Another Stay-alive problem!

[Torper]

I've just fitted a kit-built loco with a KA2 Stay-alive and a Zimo MX634 decoder.  Unfortunately, when the KA2 is fitted the decoder and motor ignore the Powercab controller and run when they shouldn't, only stopping when the power supply is switched off at the wall.  It runs fine without the KA2.   The only thing I can think of that would cause this problem is if the CVs permitted DC running.  I have CV29 set at 34 which, as far as I can see, comprises Bit 1 (speed steps) and Bit 5 (long address).  It shouldn't allow for analogue running. Am I right in this?

 

DT

Edited April 13, 2020 by Torper
Clarification

[BromsMods]

It is possible that the MX634 is not compatible with the KA2.  Have you followed the notes concerning use of the KA2 with non-TCS decoders?

 

Not all DCC decoders are Keep Alive® ready. TCS decoders use software to enable and disable KeepAlive, so if your decoder does not support Keep Alive®, it may not work. Keep Alive® supported code for TCS decoders was implemented fully by November 2012. If the date code in your decoder is before this date, Keep Alive® may never enable. CV's 250, 251, and 252 read as the month, day, and year of manufacturing.

If you would like to attach a Keep Alive® unit to a non-TCS decoder, you need to connect the wires across the output of the bridge rectifier on the decoder. You will need to consult your decoder manufacturer for a schematic or wiring diagram.

[Torper]

I don't know when the MX634 was introduced but I certainly bought mine well after 2012.   CVs 250, 251 and 252 don't make much sense as they read 240, 001, and 026 respectively.  I'm pretty sure that the MX634 should be able to work with the KA2 as apart from anything else it is equipped with two wires specifically to attach to a Stay-alive.  The manual does state that the total capacity of any connected capacitor must not exceed 5000uf, and the KA2 clearly does exceed that.  However, I have read a number of posting on this forum that say that that can be disregarded with the KA2.  Moreover, after the problems with the KA2 I replaced it with a 3300uf capacitor but the result was exactly the same.

 

DT

Edited April 13, 2020 by Torper
Correction

[Nigelcliffe]

I've used the KA1 and KA2 on a variety of Zimo decoders, and every time it works fine.   So, discount the "compatibility" issue.

 

When I've connected them, I've gone to the decoder positive and decoder negative.  This is NOT the same as the pins provided on some decoders for stay-alive where there is partial or full support on the decoder for charge management.    I took the view that the charge/discharge management was present in the KA2 (in the form of a few diodes and resistors - I've taken one apart to inspect it). 

 

The stated limit on charge in the Zimo manual makes little sense to me.  I can see how rate of current flow causes a problem with heating, but not total charge stored.   However, as I don't use those pins, its just theory...

 

CV's changes.   I usually just change the setting for Zimo's to ignore DC running.   That's CV29 as discussed.  Possibly relevant is CV153 - time to run without a valid DCC signal.     It sounds like the loco is not seeing a regular clean DCC signal, so are there other devices on the layout which could be an issue ?  (Or equivalent question, do the lack of control occur if the DCC system is only connected to a 1-yard of plain track ? ).

 

 

- Nigel

 

[BromsMods]

In that case I have a feeling that the first paragraph relates specifically to TCS decoders as CVs 250-252 are probably used by different decoder manufacturers for different purposes.

 

Was the Zimo decoder not supplied with a capacitor?

[Nigelcliffe]

CV250-252 is maker specific,  so looking up CV's provided for TCS decoders in a Zimo decoder will produce nonsense.   In Zimo they're part of identifying decoder hardware and decoder serial number.  

 

I've no idea what ships with an MX634.   However a tiny electrolytic capacitor - say few hundred to 1000uF, will have a tiny very hard to discern effect.      A KA2 is circa 140,000uF.   Run time is roughly proportional to capacitance. 

 

But, lack of stopping with a 3300uF capacitor indicates the decoder is on power, but not "hearing" a DCC signal.    That means either some track interference, or damage/fault at the decoder. 

 

 

[Torper]

It's really rather odd because it's not as though the KA2 didn't work.  It did - but then went completely over the top.

 

To go into rather more detail, I have a kit-built 0-6-0 loco fitted with the MX634.  Being kit built there are no complicated electronics such as suppressors, etc.  The loco is fitted with tender pickups on all 6 wheels and I decided these were in need of refurbishment, which meant that I ended up basically rebuilding the tender chassis.  All was well and it ran fine.
However, I then remembered that I had a KA2 stay-alive in my spares box and as there was room for it in the tender I thought I might as well fit it, connecting it to the blue and grey wires on the Zimo - they're specifically meant for a capacitor.  I decided to test this set up before fitting it to the loco so I put the tender chassis with decoder fitted on the rails (all this was done on my stand-alone programming track) connected up to a rather old can motor and the KA2, both lying at the side.  And initially it seemed to work.  Much to my gratification, when I lifted the tender chassis off the rails the motor continued to run just as it should.  But then when I put the chassis back on the rails the motor speed suddenly increased.  I reduced the speed on my Powercab controller, but the motor ignored this.  I reduced the speed on the controller to zero, but the motor continued to ignore this and the only way I could get it to stop was to turn the power off altogether, whereupon the motor gradually came to a stop.  However, when I turned the power back on, with the controller still at zero, the motor started up again and continued to run, drawing 0.4 amps, until I switched the power back off.

OK, must be the KA2 I thought, so I removed it and all was well again.  Put it back, and once again the motor became a law unto itself.  So is it the KA2 or the decoder?  As it happened I had a spare 3300uf capacitor lying around, so I connected that up to the Zimo and basically the same thing happened as it had with the KA2 fitted - the motor and decoder forgot all about the controller and the only thing that would stop it was to turn the power off altogether.  The Zimo came with a tiny 220uf capacitor - I've never tried that as I've always felt that its effect would be negligeable.

 

The loco runs well without the KA2, so that's probably how it will now remain.  But I would like to know why things happened as they did and whether the KA2 can possibly be used on another loco yet to be built.

 

DT