Constructing a diode matrix?

[Calimero]

I'm trying to construct a diode matrix which is pretty large and being the first time I've done I'm unsure as to what to use for the board. Is there a specific type of board to use to solder the diodes to, Maplin do a matrix board is that suitable?

 

Probably a straight forward answer but of course any input and assistance is much appreciated.

[Suzie]

There are several different ways to do it, and how you will go forward will depend on how you like to wire things up.

 

Probably the easiest to visualise and maintain is the simple double row tag strip with each diode wired between a pair of tags. Input from route buttons is on one side of the strip, and the outputs to the point motors on the other side.

 

If you are a bit pushed for space you can use strip board in the same way, but it can sometimes be tricky to get the 1N5401 diode's wires through the holes.

 

If you want to be really compact you can make a sandwich of stripboard with all the diodes mounted vertically between two pieces of stripboard, but you have to get it right first time - it is not possible to get it apart once you have made it!

[Calimero]

There are several different ways to do it, and how you will go forward will depend on how you like to wire things up.

 

Probably the easiest to visualise and maintain is the simple double row tag strip with each diode wired between a pair of tags. Input from route buttons is on one side of the strip, and the outputs to the point motors on the other side.

 

If you are a bit pushed for space you can use strip board in the same way, but it can sometimes be tricky to get the 1N5401 diode's wires through the holes.

 

If you want to be really compact you can make a sandwich of stripboard with all the diodes mounted vertically between two pieces of stripboard, but you have to get it right first time - it is not possible to get it apart once you have made it!

Is the 1N5401 Diode the best to use as I've also seen 1N4001 and 1N4148 quoted in various articles.

 

Plan is for the matrix to work a number of Cobalt Slow Action motors, using push to make buttons so e.g. I press a push button for platform 1 and hold and six motors will operate to do the route, any flank protection will operate as well. I've drawn out my matrix chart and will need a large number of diodes.

 

I like doing this as it's a bit similar to real practice!

[Gordon H]

You will need something more than the usual simple unipolar diode matrix if you want to use slow action motors this way.

Slow action motors need to have the polarity of their drive current reversed to changeover, so although you could easily arrange to apply (say) the positive supply to the required set of motor terminals with a diode matrix, you also need to deal with the negative supply and switch that to the other motor terminals at the same time. You effectively need two matrices, one for the positive supply and one for the negative.

[Suzie]

Since the cobalts do not need constant supply to keep them in place you can use a split rail power supply and two diode matrices - one for each power supply rail.

 

You will need DPST push buttons and the low current draw of the Cobalts (30mA) will enable you to use 1N4148 diodes (which will fit on stripboard OK). You just have to make sure you hold the button down for long enough for all the motors to move.

[jamesrj]

As you are using Cobalt motors the IN4001 is a general 1A diode, the IN4148 is what we term a signal diode          .Have a look at a previous post regarding these motors on half wave.

 

If you used a centre tapped  supply and connected one side of all your motors to the centre tap then you would need two diodes connected in inverse parallel   connected to the other motor terminal, with the Positive supply via your push button to one diode and the Neg supply to the other diode via your other push switch  (ie  both diodes connected back to back at one end only going to the motor, the other 2 free ends going thro your switches.).  Re Gordon H's info

 

james

[davidcrabb]

Hi, I use Cobalts and needed multi point operation for route selection.

I use latching relays from Heathcote Electronics with 3PDT & 4PDT switches which negates the need to hold the switches down!

They have 2 relay and 4 relay DP relays on the boards which very versatile.

I have some fairly complex point work one route selection requires 8 points to change, fortunately they are in pairs so only require one relay board, I also use a rotary switch so that one relay board can be used for multiple route selection.

[Suzie]

The 1N4148 is a very cheap diode which can be easily fitted to stripboard that is good for up to 150mA, which is more than enough current in this application where the manufacturer has quoted a maximum load of 30mA for the Cobalt. 1N4001 will be fine too and in this application 1N5401 will be overkill.

 

1N5401 is suitable for solenoids where currents are in excess of 3A which neither 1N4001 or 1N4148 will be suitable for.

[Calimero]

I'm a little confused now as some of the online stuff I've seen show different matrices some with diodes controlling feed from the motor, others controlling feed to the motors and others with a bit of both.

