and then do some work on my 'parallel project' of rolling stock construction/improvement.

(I always find it best to have an interesting task in hand as well as boring ones!)

Apologies to anyone who finds wiring intrinsically interesting.

 

If nothing else, I've discovered that my logic for the matrix was not wrong, and that I've probably been over-optimistic in my cable-run length.

Some re-design is probably indicated!

 

[steve W]

If you are using multipls of 7/02, the add a 3rd or fourth one if you've got enough wire. They won't solder to the pm but a screw connection and single wire for the last couple of inches will help

 

[RAF96]

I use a GM CDU and can confirm it will punch 6 solenoid point motors in sequence. I did not have a diode matrix in circuit as I was using Rocrail route setting which allows a mark space to be set between firings if required.  

 

One thing of note is the surface mounted point motors tend to be more current hungry than under-board types so maybe being last in the queue point D is being starved of current. I don’t suppose there is a quick two wire swap that would allow you to reverse the firing order with D first for a test firing.

[St Enodoc]

1 minute ago, RAF96 said:

I use a GM CDU and can confirm it will punch 6 solenoid point motors in sequence. I did not have a diode matrix in circuit as I was using Rocrail route setting which allows a mark space to be set between firings if required.  

 

One thing of note is the surface mounted point motors tend to be more current hungry than under-board types so maybe being last in the queue point D is being starved of current. I don’t suppose there is a quick two wire swap that would allow you to reverse the firing order with D first for a test firing.

With a diode matrix as drawn and described all the points on a given route are fired simultaneously.

[DCB]

Your photo and diagram don't match. The diagram has F to the right of E, the photo has E to the right of F.

 I reckon the fault is in 5/ 6 roads.    I wouldn't use the new peco point motors in hidden sidings, the old ones have a nice straight pull the new ones work through an ugly right angle drive

 It looks to me as if the number of motors firing is not balanced.   Roads 1/2  have 4 motors to fire and  3/4 from have 5 motors to fire.or with 5/6 connected it becomes 6 and 7 respectively.

If the CDU is a bit feeble one motor may well not fire and that may well be D.   I would disconnect the motor for road 1 on the right hand side (E or F depending on diagram or photo)  and see if that cures it. if it does get a bigger CDU.

I had the opposite problem, my capacitor was so big  that when two points were thrown the point blades sometimes  pulled out of the tie bars. I added extra point motor coils so when only two points threw there was an extra coil in circuit to take some of the current. It works well.

My Matrix is just choc blocks with diodes strung between, very compact, makes fault finding much easier and you don't damage diodes when soldering.

[Broadway Clive]

Its an interesting concept you have here. I worked on London Transport railways as they were known and took inspiration from their idea of having localised controls in the form of AMRs (automatic machine rooms) that replicated a signal box. Control of which was then centralised at a remote control panel.

So from my own control panel, buttons send  12v 'commands' to remote boards that contain relays and a CDU with connections to local points, some of which are permanently wired together.  Safety interlocking, signal control, automatic working boards, and even panel indications can also use the 12v 'commands' so they've been very useful.

[Baby Deltic]

Bear in mind that each diode will have a voltage drop of about 0.6V across it in forward bias which, if several are in conduction at once, could affect the amount of current being pushed into the solonoids. 

[drmditch]

49 minutes ago, Baby Deltic said:

Bear in mind that each diode will have a voltage drop of about 0.6V across it in forward bias which, if several are in conduction at once, could affect the amount of current being pushed into the solonoids. 

Thank you. Do you mean in series or in parallel ?

[Baby Deltic]

In series the voltage drops add up so its actually around 0.7V the same as the base-emitter junction on a BJT. So three diodes in series would give a drop of 2.1V for example. There is very little data for parallel diodes short of me testing them on a breadboard and measuring the drops. I think the fact the diodes are sharing the current would reduce the voltage drop.

Edited March 26, 2020 by Baby Deltic