I think I just fried a DCC loco chip - diagnosis assistance needed

[Pete the Elaner]

I am not sure what you find worrying?

I find it worrying that the diagram we both found on NCE's own website has such an important error on it.

 

As for track voltage:

Voltage is the electrical pressure which makes current flow in the same way that blowing through a straw moves air along it. In this case, pressure is similar to voltage & current is similar to air flow.

You will need some voltage to push the programming commands through the decoder, but not necessarily full operating voltage & in theory, these could just be short pulses when required. (I think Nigel said this 'signal as required' is what should actually happen. He knows a LOT more than me about how DCC actually works).

[NIK]

Not having an electronics background I don't understand what you did or what it proves or disproves. More explanation for an idiot please?

 

Hi,

 

My quick tests indicate for my NCE Power Cab there does not seem to be a low current limit when programming is selected.

 

So when programming a DCC decoder in a loco for the first time should there be certain wiring faults the current may be high enough to cause permanent damage to the decoder or the wiring.

 

Regards

 

Nick

[imt]

I find it worrying that the diagram we both found on NCE's own website has such an important error on it.

 

 

 

The error escapes me.  I have been referred to information published by Mark Gurries, much of which I have already read with interest.  You may wish to note that on:

 

https://sites.google.com/site/markgurries/home/nce-info/nce-accessories/nce-autoswitch

 

that the diagram is posted as a solution to a problem in the AutoSwitch - i.e.it switches off the main track power and bypasses the unit entirely for programming.

 

Unfortunately I cannot find anything from him about the actual process of "programming".  Sadly I won't be looking on Yahoo - its a jungle with unforeseeable consequences.

 

I have has volts and amps etc explained gently to me in the past - usually with reference to water pipes.

Edited September 11, 2017 by imt

[imt]

Hi,

 

My quick tests indicate for my NCE Power Cab there does not seem to be a low current limit when programming is selected.

 

So when programming a DCC decoder in a loco for the first time should there be certain wiring faults the current may be high enough to cause permanent damage to the decoder or the wiring.

 

Regards

 

Nick

 

Thanks Nick, perfectly clear.  This MAY be the cause of my fried chip, or may not.  It's on it's way to DCC Concepts who will tell me the word in due course.

Edited September 11, 2017 by imt

[Torper]

Yahoo is where Mark Gurries posted it.  I'm not going seek permission to copy his document to this forum.

Mark Gurries testing of the NCE Autoswitch is thorough.  If wanting to know what's going on with programming from a PowerCab with and without the Autoswitch, then read that document first.

  I didn't actually ask you to copy it, Nigel.  If I wanted it badly enough I could join Yahoo.  I don't want to join Yahoo, therefore I have to do without Mark Gurries' document.

 

Or do I?  Thanks to Google, I found this on a Public site - https://sites.google.com/site/markgurries/home/nce-info/nce-accessories/nce-autoswitch. Although obviously briefer and simpler than the Yahoo article, Mark Gurries appears both to confirm that the Autoswitch is little more than an automatic DPDT switch, albeit a useful one, and also confirms that the NMRA's definition of a true programming track is one where the "track's power (Current) is intentionally limited to only allow safe programming of the decoder and nothing else".  As he seems to take the view that a combination of the Powercab and the Autoswitch fulfils the NMRA definition, the voltage in the rails may be a red herring as it seems clear from what Mark Gurries says that the switch "preserves the Program Track low current operation at all times".  So I can only assume from this, and from the manual, that in Programming Track mode, the PowerCab delivers the same voltage to the track, but a much reduced amperage.

 

DT

 

EDIT:  I've just seen Imt's posting at #78 - it seems that we've been following the same lines.  The only difference is that I construe Mr Gurries' article as confirming that in Programming Track mode the PowerCab complies with the relevant NMRA definition, whileas Nick's tests referred to above tend to contradict that!

Edited September 11, 2017 by Torper

[imt]

 

EDIT:  I've just seen Imt's posting at #78 - it seems that we've been following the same lines.  The only difference is that I construe Mr Gurries' article as confirming that in Programming Track mode the PowerCab complies with the relevant NMRA definition, whileas Nick's tests referred to above tend to contradict that!

