Bachmann Class 37 decoder fitting problem

[highpeakman]

I purchased the Bachmann class 37/0 (D6801 green) a few months back. It was DCC ready with a 21 pin socket. The loco ran smoothly on DC with the cab lighting, front and rear running lights working perfectly and directionally.

 

I have just fitted a Bachmann 36-554 21 pin decoder as I figured that should be a straightforward fit without problems.

 

I did nothing except remove the body carefully, remove the blanking plug and then plug in the brand new decoder. I should say at this stage that I have previously fitted decoders to about a dozen locos including older models and split chassis types without problems. I have also fitted sound decoders in two locos without problems. I use a NCE Power Cab but also make use of a SPROG on a specially built test rig.

 

Since fitting the decoder and refitting the body the 37 runs basically correctly on both NCE and SPROG. The problem is that the running lights do not work anymore in either direction. Other Functions seem OK i.e F1 turns on the cab lights on/off, F3 switches to shunting mode OK. It seems from the instructions that the lights should work against F0 (or F10) but they do not react to these commands.

 

[No attempts have been made to alter any CV settings (yet).]

 

I note that on this model the LED boards for the running lights (fitted the the body) connect to the chassis via spring connectors (unlike earlier 37s). These seem OK but may not be connecting properly although I was really careful when refitting the body.

 

I have carefully checked the Bachmann site for clues and searched this forum but cannot see any clues. All the fitting guides I can find refer to earlier editions of the loco that do not have the 21 pin decoder or spring connections. I note that on one Bachmann instruction sheet it mentions about "reversing" the decoder (it is fitted the right way up according to instructions) but I am wary about doing this in case of more problems (besides the other funcions appear OK).

 

I feel I may be missing the obvious but has anybody any suggestions please before I remove the body again? Thanks

[DIRTY DIESEL]

Hi

,

Just a thought as iv did something similar with one of my class 37's but there should be two tiny screws (one each end) at the nose end under the coupling, I left these off and had the same problems, don't think the body makes proper contact if screws are not screwed in

 

cheers

 

Stu

[highpeakman]

Thanks Stu

 

I think you are right as I have now got to the seat of the problem. After removing the body again it appears that the problem stems from the spring contacts on the LED boards that house the red lights and route indicator panel lights. These contacts press down onto pads on the chassis. It would seem that the contacts in this case did not make contact again when the body was refitted. Surprisingly this happened at both ends of the loco and yet considerable care was taken during the process of removal and refitting. Careful bending of these contacts to apply more force appears to have cured the problem but it does seem to me that they must be very sensitive and vunerable to have lost contact in this manner.

 

Stu I guess that your adjustment of the screws just gace the extra pressure necessary so it sounds they they are only just contacting.

 

So if anyone is removing this body then please be careful on this later type of 37 body.

 

Don

[DIRTY DIESEL]

No problem Don, glad you got it sorted.

 

Stu

[highpeakman]

Just to clarify the above. The body fixing screws had been fitted and tightened originally when the problem was first experienced.

 

Body fully refitted now and all is working as it should but the contacts had to be bent quite a lot.

 

The other problem with this particular model was that it is swimming in oil. Almost dripping off the motor unit and getting all over the bogies and body sides during the fitting of decoder. Way overdone at trhe factory. Even after a lot of wiping down it should not need oiling again for years!

[highpeakman]

Problem NOT solved after all!

 

When using Power Cab the lights work correctly at some speed settings but go off at others! As you increase the speed (in forward or reverse) the lights (at both ends) go off at certain speeds then come on again then go off again and on again. It is speed (throttle setting) related not a function of time - if you leave loco running at same speed then the lights do not change - only when speed is changed then they change and maintain state while speed is set. Obviously it is not a function of the spring contacts this time. Not sure where I am going with this one now!

 

 

 

 

On SPROG Throttle I cannot make the lights work at all and yet the F1 function (cab lights) and F3 (shunt speed) work perfectly.

 

Any ideas please? Anything to do with the Bachmann decoder? I have not suffered such problems previously with other makes.

 

Don

[Nigelcliffe]

Problem NOT solved!

