NCE Power Cab + DCC-C Cobalts = short circuits?

[Suzie]

I think the issue with the EB1 is that the PowerCab is more sensitive to the overload and trips first hence why they can't be used. The CP6 is not ideal, but without adding a booster is probably the only realistic option.

 

Without knowing exactly what has caused the failure of the cobalt (overload welding the switch contacts together, burning out the switch contacts so they are open circuit, motor taking too long to run and overheating the series resistor or motor) it is not easy to put a finger on the exact root of the problem, but there are a few things that can be looked at:-

 

1. Low voltage on the accessory bus - The way that the CP6 works means that as the load increases the available voltage reduces, and not in a linear way, so that when the bulb starts to glow the voltage will be too low to move the cobalts reliably. When all is sorted (see rest of post!) it will be best to run the accessories directly from PowerCab and not via the CP6. In normal operation there is unlikely to be anything that will cause an overload on the accessory bus, and if it does the PowerCab will shut down and you will know what has happened and that something serious needs to be looked at.

 

2. Shorts when Electrofrog points are operated - If you are powering the frog you must cut the links from the frog to the switch rails and wire the switch rails to the stock rails. There is no way to guarantee that the switch in the cobalt will operate while the switch rails are in mid travel, especially if you have left the over-centre spring in place. Since you are powering your cobalts from the accessory bus you need to use the auxiliary switch contacts on the cobalt to power the frog from the local track feed to the stock rails and not use the defined 'Frog' connection. If you don't rewire the point at least move over to using the auxiliary switch because that will just light the appropriate lamp on the CP6 while the point is moving rather than upset the cobalts.

[Mallard60022]

I've turned on the Cab meter and it is great. . About 0.4A draw with no loco's working but Juicers and a couple of motors in place, then around 0.7A with two big beasts hauling long trains and two loco's parked. I doubt there will ever be more than 4 loco's working (double heading) so I think I'm OK and thanks to Bromsgrove for his advice.

I forgot to mention that my 10 or so DCC 80 Juicers were overheating in that problem time a week or so back, as there was not enough power to 'flip' the relay (didn't even know the DCC ones had relays). If we had not noticed I fear they may have burnt out eventually. Could that have happened to your point motor, the overheating due to not enough power to throw the relay I mean? 

ATB with your problem imt.

Phil

Edited July 18, 2018 by Mallard60022

[imt]

Which Cobalts are you using? I have again checked my IP Digital motors and the figures I quote earlier are still what I see (30mA changing and 45mA stall), yet you are stating that your current draw is 10 times what I am seeing 

 

Sorry for the delay in replying.  I have several kinds of Cobalts from 2 early analogue ones controlled through an ADFX2 board through to the latest IP ones.  I hear what you say.  DCC Concepts have been unfailingly supportive and provided useful advice as I waded in to DCC control.  I have looked at their leaflet for the IP range which says "Power: 9~23v DCC or DC. static <5mA, active average <40mA".  I am only reporting what I have seen on the digital ampere display on the NCE PowerCab handset as suggested by BromsMods above.  It has only 2 decimal places to show, but isn't 30 thousandths the same as 3 hundredths which would therefore register as .03 on the handset?  I am well known for my oblique stupidity so I could be wrong again!  So a single point throwing would raise the background load by .04, multiple ones by a varying amount over time since they all start throwing in a staggered way.  Throwing a single point I see the load go from. 0.36 amps to a peak of 0.42 amps or a throw load of 0.05 amps (50 milliamps) - which seems to be a bit higher than expected.  I may be my idiocy, the relative accuracy of the simulated ammeter or other factors introduced by this happening with the whole layout attached.  What I intend to do is set up a test rig with ONLY two Cobalts on to see what registers on the load and what changes when 1 is fired or 2 are fired simultaneously.  Then at least I shall have something exact to talk to DCC Concepts about.

[imt]

.....

I forgot to mention that my 10 or so DCC 80 Juicers were overheating in that problem time a week or so back, as there was not enough power to 'flip' the relay (didn't even know the DCC ones had relays). If we had not noticed I fear they may have burnt out eventually. Could that have happened to your point motor, the overheating due to not enough power to throw the relay I mean? 