 

Yes I think we are, but none the worse for that - we both seem to be finding out more as we go on.

 

You may also have had the experience I have had, where having decided to use NCE mini-panels to automate my layout, I found the NCE documentation poor and confusing.  This seems just another example.  That's how I came across Mark Gurries originally: he had written some amendments for the manual.  It seems that NCE are good engineers, but not good at documenting things (not unusual in my experience).  So, having a lot of experience of using their stuff, and having read what you have read, I agree with you that it very probably does do what it should do - I just cannot prove it.  Most of the other people posting on here seem also not to be able to prove things one way or another, or indeed explain what should happen in programming a DCC chip.

[NIK]

Thanks Nick, perfectly clear.  This MAY be the cause of my fried chip, or may not.  It's on it's way to DCC Concepts who will tell me the word in due course.

 

Hi,

 

I don't have any definite info on why your decoder stopped working. Your decoder failed after it was programmed so any lack of current limiting in programming mode was probably not involved.

 

I have seen poor soldering on the underside of commercial DCC sockets and I've seen a 8 pin decoder emit smoke after being plugged in (presumably some of the decoder pins touched the ballast weight below the socket). I've had programming a DCC decoder fail due to a spike of solder under one of the commercial DCC sockets pierce the insulation of a wire going to the motor. In this case the decoder did not burn out possibly because I was programming using the programming output of an NCE Power Pro which I think has a low current limit. 

 

So my best guess with your case is that an accidental short circuit was formed when the loco was turned upside down.

 

Please excuse me if this question has been asked before but the NCE Power Pro was physically disconnected from the programming track before connecting the DC controller to check DC performance of the loco?.

 

Regards

 

Nick

[imt]

 

Please excuse me if this question has been asked before but the NCE Power Pro was physically disconnected from the programming track before connecting the DC controller to check DC performance of the loco?.

 

...

 

Nick

 

No it hasn't.  My programming track is a piece of ST 604 on a piece of plywood.  There is a red wire soldered to one rail and a black wire to the other (black to the back - my mantra as I wired the droppers on my main system).  The 2 wires go to a piece of choc block.  If I want DC I take my Gaugemaster Combi outputs and screw them colour coded into the chock bloc (I had thought of using the plug in style but I had run out).  I then plug the wall wart output into the Combi and then the wall wart into the wall socket. I finally put the loco on the track. Hey presto a DC test bed.  If I want DCC I take my box with an NCE PCP set into it and take the outputs (colour coded again) and screw them into the choc bloc - I won't bore you with the rest.

 

It is of course possible to forget which is which - i.e. to set up a DC test and put a DCC loco on the rails.  I am not aware of having done so - but I get old and frail.

[Pete the Elaner]

The error escapes me.  I have been referred to information published by Mark Gurries, much of which I have already read with interest.  You may wish to note that on:

 

https://sites.google.com/site/markgurries/home/nce-info/nce-accessories/nce-autoswitch

 

that the diagram is posted as a solution to a problem in the AutoSwitch - i.e.it switches off the main track power and bypasses the unit entirely for programming.

 

That article describes version A & version B, stating that version A does not actually work properly - it provides power to the programming track at all times. The wiring diagram is to make a version B board (which works correctly) work like a version A board. I don't see the point of this.

 

If you have a board marked REV B, then it should be connected like this:

https://www.dccconcepts.com/manual/nce-owners-manual-program-track-auto-switch/

 

If you have a board marked REV A then according to the Mark Gurries' page, you can connect the PROG wires to the 2 unused terminals of the relay, which will then switch on/off the programming track properly.

 

If you have recently bought a REV A board in the UK then surely under consumer rights, you could take it back for a refund/replacement because it simply does not work properly & is therefore unfit for its intended (designed) purpose? Your rights are with the retailer, not the manufacturer.

[imt]

That article describes version A & version B, stating that version A does not actually work properly - it provides power to the programming track at all times. The wiring diagram is to make a version B board (which works correctly) work like a version A board. I don't see the point of this.

 

If you have a board marked REV B, then it should be connected like this:

https://www.dccconcepts.com/manual/nce-owners-manual-program-track-auto-switch/

 

If you have a board marked REV A then according to the Mark Gurries' page, you can connect the PROG wires to the 2 unused terminals of the relay, which will then switch on/off the programming track properly.