 

When using Power Cab the lights work correctly at some speed settings but go off at others! As you increase the speed (in forward or reverse) the lights (at both ends) go off at certain speeds then come on again then go off again and on again. It is speed (throttle setting) related not a function of time - if you leave loco running at same speed then the lights do not change - only when speed is changed then they change and maintain state while speed is set. Obviously is not a function of the spring contacts. Not sure where I am going with this one!

 

Don

 

First thing I'd check here is the setting in CV29. I'd be suspecting the chip set to 14 step and the throttle set to 28/128 step mode.

 

- Nigel

[highpeakman]

OK. I'll have a look at that - Thanks.

 

I had just put the DC board back in and ran the model. Lights and everything run perfectly at all throttle settings.

 

However I just looked at the SPROG readings of the chip and you are absolutely correct - chip settings are at 14 step. I missed that before but am a bit surprised if that is how it comes in the box.

 

I will now refit the DCC chip (again!) and reset CV29. It will have to wait a couple of hours now before I can do that but Thank you very much for that suggestion Nigel.

 

.

Don

[highpeakman]

Refitted decoder and set CV29 as Nigel suggested and now running lights work as normal. Curiously they are now switched on/off by Function 0 whereas before they worked part time as described above but with no control from the Function 0 switch. I am not sure I understand why the speed step function CV29 setting should affect the way the lights operate. Oh well I live and learn!

 

Having reassembled body again and alll the running lights working etc I now note that the cab end lights at No 1 end have ceased to work - a function I am sure of the spring connectors (different ones to the running lights) for those. It seems to me that the spring material used by Bachmann just does not have suficient "spring" in it and it seems that in practice there is a need to reset all of these contacts every time the body is removed.

 

Thanks for the help.

 

Don

[giz]

When I've fitted decoders to Bachmannn 37s it always seems to take at least four attempts at fitting the body before I can get all the lights to work.

 

The end screws are also a pain, too tight seems to be as bad as too loose when making the connection.

[Nigelcliffe]

Refitted decoder and set CV29 as Nigel suggested and now running lights work as normal. Curiously they are now switched on/off by Function 0 whereas before they worked part time as described above but with no control from the Function 0 switch. I am not sure I understand why the speed step function CV29 setting should affect the way the lights operate.

 

Well, you did ask !!!

 

Simple stuff, with no technicalities as to why, see an article I wrote ages ago: http://www.2mm.org.uk/articles/cv29%20calculator.htm

 

 

Or the more complex "why". Within the NMRA specification for DCC are two primary settings for speed steps; 14 step and 28/128 step. The throttle and the decoder need to be set the same for things to work correctly. Now, if there is a mis-match, as in your loco above...

 

The data packet sent over the rails to the loco contains a sequence of bits. Those include the loco address and the instructions to the loco. The instructions may be speed or some other commands. If we look at the structure of a speed packet we find the following... If sending a 28 step packet, the fourth bit in the packet is used for speed. If sending a 14 step packet, the fourth bit in the packet is used to control the front light output (FL). This tells us that the same bit in the instruction on the track can be interpretted differently depending on the setting in the decoder in CV29 for speed step.

So, if you have a controller sending 28 step instructions to a decoder set for 14 step, and slowly accelerate the controller, the decoder will receive instructions which it interprets as "lights on", followed by a small speed change and "lights off", then "lights on" again, and the next is another small speed change and "lights off".

 

 

references for this stuff:

http://www.nmra.org/standards/DCC/standards_rps/rp921.html#Speed%20and%20Direction%20Instructions

http://www.nmra.org/standards/DCC/standards_rps/S-92-2004-07.pdf

 

 

 

- Nigel

[highpeakman]

Nigel

 

As you say, I did ask!

 

Actually I do understand your explanation now although I had not previously realised the relationship between these particular packets. Just a constant process of learning I guess. However, to be honest, getting into the depths of the codes and packets is not my forte even after a lifetime in electronics (it was microwave - high frequency stuff and I only had a brief experience with networks late in my career).