....

 

I think you are probably right in some way or another, though I am not really qualified to say so.  I will be following this up with DCC Concepts when I have some clear measurements without too many extraneous other factors.

[imt]

.....

Without knowing exactly what has caused the failure of the cobalt (overload welding the switch contacts together, burning out the switch contacts so they are open circuit, motor taking too long to run and overheating the series resistor or motor) it is not easy to put a finger on the exact root of the problem, but there are a few things that can be looked at:-

 

1. Low voltage on the accessory bus - The way that the CP6 works means that as the load increases the available voltage reduces, and not in a linear way, so that when the bulb starts to glow the voltage will be too low to move the cobalts reliably. When all is sorted (see rest of post!) it will be best to run the accessories directly from PowerCab and not via the CP6. In normal operation there is unlikely to be anything that will cause an overload on the accessory bus, and if it does the PowerCab will shut down and you will know what has happened and that something serious needs to be looked at.

 

2. Shorts when Electrofrog points are operated - If you are powering the frog you must cut the links from the frog to the switch rails and wire the switch rails to the stock rails. There is no way to guarantee that the switch in the cobalt will operate while the switch rails are in mid travel, especially if you have left the over-centre spring in place. Since you are powering your cobalts from the accessory bus you need to use the auxiliary switch contacts on the cobalt to power the frog from the local track feed to the stock rails and not use the defined 'Frog' connection. If you don't rewire the point at least move over to using the auxiliary switch because that will just light the appropriate lamp on the CP6 while the point is moving rather than upset the cobalts.

Thanks Suzie.  As ever very helpful. No details were given by the DCC-C engineer and I expect it's too late to ask now.

 

1. This is an interesting warning. At the moment - probably by accident or also maybe because there is no loco load on the track - all the points seem to throw, even in multiples.  However I 'll do as you say.

 

2. A good point for others reading this - I don't have any frog switching, display lights, switches or anything else attached to my Cobalts - just 2 wires from the accessories bus for power and command.

[WIMorrison]

Thank you for your help.  I have checked the manual and it does indeed say 2 amps.  Sorry I should have read more carefully.  I have also switched the ammeter on (if all else fails read the operating instructions!).  I have a background at my reduced level (described in post #17) of 0.16 amps.  Changing a crossover sees the draw go up at peak to 0.42 (so much for 30 milliamps!). Causing the reset macro to run as described above (multiple Cobalts fired one after the other) I get a max reading of 0.8 amps and a beautifully lit bulb. So the explanation is clear.  

 

Adding most of the rest of the load, including my 2 NCE Mini-panels gets me up to a background 0.34 amps.  Strangely the peak firing loads are 0.54 for a crossover but the maximum of 0.8 for the macro doesn't seem to change(??).

...

 

 

Sorry for the delay in replying.  I have several kinds of Cobalts from 2 early analogue ones controlled through an ADFX2 board through to the latest IP ones.  I hear what you say.  DCC Concepts have been unfailingly supportive and provided useful advice as I waded in to DCC control.  I have looked at their leaflet for the IP range which says "Power: 9~23v DCC or DC. static <5mA, active average <40mA".  I am only reporting what I have seen on the digital ampere display on the NCE PowerCab handset as suggested by BromsMods above.  It has only 2 decimal places to show, but isn't 30 thousandths the same as 3 hundredths which would therefore register as .03 on the handset?  I am well known for my oblique stupidity so I could be wrong again!  So a single point throwing would raise the background load by .04, multiple ones by a varying amount over time since they all start throwing in a staggered way.  Throwing a single point I see the load go from. 0.36 amps to a peak of 0.42 amps or a throw load of 0.05 amps (50 milliamps) - which seems to be a bit higher than expected.  I may be my idiocy, the relative accuracy of the simulated ammeter or other factors introduced by this happening with the whole layout attached.  What I intend to do is set up a test rig with ONLY two Cobalts on to see what registers on the load and what changes when 1 is fired or 2 are fired simultaneously.  Then at least I shall have something exact to talk to DCC Concepts about.