 

If you have recently bought a REV A board in the UK then surely under consumer rights, you could take it back for a refund/replacement because it simply does not work properly & is therefore unfit for its intended (designed) purpose? Your rights are with the retailer, not the manufacturer.

 

Well, as far as I can see the power for the program track was taken off before the auto-switch for Rev A.  But anyway all of this does not in any way indicate either whether the NCE Power Cab programs at "reduced energy" as defined by NMRA or whether (If it does not) the auto switch actually does anything to help.

 

I believe 1) that the NCE PowerCab performs correctly and 2) the auto switch is what it says and contributes nothing to the process (or the argument for that matter).

 

Given that obviously neither of us knows for a fact what the situation is this is all a bit pointless.

Edited September 12, 2017 by imt

[Pete the Elaner]

Sorry, I was only trying to help.

I know I would want an accessory such as this connected correctly.

[imt]

Sorry, I was only trying to help.

I know I would want an accessory such as this connected correctly.

 

I apologise if I sounded scratchy.  I am sure that you were helpful to people reading this thread who have one of the auto switches: that does not include me, nor does it answer the question which I really want an answer to. Sorry if I was being proprietorial about my problem.

[imt]

The word is back from DCC Concepts (and it's not good!):

 

"Decoder received and tested.

 

All CVs are unable to be altered - all reading as 255 - including address, manufacturer ID, acceleration, deceleration etc. It will not do a decoder reset.

 

This is one sign of a processor that's been fried with either voltage spike or incorrect connection. There's no sign of any overload/heating."

 

They are very generously replacing the unit as the only person who could have caused this was me.  Sadly I am still none the wiser as to what I did wrong.

 

​Anybody got any advice please?  Remebering the question:

 

1. ​I installed an 8 pin DCC socket with harness into an old Lima chassis and tested that everything worked on DC with a blanking plate.
2. I put a generic DCC chip in and everything worked
3. I put in a DCC Concepts Zen 360 8 pin direct chip and it worked well, better with adding a DCC-C stay alive and then changing CVs 2,6,5.  Impressive slow crawl and good speed response.
4. Removed the chip with stay alive attched so as to be able to cut off and replace the old tension lock couplings with NEM pockets (I use Kadees).  There was a lot of cutting and sticking but no metal was involved - all plastic.  I also eased the chassis to accept a Hornby body.
5. Replaced Zen chip with it's stay alive still attached, and it wouldn't work again.
6. Inserted generic chip and it worked.

What could I have done before 5 that could have caused the problem and then between 5 and 6 reversed it?

 

By the way the loco still runs on the generic chip - but not as well (as I would expect without a stay alive and no CV changing)

[smokebox]

Other than static electricity causing the damage, nothing obvious springs to mind. Just put it down to "one of those things".

[Torper]

And of course you'll now be worried that if you put the replacement Zen decoder into the loco it'll rush off again in a puff of smoke....

 

I can only suggest that after you install the new decoder you first put the loco onto the programming track with the command station in programming track mode and take it from there.......

 

Incidentally, Sod's law always ensures that if you have a loco that runs well but then carry out some alteration to it, however unconnected that may be to the mechanism, the loco will never run as well when you eventually put it back on the track as it did beforehand. (possibly this applies more to kit built locos rather than RTR ones)

 

DT

[imt]

......

 

Incidentally, Sod's law always ensures that if you have a loco that runs well but then carry out some alteration to it, however unconnected that may be to the mechanism, the loco will never run as well when you eventually put it back on the track as it did beforehand. (possibly this applies more to kit built locos rather than RTR ones)

 

DT

 

Darn!  There you are I KNEW I had forgotten something simple.  Thanks Torper - at least you made me laugh!

 

Yeah - I will certainly be careful as to how I handle the replacement chip.  Write down what you are going to do, read it through, check with a friend, do it under controlled conditions.

Edited September 13, 2017 by imt

[Izzy]

The word is back from DCC Concepts (and it's not good!):

 

"Decoder received and tested.

