Can I wire two CDUs in parallel ?

[Largechris]

I’ve been working on a fiddle yard and I’m using a diode matrix for route selection, due to various track layout limitations ideally I’d like to power 7 turnouts (n gauge surface mount motors) at one time. My 4700uF CDU isn’t coping, so I have a 8800uF one on order.

 

If the 8800 still won’t work can I wire both CDUs in parallel (ie in parallel off the power supply then both feeding into the diode matrix board) or is that going to cause problems? (LED warning lights are in place to alert to the CDU being charged).

[kevinlms]

7 hours ago, Largechris said:

I’ve been working on a fiddle yard and I’m using a diode matrix for route selection, due to various track layout limitations ideally I’d like to power 7 turnouts (n gauge surface mount motors) at one time. My 4700uF CDU isn’t coping, so I have a 8800uF one on order.

 

If the 8800 still won’t work can I wire both CDUs in parallel (ie in parallel off the power supply then both feeding into the diode matrix board) or is that going to cause problems? (LED warning lights are in place to alert to the CDU being charged).

What is your input voltage and is it AC or DC? What length and what gauge wiring are you using? Those answer are far more relevant.

[Largechris]

6 hours ago, kevinlms said:

What is your input voltage and is it AC or DC? What length and what gauge wiring are you using? Those answer are far more relevant.

12V DC or 16V AC input Voltages, I'm not worried about charging time, I get a light telling me the capacitors are charged whichever I use.

Wire lengths are short, less than 0.5m to each turnout as I'm keeping the controls local to the fiddle yard board. 0.5mm wire for each point motor coil connection (there's only 0.5mm coming out of each Peco surface mount motor anyway so that's fixed) and 1mm for the return. 

 

 

[Michael Hodgson]

Sorry, simple answer is the chances are it won't work. 

 

Your problem isn't inadequacy of the capacitance available to do the job, which is what you are seeking to address by wiring two CDUs in parallel, the problem is that you are trying to move the 7 solenoids wired in parallel.  The trouble with that is that no two solenoids have exactly the same electrical characteristics, so some are drawing more current than others (albeit only briefly).  More importantly, each diode in your matrix is reducing the CDU's output voltage slightly - and you will have more diodes feeding one point in the route than the next one.  So some of the points are given a higher voltage kick than others.  The net result is that some of those solenoids aren't getting enough of a kick, whilst others are getting too much.  

 

I've successfully run both ends of a crossover (ie two points) sometimes three points at once quite happily off a single CDU.  But 7 at the same time is way too many.

 

The recommended approach is one (smaller) CDU per point, and the diode matrix should be triggering those individual CDUs rather than distributing the load of a big single one.

[Largechris]

52 minutes ago, Michael Hodgson said:

Sorry, simple answer is the chances are it won't work. 

 

Your problem isn't inadequacy of the capacitance available to do the job, which is what you are seeking to address by wiring two CDUs in parallel, the problem is that you are trying to move the 7 solenoids wired in parallel.  The trouble with that is that no two solenoids have exactly the same electrical characteristics, so some are drawing more current than others (albeit only briefly).  More importantly, each diode in your matrix is reducing the CDU's output voltage slightly - and you will have more diodes feeding one point in the route than the next one.  So some of the points are given a higher voltage kick than others.  The net result is that some of those solenoids aren't getting enough of a kick, whilst others are getting too much.  

 

I've successfully run both ends of a crossover (ie two points) sometimes three points at once quite happily off a single CDU.  But 7 at the same time is way too many.

 

The recommended approach is one (smaller) CDU per point, and the diode matrix should be triggering those individual CDUs rather than distributing the load of a big single one.

 

Thanks that's helpful. If I have more smaller CDUs triggered by the diode matrix, then these would be downstream of the matrix, and wouldn't charge up as the route button is a momentary contact?  

 

(The diode matrix is the DCC concepts one which is setup to have the CDU on the input side)

[Oldddudders]

In the prototype, be the signalman using levers, miniature levers, button or a mouse, multiple sets of points involved in setting up a route do not all throw at once, but in sequence. This is true even for a simple crossover. Reducing the number of points fed by one CDU would ensure reliability and be more authentic. 

[kevinlms]

2 hours ago, Largechris said:

12V DC or 16V AC input Voltages, I'm not worried about charging time, I get a light telling me the capacitors are charged whichever I use.

Wire lengths are short, less than 0.5m to each turnout as I'm keeping the controls local to the fiddle yard board. 0.5mm wire for each point motor coil connection (there's only 0.5mm coming out of each Peco surface mount motor anyway so that's fixed) and 1mm for the return. 

 

 

A 12 Volt DC input is vastly inferior to 16 Volt AC input for a CDU. 

You can check the voltage at the output of the CDU with your multimeter to confirm the difference. 

