Problem reading Loksound CVs from Bachmann 37

[Will Vale]

Hi all,

 

I've just started playing with DecoderPro and my Powercab, using the NCE USB interface. So far so good, I upgraded the firmware, did a complete read of the Howes ESU chip in my class 60, and changed a few simple CVs to check it all worked. I was also able to read the 21-pin chip in a class 66 with no problems.

 

I then went to fill in the rest of my loco roster, and came across a couple of issues:

 

1) The Howes Loksound Micro in my Hornby 08 isn't recognised - is there a closest match decoder? (I tried reading it as a LokSound 3.5 but the Powercab crashed eventually - possible while reading nonexistent CVs?)

 

2) The SWD Loksound 3.5 in my (21-pin chassis) Bachmann 37 doesn't read at all - it usually produces garbage values (often 255) although once it managed to read the address. I tried a different chip (Bachmann factory 37 sound) with the same result.

 

I tried reading a sample of CVs using the handset and that had the same problem, giving either 255 or Can Not Read CV messages. The motor twitches suggesting a read attempt is happening though. Also writing CVs with ops mode is fine, and the loco runs properly, all functions present and correct.

 

Given that the loco and chip work OK bar reading, and that I can read 8-pin and 21-pin Loksound chips in other locos, it seems like there must be something weird about the 37 chassis which makes it hard to read CVs? Should I be looking to remove capacitors at this point, or is there anything else I can try first?

 

Many thanks in advance for any help!

 

Cheers,

 

Will

[Nigelcliffe]

You've done the obvious which is to prove the problem is with certain locos on the programming track - if the PowerCab handset cannot read the CV's then JMRI/Decoder Pro cannot either (all the software does is the equivalent of pressing buttons on the handset very quickly).

Are you using a short programming track with a direct electrical feed from the PowerCab, or is this the entire layout ? I'd try with a short programming track and a dedicated short electrical feed if not already doing so.

It might be just programming load, sometimes putting a resistor in series or in parallel with the track helps - Google to find suitable values, what is used for series is different to parallel.

 

 

For the LokSound Micro, raise its absence with the JMRI-Users Yahoo group. Someone will add it (from the manuals, a Micro is identical in CV structure to a 3.5 and an XL, so the work required should be minimal to add another decoder family to the existing files).

 

You should be able to use the 3.5 file without problems. However there is a technical issue on command timing with respect to NCE USB devices in a discussion in the JMRI-Users group (I'm not following that thread in detail as I don't run NCE kit). This might be part of what you are seeing, so it may be worth trying it more slowly; read each page one at a time rather than "read all CVs in decoder".

 

- Nigel

[Will Vale]

Thanks Nigel, I was trying to program on the layout plus there's a breaker between the track feed and the track bus which might be having an effect? I'll try and hook up a standalone programming track and see how that goes. Likewise thanks for the tip about page by page reading. I'll post back here when I've got some results to report.

 

 

Cheers,

 

Will

[Will Vale]

Well I've tried a few things but not with much success.

 

1) Created a standalone programming track

2) Tried 1K resistor in parallel, and a 22R resistor in series with the programming track (these seemed to be common "it worked for me" values from my Google searches).

3) Swapped the decoder in another chassis (8 pin with adaptor) which was working fine with its existing Sapphire decoder - still unable to program.

4) Got slightly better results (some CVs read correctly some of the time) setting Direct Byte mode manually in DecoderPro/JMRI.

 

I also realised that the motor pulses for read back are audible binary with a sync. bit at the start of the value. So I now wonder if the problem is the decoder not reporting the value correctly when requested - it was occasionally reporting three as "dit" rather than "dit-dit-dit", which matched up with the zero the Powercab was seeing. So maybe it's a bad decoder after all?

 

I've emailed NCE (who hadn't come across the problem) and SWD (who are on holiday for another week or so I think) so I'll see what comes back from that. In the meantime I can probably figure out the CVs I want to program (function/sound slot remapping) on another loco (where I have read-back values to undo with) and just blindly apply them to the 37, given that it runs and programs fine in ops mode. Do SWD decoders have factory-reset-to-sound-provider-values on CV8 like Howes now do, or will that mess up the sound setup?

 

 

I did manage to get the micro decoder read back properly though - going sheet by sheet seems to work.

 

Cheers,

 

Will

 

[Pete Harvey]

Will

 

You could try contacting DCC Concepts who I believe is on you side of the planet he may be able to help as he has worked a lot with sound locos?

 

It is worth a try?

 

Pete