 

I think we might have a decimal issue somewhere here as you quote 0.16A as static load with 0.42A as firing load for a crossover which is .26A or 260mA, and you later suggest that you are seeing 0.54A for a crossing (same one) firing with 0.8A for the macro, presumably several motors firing at the same time?

 

This is why I am suggesting that your figures are significantly higher than expected by the docs from DCC Concepts and my test data which falls exactly in line with what they suggest, both of which are much lower than you have documented in posts - though I agree your last figures suggest a load of 60mA (0,06A not 0.05 ), but this is still 100% higher than I am seeing.

[imt]

Well - I have set up my test rig.  It is my (separate from the layout) programming/testing track with a PCP faceplate and PowerCab attached, and two pairs of crocodile clip flying leads to two point motors. NOTHING but NOTHING else is attached and the only load is 2 Cobalt Digital IP point motors. The point motors are still fixed under their baseboard and hence moving the points to which they are attached (have a physical load). Photo of rig attached.

 

The static load is .01 - as expected (two at .005).  The switching performance is amazing!  Sending a change direction to a single point (say from normal to reverse) results in .06 for some time during which time it moves the point mechanism, .09, .13 until stall, .04, .01 (the numbers vary a bit from time to time by the odd +/- .01).  Sending a command to reverse when already reversed results immediately in .13 again for some time (until it stalls again?) and then .01.  I would have thought if the point was in the correct alignment that wouldn't happen (i.e. commands would be ignored) - but it does this repeatably.

 

Now I repeat, I am reading the ammeter available on the PowerCab handset - which may not be terribly accurate.  I don't have access to a recording oscilloscope so I am writing down what I see. It's all very crude and amateur - but I think the readings are reasonably fair and correct.

 

I have not bothered with firing the 2 Cobalts at the same time yet.

 

Thoughts?

 

 

[WIMorrison]

In summary I think that what you have is;

 

• Static load = 5mA per motor, exactly as stated by DCC Concepts
• Stall current = 130mA - about 3 times what DCC Concepts suggest
• moving current = 60, 90, 10 or 40mA which would average 50mA which is about double I see and three times what DCC Concepts suggest.

 

I am not sure on the last point of moving current though - perhaps you could confirm the moving current.

 

The challenge for your light bulb is on the stall current as it doesn't take much to light it up and if the stall current is as you suggest then that will be your issue.

[imt]

In summary I think that what you have is;

 

• Static load = 5mA per motor, exactly as stated by DCC Concepts
• Stall current = 130mA - about 3 times what DCC Concepts suggest
• moving current = 60, 90, 10 or 40mA which would average 50mA which is about double I see and three times what DCC Concepts suggest.

 

I am not sure on the last point of moving current though - perhaps you could confirm the moving current.

 

The challenge for your light bulb is on the stall current as it doesn't take much to light it up and if the stall current is as you suggest then that will be your issue.

 

I agree with that.  Understanding why it is what it is gives me a headache.  Can you spot a fallacy in the test? Do your Cobalts have a physical load when you test them?

 

I know of no reason why my Cobalt Digital IPs should be different to anybodyelse's.

 

[additional]

 

The movement current is a list of values I see from the moment of pressing the command button to the return to 0.01.

 

Causing 2 Cobalts to fire at once produces a maximum load of 0.28 for a short time.

Edited July 19, 2018 by imt

[scottystitch]

Could it be higher because the springs are still in the points? i.e.it requires a bigger grunt to overcome the spring resistance?

 

Just thinking out loud.....

[imt]

Could it be higher because the springs are still in the points? i.e.it requires a bigger grunt to overcome the spring resistance?

 

Just thinking out loud.....

 

Yes the springs are still in. I have not had any problems with that until recently.  It needs a grunt to "push" past the centre but then the spring helps by joining in the push.  I generally don't muck about with my trackwork.

