Loksound 3.5 Micro Keep Alive

[peteski]

Yes, ESU doesn't mention it, but out of curiosity did you try writing some value to CV11 and see if it was retained?  I know, it is unlikely.

If CV11 is not implemented by ESU then I can only assume that the firmware of this decoder is programmed to stop the model when it is no longer receiving DCC packets through the track.

If you still have the KA connected, can you run the model with the headlight turned on and check if when lifted off the track with the KA supplying power, does the headlight remain on, or does it off when the motor stops?

[Chinapig]

This one had me scratching my head.  I've just gone and taken off the Kapton tape and checked my wiring.  It's fine.  The KA charges but fails to discharge.

I've read CV11 and it's zero - that's even after supposedly writing a number to it using my Sprog.

The front light goes out immediately on breaking the power to the loco.

[peteski]

The KA might be discharging but slowly since it only powers the microcontroller circuit.  Can you hook up an ammeter in series with the KA's positive lead?  The ammeter's positive lead to the decoder's +VE and negative to the KA's +VE wire.  the 200mA range (if that is one of the ranges on your meter) should work.

That way you can monitor charging and discharging current.   When you power up the decoder, you will see a momentary higher positive current charging the KA. When you remove track power you can then observe if there is any negative (discharge) current coming out of KA. 

[Chinapig]

Sounds dumb but I'm not sure which settings to put my multimeter on.  Are you able to direct me from the pic, please

[MK]

Ted, the lower quadrant with the big "A" (around 7 o'clock position, OFF being 12 o'clock), turn the indicator to the "200m" setting.

[Chinapig]

Yes, they were the multimeter settings I used previously but the read-out was zero on charging and zero when I terminated the track connection.  I've just tried it again and got the same result.
However, when testing the voltage between the two sides of the KA, it measured about 0.5v before turning the track power on and remained at about 12.5v after turning the track power off.
Maybe something's not working correctly on the Amps side of my multimeter?

[John]

I recommend you watch this video .. you said your KA is home made -- double check your diodes in the KA to make sure they are actually charging the KA ..  a simple check could be to "charge" off your track, then discharge by putting an appropriate LED across the output .. see if it lights up

/>

[Chinapig]

Interesting video, thanks.
I've just tested it with an LED and the LED remains lit for quite a while.

[peteski]

Yes, they were the multimeter settings I used previously but the read-out was zero on charging and zero when I terminated the track connection.  I've just tried it again and got the same result.
However, when testing the voltage between the two sides of the KA, it measured about 0.5v before turning the track power on and remained at about 12.5v after turning the track power off.
Maybe something's not working correctly on the Amps side of my multimeter?

That makes no sense.  I wonder if at some point when you last used the low-ranges of Ammeter (not the 10A range) you accidental hooked the meter up like you were going to measure voltage?  The low-ranges of Ammeter are protected by a fuse.  If you connected it across a voltage source, that fuse likely blew.

If you have a 9V battery and any 12V light bulb used in older locos for headlight (those are usually 40mA bulbs), hook the Ammeter (on 200mA range) in series with the battery and bulb.  If the bulb does not light and meter shows "0", then the fuse is likely blown.  The fuse is usually inside the battery compartment.

[Chinapig]

I borrowed a multimeter and wired it to measure milliamps with the following result.
When the power was turned on it went from zero to 5.2 in a couple of seconds.  When the power was turned off it went straight back to zero. 
I'm not sure what was meant to happen or what this tells me but am interested to know.

[peteski]

So you used another multimeter? That also seems wonky.  5.2 what?  Milli Amps?  That seems like an awfully low current for charging the Super Caps.

What are the capacitor values you used and what is the resistor value?
I just realized that your KA is probably not a true Super Caps based KA, but just a stay-alive circuit using standard tantalum caps rated few hundred micro Farads.  If that is the case, it will not keep the decoder running for more than a fraction of a second.

Peteski

[Chinapig]

Yes, it's a stay-alive built according to the diagram I posted previously.  I've upped the capacitors to 20 of the 220 mf to give it a proper chance. 
As I mentioned before, this particular stay-alive worked well on another loco and left the SA with zero volts when tested with the multimeter set to volts.  After cutting the current on the LS 3.5 Micro the voltage was still about 12.5 volts.
I don't know but perhaps the issue isn't with my SA but where I've wired it up to on the decoder?
It's a bit of a mystery but thanks for your time trying to resolve this issue.

[peteski]

Ok, so with 20 220 uF you have total of 4400 micro-Farads (or since the "micro" symbol "µ" is not on a keyboard, it is usually written as uF).  mf or actually mF (upper case F is a symbol for Farads) would mean milli-Farads and it is1000 times larger than uF.  I just mentioned that because I feel that it makes sense in electronics to use proper terminology.

Anyway, while 4400 uF is quite a bit more than the typical (<1000 uF) stay-alive caps modelers add in N scale models, a typical small KA has capacitance of 200,000 uF.

But if that exact KA seems to work in other decoders then we have to assume that it is working properly.

From what I can tell, you are connecting the SA in the correct way.  So just to make sure I understand, you are connecting the positive wire of your SA to the decoder's blue wire, and negative wire to the diode's anode (as indicated by a green rectangle in the diagram I posted)?

Then, when you power the decoder/SA combo the decoder works as designed (making sound, powering motor  and functions), and the SA gets charged (there is 12V across the capacitors)?

When you remove the DCC power form the decoder and measure the voltage on the decoder at the points the SA is connected to, you still see 12V?

If that is the case, when you check again after 15 or 30 minutes with the decoder still powered only by the SA), does the voltage bleed off to a lover value, or stays at 12V?

[Chinapig]

Thanks for the explanation; you've just about summed it up.
When the loco is powered it reads about 14v across the capacitor terminals.  When the power is removed this drops to 8.0v, 3.5v and 0.7v after 15 seconds, 30 seconds and 60 seconds respectively.

[peteski]

Ah ok, that sounds more like what I expected.  So the SA pack works as designed (charges when powered by the decoder, and gets slowly discharged by the decoder when the track power is removed). That would be expected as even a stopped (idling) decoder's electronics still consume power.

The only problem is that the decoder is acting as if it lost power (basically stops functioning).    It is possible that as I mentioned, it is designed to stop running when is is no longer receiving DCC packets from the rail (even when the SA circuit provides enough power for it to run for a bit).