 

Some show push buttons others switches!

 

I'll try and post a pic of my wiring diagram if that helps.

[Suzie]

A diode matrix is always going to be a bit of a bird's nest, and doubly so when using slow motion reversible motors. Programmable route setting solutions are available if you do not want to face the complexity of a diode matrix.

 

Here is a little diagram showing a sample matrix for cobalt and similar motors. The three switches 1-3 are double pole push buttons and each one can set a route. Point motors are A - C.

 

 

Each motor that is part of a route has a single diode going to the switch for the route, it can either go to the negative or positive depending on which way the motor needs to go for that route. You can see from my diagram that while A and C are part of all three routes, motor B is only part of two routes and it will not change when button 2 is pressed.

 

It is quite easy to build your matrix once you get started, you just need one diode for each point moter per route and one double pole pushbutton per route. just add route buttons and motors as required.

[Calimero]

Slowly getting my head round this I think?

 

 

I've drawn up my matrix plan (attached), 1 to 10 on the left feed from a push to make switch or a relay (still deciding or a push to make which then operates a relay), A to J across the top are the point motors either in normal or reverse. Diodes go to the relevant detection of the points either N or R dependent on route set.

 

Now I know this should work with a solenoid motor but I still am thinking it would work with a cobalt, some said you need to have a matrix for the opposite side would those diodes go the opposite direction to the ones shown?

 

I'd be grateful if anyone could amend the diagram to show what I need to do??? I'm still thinking e.g if I press (and hold) route 1 power will feed through via diodes to say motor A and the motor will move to the required setting along with any others related to that route. That button doesn't need to move A to reverse as that would be done by say route setting button 7a which would place those motors to the opposite direction.

 

Again thanks for your help, my head is going to go pop!

[Suzie]

With a slow motion motor fed from a split supply (as I have shown above) one end of the motor is common, and the other end is fed with positive or negative depending on whether you want normal or reverse. For each motor that is part of a route you will have a forward diode from a positive feed of the point has to go to normal, and a backwards diode from the negative feed for any points that have to be reversed. That is why you need double pole pushbuttons (you can buy them, and they will be cheaper than single pole buttons with additional relays).

 

There are other less efficient ways of doing it with a big bank of resistors and a higher voltage power supply, but it is probably only suitable for the very keen since it is reliant on knowing the current consumption of the motors reasonably accurately for consistent and reliable results..

[Dazzler Fan]

+

This tread seems to have come to a stand still some time ago,

just when it was getting interesting.

 

In the "Shows You How Series", Peco issued a booklet 'Wiring

the layout - Part 2 for the more Advanced. Under Advanced Point

Switching the steps are given to translate routing using a Diode

Matrix to a piece of Veroboard using the "N" & "R" notation.

 

Each of the routes A-D fires 4 solenoids which takes a lot of current.

via a CDU. The tracks on the Veroboard are not heavy enough with

all the holes to take the current being supplied on the Bus for A-D.

The solution of course is to tie the tails of the Diodes on each track

and take them back to their respective terminals A-D. On the diagram

it would be best to invert them as D-A so they have a clear uninterupted

path to their terminals.

Peco have not re-printed the booklet for obvious reasons, but it still stands as a good

[28XX]

0.1" Vero board is only good for about 0.5A. While the load is admittedly only momentary, I think this is a weakness in design.

[Dazzler Fan]

Agreed.

 

Somewhere I have the spec from Veroboard.

[gordon s]

A diode matrix is always going to be a bit of a bird's nest, and doubly so when using slow motion reversible motors. Programmable route setting solutions are available if you do not want to face the complexity of a diode matrix.

 

Here is a little diagram showing a sample matrix for cobalt and similar motors. The three switches 1-3 are double pole push buttons and each one can set a route. Point motors are A - C.

 

Diode routing matrix.png

 

Each motor that is part of a route has a single diode going to the switch for the route, it can either go to the negative or positive depending on which way the motor needs to go for that route. You can see from my diagram that while A and C are part of all three routes, motor B is only part of two routes and it will not change when button 2 is pressed.

 

It is quite easy to build your matrix once you get started, you just need one diode for each point moter per route and one double pole pushbutton per route. just add route buttons and motors as required.