 

[additional]

 

I have just moved a set of jumpers to another Cobalt Digital IP (for those who have been reading all of this - the one which "shorted" and was repaired by DCC-C).  It is on a trap point.  Thus it is half a point AND this ones spring has been removed.  The performance is exactly the same as one with a spring - peaking at 0.13.

Edited July 19, 2018 by imt

[WIMorrison]

It does say on the DCC site that you should remove the detent springs - and that make sense as you will be getting a snap action on the blades rather than the smooth motion of the blades which is afterall why we buy the Cobalt (or Tortoise) motors

 

 

It is however easy to spring them out if you have accidentally left them in place.

[imt]

It does say on the DCC site that you should remove the detent springs - and that make sense as you will be getting a snap action on the blades rather than the smooth motion of the blades which is afterall why we buy the Cobalt (or Tortoise) motors

 

 

It is however easy to spring them out if you have accidentally left them in place.

 

Well I use them because I have a complex control set up and they are easy to use, and only 2 wires are required (yes those mythical two wires of DCC!)

 

Your post probably cross with my amendment above.  It don't make a difference to my Cobalts on my layout (performance is the same spring in spring out).

[WIMorrison]

The other reason to remove the spings is that it stops the point blades shooting over before the Cobalt switches have changed resulting in a short on the layout

[imt]

The other reason to remove the spings is that it stops the point blades shooting over before the Cobalt switches have changed resulting in a short on the layout

 

I appreciate that information is good for the general reader, but in my case since I don't do frog switching (or anything else) with the Cobalts it doesn't matter to me.

[imt]

I have reported my findings to DCC Concepts and asked if they are untypical.  I don't think it appropriate to print their personal reply, however they have confirmed that my findings are typical, which I summarise as follows:

 

Base load = 0.005 amps per unit. Which is what I found.

Switching current = 0.05 amps. For me it was 0.06 amps.

Stall current 0.12/0.13 amps for a short period.  Which is what I found.

The stall current applies whether the unit has a need to move or not. So sending a "reverse" to an already reversed point will see a 0.12/0.13 pulse.  Also what I found.

 

Given the crudity of my measuring device I accept that some of my slightly higher measurements are within tolerance

 

So now we all know.

[scottystitch]

I have reported my findings to DCC Concepts and asked if they are untypical.  I don't think it appropriate to print their personal reply, however they have confirmed that my findings are typical, which I summarise as follows:

 

Base load = 0.005 amps per unit. Which is what I found.

Switching current = 0.05 amps. For me it was 0.06 amps.

Stall current 0.12/0.13 amps for a short period.  Which is what I found.

The stall current applies whether the unit has a need to move or not. So sending a "reverse" to an already reversed point will see a 0.12/0.13 pulse.  Also what I found.

 

Given the crudity of my measuring device I accept that some of my slightly higher measurements are within tolerance

 

So now we all know.

Thank you for reporting back their/your conclusions.

[WIMorrison]

I wonder if they had any explanation as to why I see values much lower than yours, excepting the static load?

 

The current draw by a motor that is already thrown makes absolute sense as these are stall motors and therefore will always draw current to reach the stall position at which time they switch off.

[imt]

I wonder if they had any explanation as to why I see values much lower than yours, excepting the static load?

 

The current draw by a motor that is already thrown makes absolute sense as these are stall motors and therefore will always draw current to reach the stall position at which time they switch off.

 

I didn't ask them, but it may be variations in the accuracy of the equipment/method.  I am assuming you have better equipment than mine.

 

I bow to your greater knowledge - though why something which is in the right position needs to powered so as to move I am not clear.

[WIMorrison]

The reason it will draw current on an already thrown motor is because there aren’t any switches to physicallystop the motor drawing current - that is why you see the residual 50mA draw.

 

When you tell the point to throw it tries to move it, if it already fully in that direction it will draw stall current until the circuit realises it has stalled at which point it drops to residual. Had the signal been for it to move then the motor would have moved until it hit stall, and the cycle goes on.

 

One reason you may be seeing a higher current is I think you have a lower voltage, IIRC you said you use 12v at the track which will draw 35% more than me as I use 18v at the track.