 

What voltage rating is the capacitor?

 

The time taken to recharge the CDU is usually not a major issue, but I found that a transistor based one charges up almost instantly, is superior, if you accidentally select the wrong route and want to correct your error.

[SamThomas]

2 hours ago, Largechris said:

12V DC or 16V AC input Voltages, I'm not worried about charging time, I get a light telling me the capacitors are charged whichever I use.

Wire lengths are short, less than 0.5m to each turnout as I'm keeping the controls local to the fiddle yard board. 0.5mm wire for each point motor coil connection (there's only 0.5mm coming out of each Peco surface mount motor anyway so that's fixed) and 1mm for the return. 

 

 

Do you have a common return for the solenoids ?

 

If so & even if your CDU is capable of changing 7 x solenoids at the same time you need to have a common return wire much larger than 0.5mm as it is carrying 7 x times as much current as the individual wires to the point motors.

[Chris M]

I move up to six point motors via a diode matrix at the same time with absolutely no problem.

Two things I have done:-

1. Arranged the storage sidings into a 3+3 arrangement rather than a ladder of 6. This reduces the number of points that need to be thrown at any one time.

2. I do have two CDUs. I have arranged the wiring so that each CDU only feeds half the points. Both CDUs feed to the cheap push to make switch, this is then split to two diode matrices, one for each half of the points. The feed then goes to the point solenoid and the common return goes to one CDU for half the points and the other CDU for the other half. No problems at all so far.

 

There is another simple solution. Use the diode matrix to activate relays and run the feed from the CDUs through the switch part of the relay. That way it is easy to create two separate circuits.

[Chris M]

This shows my how my storage sidings are arranged to reduce the number of points being fired at once.

[Largechris]

11 minutes ago, SamThomas said:

Do you have a common return for the solenoids ?

 

If so & even if your CDU is capable of changing 7 x solenoids at the same time you need to have a common return wire much larger than 0.5mm as it is carrying 7 x times as much current as the individual wires to the point motors.

 

Yep, a common return, I saw there was already a lively debate on this board about core sizes required for pulsed current with answers varying between size for biggest pulse (21 Amps in my case maybe so at least 2mm wire) and sizing for average current... anyway that's no big deal I will triple the return wire to 3mm.

[Largechris]

33 minutes ago, kevinlms said:

A 12 Volt DC input is vastly inferior to 16 Volt AC input for a CDU. 

You can check the voltage at the output of the CDU with your multimeter to confirm the difference. 

 

What voltage rating is the capacitor?

 

The time taken to recharge the CDU is usually not a major issue, but I found that a transistor based one charges up almost instantly, is superior, if you accidentally select the wrong route and want to correct your error.

 

Ok I will check the CDU output Voltage tonight and see if it varies according to 16V AC input or 12V DC input cheers

[Largechris]

14 minutes ago, Chris M said:

I move up to six point motors via a diode matrix at the same time with absolutely no problem.

Two things I have done:-

1. Arranged the storage sidings into a 3+3 arrangement rather than a ladder of 6. This reduces the number of points that need to be thrown at any one time.

2. I do have two CDUs. I have arranged the wiring so that each CDU only feeds half the points. Both CDUs feed to the cheap push to make switch, this is then split to two diode matrices, one for each half of the points. The feed then goes to the point solenoid and the common return goes to one CDU for half the points and the other CDU for the other half. No problems at all so far.

 

There is another simple solution. Use the diode matrix to activate relays and run the feed from the CDUs through the switch part of the relay. That way it is easy to create two separate circuits.

 

Yep that makes sense, the bigger CDU is on order anyway so depending on the results of that and using 16V AC, plus the bigger return wire (two minute job), I will price up relays versus another diode matrix, thanks

[NinOz]

If more capacitance is required why not just wire additional capacitors to your existing CDU.

 

Not an approach I would use to fire 7 with one blow.  I use individual CDUs, one for each solenoid, maybe two at a pinch, via a diode matrix.

[Largechris]

33 minutes ago, NinOz said:

If more capacitance is required why not just wire additional capacitors to your existing CDU.

 

Not an approach I would use to fire 7 with one blow.  I use individual CDUs, one for each solenoid, maybe two at a pinch, via a diode matrix.

 

I was wondering that, there's a power transistor on the CDU board, so didn't know if other components in the CDU would need uprating.

 

Individual CDUs sound like an option but you must need relays, which type do you use?

 

[RAF96]

Upping the supply voltage will have more effect than anything else.

Imagine the CDU caps as a tank of water feeding a hose (i.e. your solenoids). The pressure at the hose is dependant upon the height of the tank above the hose (i.e. supply voltage) not its capacity.

[Crosland]

The effect is dependent upon both the voltage and the capacitance. The difference is that the energy stored in the CDU increases with the square of the voltage whereas it increases linearly with capacitance.

