Really odd behaviour from Zimo chip

[MarshLane]

Evening all,

I am very baffled with this one, so if anyone can help shed light it would be appreciated.

 

Several months ago I fitted a Zimo MX696KS (Paul Chetter Sound Project) to a mates Heljan O Gauge Class 31.  All worked fine, sounds great.  Dispatched off to him (he lives on the sound coast, I am up 'norf).  A few weeks ago, without doing any changes to the CV values at anytime, the decoder started acting up.

 

If it is started up each day, it starts as it should and then runs normally.  However, if a period of 36-48 hours elapses between the sound being switched on, pressing the 'Sound On' function key on his NCC ProCab does absolutely nothing. No audio can be heard.  Movement and lights are fine and work as normal, and the loco can be run and controlled silently.  If the 'sound on' is left on .. then a period of time later (anywhere between 15-25 minutes) it just randomly kicks into life.  Shut it down, start it up again, works fine.

 

Body off, check for loose wires, all is fine and connected.  Reset the decoder, all appears fine, but the problem is still there.  There is a Zimo stay-alive fitted to it, which seems to be working ok, as if the loco is lifted off the track the audio continues for a few moments as it should.  Like I say, start it the next day (with power to the layout having been switched off overnight) its fine, but if there is a gap of at least a day in the loco being used, it does nothing.  There doesn't seem to be any logic to the delay, the shortest experienced is about 15 minutes .. the longest delay before the sound starts has been about 48 minutes.  It does not make any difference if anything else is running on the layout or not either.

 

Does anyone have any ideas?  I can understand a short (few seconds) random delay between pressing the button and starting to simulate a cold start but half-an-hour isn't right! It also didn't do this for the first few months after being fitted.  The chap that owns it is not confident at changing CV values at all (I do it for him!) so I know nothing has been changed, even so a reset hasn't sorted it.

 

Very bewildering! I am open to any thoughts/ideas....

 

Rich

[Ray H]

That does sound odd. It almost seems that the stay alive takes a long time to discharge and is, thus, keeping some kind of power flow to the decoder - which is what it is there for in a roundabout way.

 

Similarly, it sounds as though there needs to be a certain (time) amount of current flowing through the decoder (and, possibly, warming the decoder) before it'll work.

 

Could it be a dodgy solder joint on the decoder itself? I'd be tempted to refer the problem to the place of purchase if that is possible.

[WIMorrison]

I had a Zimo MX645 that would periodically refuse to run forwards on DCC but would always run correctly on DC, and always ran in reverse on DCC or DC. I checked it mechanically and could find no fault. Eventually I gave up and replaced the decoder and the issue was instantly cured.

 

You can send the decoder back to Zimo for repair/replacement, and as it is a sound decoder this may be worthwhile doing. In my case, the cost of repair/replacement was the same as purchasing a replacement, and it meant that me loco was not out of service for many months whilst Zimo worked their magic.

[Harlequin]

Is the decoder in a location where heat can't escape easily?

 

Zimo decoders seem to automatically turn down the audio amplifier circuit when they get hot.

 

I had similar symptoms as you describe on two separate Dapol GWR Railcars with MX645 decoders and the solution in both cases was to cut ventilation holes around the decoders.

 

Edited January 11, 2023 by Harlequin

[MarshLane]

2 hours ago, Ray H said:

That does sound odd. It almost seems that the stay alive takes a long time to discharge and is, thus, keeping some kind of power flow to the decoder - which is what it is there for in a roundabout way.

 

Similarly, it sounds as though there needs to be a certain (time) amount of current flowing through the decoder (and, possibly, warming the decoder) before it'll work.

 

Could it be a dodgy solder joint on the decoder itself? I'd be tempted to refer the problem to the place of purchase if that is possible.

 

Hi Ray,

Although I can follow your logic there, I dont think that is the issue.  If the loco has been run (for example) Monday afternoon, then come to it Tuesday afternoon and it works fine (and thats with all power to the layout switched off overnight so 12+ hours).  But if you didn't run it Tuesday and ran it Wednesday, the problem occurs. So while I thought about the decoder being 'warmed up' there's no supporting evidence of that.  But thanks for the idea, I have sent an email to @pauliebanger to see if Paul has any ideas as well.

 

29 minutes ago, Harlequin said:

Is the decoder in a location where heat can't escape easily?

 

Zimo decoders seem to automatically turn down the audio amplifier circuit when they get hot.

 

I had similar symptoms as you describe on two separate Dapol GWR Railcars with MX645 decoders and the solution in both cases was to cut ventilation holes around the decoders.

 

 

Hi Phil,

Yes its positioned in the middle of the Class 31, there is space above and below it - plus heat could escape through the roof fans and side vents on the loco.  But the same thing applies here, it would be stone cold after 12+ hours with no power.  But within 24 hours (roughly) of being used previously it works fine, more than that (ie 48 hours) the problem occurs, so I dont think its heat related.  Although again, thanks for the idea, and useful to know about the audio being turned down when they get hot!

 

Rich

[jpendle]

6 hours ago, MarshLane said:

Although again, thanks for the idea, and useful to know about the audio being turned down when they get hot!

The lights on the loco will flash as well, not much help in a kettle, but as it a diesel, it's probably not getting too hot.

 

Regards,

 

John P

[MarshLane]

Spoken to Digitrains this morning, who confirm that its something of a weird problem they haven't come across before!  Decoder is going to them in the next few days for testing and see what happens.

 

Thanks everyone for your thoughts and comments.

 

Rich