 

All CVs are unable to be altered - all reading as 255 - including address, manufacturer ID, acceleration, deceleration etc. It will not do a decoder reset.

 

This is one sign of a processor that's been fried with either voltage spike or incorrect connection. There's no sign of any overload/heating."

 

They are very generously replacing the unit as the only person who could have caused this was me.  Sadly I am still none the wiser as to what I did wrong.

 

​Anybody got any advice please?  Remebering the question:

 

1. ​I installed an 8 pin DCC socket with harness into an old Lima chassis and tested that everything worked on DC with a blanking plate.
2. I put a generic DCC chip in and everything worked
3. I put in a DCC Concepts Zen 360 8 pin direct chip and it worked well, better with adding a DCC-C stay alive and then changing CVs 2,6,5.  Impressive slow crawl and good speed response.
4. Removed the chip with stay alive attched so as to be able to cut off and replace the old tension lock couplings with NEM pockets (I use Kadees).  There was a lot of cutting and sticking but no metal was involved - all plastic.  I also eased the chassis to accept a Hornby body.
5. Replaced Zen chip with it's stay alive still attached, and it wouldn't work again.
6. Inserted generic chip and it worked.

What could I have done before 5 that could have caused the problem and then between 5 and 6 reversed it?

 

By the way the loco still runs on the generic chip - but not as well (as I would expect without a stay alive and no CV changing)

 

 

I have no experience of using stay-alives, but is there any chance it was no.4? That removing the chip with the stay-alive connected - if it had a charge in it perhaps? - could have caused a short that fried the decoder. What, if any, tool did you use to ease the chip from the socket, I find I need to use something, the NEM 652 8-pin being quite tight when fully inserted unlike the Nem 651 6-pin type. If it was metal?  It's just a thought looking at the list.

 

Izzy

[imt]

I have no experience of using stay-alives, but is there any chance it was no.4? That removing the chip with the stay-alive connected - if it had a charge in it perhaps? - could have caused a short that fried the decoder. What, if any, tool did you use to ease the chip from the socket, I find I need to use something, the NEM 652 8-pin being quite tight when fully inserted unlike the Nem 651 6-pin type. If it was metal?  It's just a thought looking at the list.

 

Izzy

 

That's some good thoughts.  DCC Concepts say the stay alive shouldn't have caused problems, but I did nothing to discharge it before I took the chip out so maybe that's it. I have a chip remover which is plastic (kind of tongs), but it takes some fiddling about.  As you say they are quite a tight fit.  Maybe I bridged something - but that would mean I needed power about, as you say maybe the stay alive.

 

I am going to have to be so careful next time!

[John ks]

Electronic components generally last for a long time

But there is a very small percentage that fail(for no apparent reason) very early in their life

You may have been the unlucky person to get one of these

 

John

[imt]

Thank you everybody for your help and encouragement.  There is a happy ending after all (apart from DCC Concepts generously replacing the damaged chip).

 

Having had this experience I am very wary now of what I do and how I do it.  

 

I have now re-fitted the Lima 37 chassis with a Zen 360 Direct and its stay alive.and it works very well just as the original one did. This one came from another source as decoders seem to be becoming scarce - or at least DCC Concepts ones are.

 

I was scrupulous about keeping the chassis on the bench, wearing a static lead and insulating EVERYTHING.  I could probably have covered the full size one with all the electricians' tape and Kapton I used (I exaggerate of course, but only a little!).  The weight is well stuck down and all the wires tucked away and stuck down, except for the orange and grey ones attached to the motor bogie which requires freedom of movement.  Only when I was sure that eveything was well insulated and stuck did it get anywhere near my programming track, which was pre-set into program mode.

 

The result is that the Zen provides a very controllable model with a response which is far better than I expected given that it is only an old pancake motor (and I am aware of the slagging off they receive!).  Now I hope it has a long and useful life ahead of it.

[Torper]

Of course, you won't have done a proper test until you've taken the loco to bits again, carved it about, tried to fit a new body, and then put it all back together again.........

 

DT

[imt]

Of course, you won't have done a proper test until you've taken the loco to bits again, carved it about, tried to fit a new body, and then put it all back together again.........

 

DT

Thanks so much for the suggestion, but I don't think I'll bother!