 

Some time ago, I built a diode matrix for solenoid motors and it worked very well, but having changed over to Tortoise motors, I was really struggling with the split voltage design, so pushed it to one side.

 

Just seen Suzie's diagram, but as my electronics knowledge is fairly limited, I have a question.  

 

Route setting itself is not an issue, but will your circuit deal with individual switching as well as route setting.  Take turnout A for example.  It may be used in several routes and also be used as a stand alone turnout outside it's route setting capability.  There may be an occasion where you simply want to switch turnout A on it's own and not part of a pre programmed route.

 

Does that circuit address that issue?

 

With solenoid motors it did not matter which way a turnout was set as setting a new route automatically reset turnouts to suit the route.  With a split voltage circuit, I don't have sufficient understanding to ascertain whether that would work if one turnout in the circuit had been manually set to be in the 'wrong' direction. 

 

I hope that makes sense.....:-)

[Suzie]

It will work for setting an individual turnout - you just create it as a route with one turnout in it. It will just be a single pole momentary switch or pushbutton with a single diode in series connected to the relevant motor and supply.

[gordon s]

Apologies, Suzie, but I was under the impression Tortoise motors need a continuous supply and are operated via switches, not push buttons.  How will they work with a momentary switch?

 

This is how mine are wired.

 

 

 

[Suzie]

If using a matrix you will need motors that stay put when power is removed.

[gordon s]

I seem to recall this was why I abandoned a diode matrix approach last time I looked at this problem.  I believe Tortoise motors will only work with changeover switches and not push buttons.  The odd push button circuit I saw had to work via relays or decoders.

 

My understanding is that a Tortoise motor will stay in position once changed, but needs a continuous supply to move it from one side to the other.

 

If anyone can come up with a 'routing' circuit that will work with Tortoise motors and a 12/0/12v supply then I'm all ears...

 

I'd also want individual switching as well..

[gordon s]

After much searching I found my original post on this subject.  Sue and Fred Horne had a circuit for split rail Tortoise motors, but that couldn't accommodate individual switching of one turnout that may be part of several routes.

 

http://www.rmweb.co.uk/community/index.php?/topic/20702-diode-matrix-and-stall-point-motors/

[Suzie]

A latching relay will make any stall motor compatible with a solenoid matrix. It is probably only tortoise you have to worry about, end-cut-off type motors will stay put.

[brigo]

Funny old world. Over 25 years ago I had a friend who had a large layout with two large storage yards and he wanted simple push button selection of the various tracks, ie a diode matrix. Trouble was he was using Bemo point motors. I designed and built a standard PCB module which took momentary inputs from switches and drove the point motors accordingly. Each module could drive 6 point motors. It basically consisted of a couple of logic gates as a flip-flop with a transistor bridge output, throw in a few protection diodes and a voltage regulator to power the lot. ( Voltage regulation is useful with most motor driven point motors as it controls the amount of overrun.) To operate, the inputs were  momentarily connected to the 0v rail either directly through a push switch or via a diode matrix.

I did try at the time to sell it as a product but there was no interest, and now a lot of people are into DCC it would probably not be technical enough.

I still have some bits somewhere and they will be used on my next layout, when I can get around to it.

 

Brian 

[DCB]

All my matrixes have been two parallell blocks of screw connector strips screwed to hardboard with 1N4001 or similar 1 amp diodes strung between them, the diodes left full length and sleeved to insulate them.  They have been working faultlessly for circa 20 years with big beefy current hungry H&M point motors.    There has to be an easier way to wire slow action point motors than using two complete matrixes.   Needs some lateral thinking.

[Suzie]

... There has to be an easier way to wire slow action point motors than using two complete matrices.   Needs some lateral thinking.

 

The lateral thinking is to use a split rail supply, and have two half matrices. Each motor has a common wire to the PSU 0V, and the feed wire goes to the matrix via a diode (one for each route) just round the correct way for the direction that route requires.

 

If you want to do away with having two half matrices and the split supply - just use AC instead at double voltage. Use fast diodes and it may be possible to use the DCC track supply in place of industrial frequency AC.

 

Lots of ways to do it, just not many people tend to bother when using slow motion motors so I am not aware of anyone having used any of these methods.