 

Switching from 12V DC to 16V AC (~22V peak when rectified) is almost a doubling of the voltage so almost a squaring of the energy stored in the CDU. The caps should be rated at least 35V to guard against the AC supply being a bit higher when not loaded.

[Largechris]

5 hours ago, Crosland said:

The effect is dependent upon both the voltage and the capacitance. The difference is that the energy stored in the CDU increases with the square of the voltage whereas it increases linearly with capacitance.

 

Switching from 12V DC to 16V AC (~22V peak when rectified) is almost a doubling of the voltage so almost a squaring of the energy stored in the CDU. The caps should be rated at least 35V to guard against the AC supply being a bit higher when not loaded.

Thanks, tonight I’ve tested on my particular CDU and on 12V DC input I get 19.2V out, and on 16V AC input I get 21.0V out. 
 

I also doubled up the return wire, didn’t make a difference. However after poking around with the meter it seems I have a few shorts on the diode matrix that I need to sort out as I suspect I’m trying to drive more than the intended number of points on each push button switch.

 

If I do end up with small CDUs for each point I could still do with some recommendations on suitable relays please?

[RAF96]

I have used normal 12v car relays, such as those used for aftermarket spot lights, etc. These were wired so that the output of my DCC accessory decoder fired into the relay coil and the main contacts were used to port the output from my CDU to the solenoids. The pulse from the acc dec was long enough to pull in the relay and fire the CDU.

The problem is you need one relay per solenoid direction, so my 36 relay ‘plank’ was three feet long.

You can get small multi relay modules but you need to research these as not all are suitable for this type of use. E.g. solid state relays (SSRs) normally only operate from AC loads as they need the reverse cycle to reset the internal circuitry. You can get AC-DC SSRs but at greater cost. Mechanical relays handle both but also have their foibles.

Many such modules are designed to work with the hi-lo logic outputs from Arduino or RasPi and can have various voltage relays, so as stated do your homework before ordering up.

[Largechris]

1 hour ago, RAF96 said:

I have used normal 12v car relays, such as those used for aftermarket spot lights, etc. These were wired so that the output of my DCC accessory decoder fired into the relay coil and the main contacts were used to port the output from my CDU to the solenoids. The pulse from the acc dec was long enough to pull in the relay and fire the CDU.

The problem is you need one relay per solenoid direction, so my 36 relay ‘plank’ was three feet long.

You can get small multi relay modules but you need to research these as not all are suitable for this type of use. E.g. solid state relays (SSRs) normally only operate from AC loads as they need the reverse cycle to reset the internal circuitry. You can get AC-DC SSRs but at greater cost. Mechanical relays handle both but also have their foibles.

Many such modules are designed to work with the hi-lo logic outputs from Arduino or RasPi and can have various voltage relays, so as stated do your homework before ordering up.

 

Many thanks,

 

Slightly terrifying... yes in my head I thought one relay per PM coil would sort out the change over problem, but blimey that's so many components to wire up correctly.

 

(Sometimes I wonder if my blind loyalty to Peco is not helping, and whether using a more complete and engineered turnout solution eg Kato, Rocco etc. , even it costs a lot more,  would save so much faff.....)

[Crosland]

15 hours ago, Largechris said:

Thanks, tonight I’ve tested on my particular CDU and on 12V DC input I get 19.2V out, and on 16V AC input I get 21.0V out. 

 

That makes no sense  Are you certain its 12V DC input?

[Largechris]

1 hour ago, Crosland said:

 

That makes no sense  Are you certain its 12V DC input?

What can I say, I have a magic touch with electrics...

 

Should be 12V DC off the right hand side of a trusty H+M Duette, and 16V AC off the left hand side... I will double check the power supply Voltages tonight....

[newbryford]

2 hours ago, Largechris said:

What can I say, I have a magic touch with electrics...

 

Should be 12V DC off the right hand side of a trusty H+M Duette, and 16V AC off the left hand side... I will double check the power supply Voltages tonight....

 

12vDc from a Duette.....

 

More likely 14v or more with lots of background noise with higher peaks than that - hence the 19v from the caps with a 12v supply.

[Largechris]

7 hours ago, newbryford said:

 

12vDc from a Duette.....

 

More likely 14v or more with lots of background noise with higher peaks than that - hence the 19v from the caps with a 12v supply.

Tried to get a measurement but don’t think my Voltmeter was really upto it if it’s that lumpy, was reading maybe 16V DC and 17.2V AC but would need an oscilloscope or something to prove some numbers.

[newbryford]

15 minutes ago, Largechris said:

Tried to get a measurement but don’t think my Voltmeter was really upto it if it’s that lumpy, was reading maybe 16V DC and 17.2V AC but would need an oscilloscope or something to prove some numbers.

 

So much for it saying 12vDC on the Duette!