 

While I may have set speed steps to 128 previously when I have installed decoders (usually Digitrax DZ125) I do not recall doing so but I note that all my other locos are set at 28/128. Given the differences in performance I would assume that 28/128 is the most frequently used so I a bit puzzled as to why the Bachmann decoder came set to 14 - is that normal?

 

Also, with the Bachmann 36-554 is there a problem using the speed table settings on the SPROG? The Bachmann data sheet only mentions the 2 CVs for start and max voltage - it does not mention even the mid voltage setting so is there a limit on what you can set. I have a G2A 0-8-0 that came with this decoder and initially had some problems which seem to result from trying to set these values. I have got it back to behaving normally but am wary of trying to set the table again.

 

Thank you very much for your help and the references which I have downloaded for browsing - later.

 

Don

[Nigelcliffe]

Nigel

 

As you say, I did ask!

 

Actually I do understand your explanation now although I had not previously realised the relationship between these particular packets. Just a constant process of learning I guess. However, to be honest, getting into the depths of the codes and packets is not my forte even after a lifetime in electronics (it was microwave - high frequency stuff and I only had a brief experience with networks late in my career).

 

While I may have set speed steps to 128 previously when I have installed decoders (usually Digitrax DZ125) I do not recall doing so but I note that all my other locos are set at 28/128. Given the differences in performance I would assume that 28/128 is the most frequently used so I a bit puzzled as to why the Bachmann decoder came set to 14 - is that normal?

 

I am surprised that a chip came set to 14 step, that is not normal. Default setting on any new chip should be 28/128.

The 36-554 is an antique chip design by ESU, bought cheaply and rebadged by Bachmann.

 

 

14-step is obselete for nearly all users, so don't use it.

 

 

 

Also, with the Bachmann 36-554 is there a problem using the speed table settings on the SPROG? The Bachmann data sheet only mentions the 2 CVs for start and max voltage - it does not mention even the mid voltage setting so is there a limit on what you can set. I have a G2A 0-8-0 that came with this decoder and initially had some problems which seem to result from trying to set these values. I have got it back to behaving normally but am wary of trying to set the table again.

 

 

For automatic decoder identification in JMRI, the software reads the values in CV7 and CV8. These contain the maker identification and a version number for the decoder. But many makers don't set the version number CV to anything useful, so JMRI cannot uniquely identify the decoder and has to pass the decision making back to the user to select from a list.

 

As far as I can see, the Bachmann decoder file is correct in JMRI, it just offers two speed settings (start and max), and no other speed table options. The decoder is listed under ESU, the decoder maker, the Bachmann chip is a rebadged ESU LokPilot V1.

If you selected a different decoder which did have a complex speed table, you would get the option to set CV values which are meaningless to the decoder (or worse, actually set something else!!).

 

 

- Nigel

[highpeakman]

Nigel

I had noted that the JMRI came up with a wide choice of ESU decoders (and I also noted this comment in a previous thread about the merits/otherwise of this chip) so I accepted that it was an ex-ESU chip - the problem was to sort out which one of the selection offered. I had clicked on the Lokpilot 3 which was, with the benefit of your input, the wrong one. Presumably this why JMRI offerred the choice of speed table setting? Should I have forced it to the Bachmann choice or use Lokpilot 1? I note that there seemed to be no choice of different types of Bachmann, just EZ command.

 

To be honest at this moment I am starting to consider replacing this 21 pin decoder as the experience with the 37 and G2a seem to indicate I will not get the flexibility and control I want. The 36-554 is probably a good value chip and will suit many applications but does not appear to offer me the control I prefer. It is basically "pay more to get more". Luckily I have no more waiting to be fitted.

 

BTW I had definitely not set it to 14 steps. I just took it out of the original packing and put it into the 37 which was why I had not checked the setting.

 

Once again Thank you very much for enlightening me. I learn and move on to different mistakes!

 

Don

[Nigelcliffe]

Nigel

I had noted that the JMRI came up with a wide choice of ESU decoders (and I also noted this comment in a previous thread about the merits/otherwise of this chip) so I accepted that it was an ex-ESU chip - the problem was to sort out which one of the selection offered. I had clicked on the Lokpilot 3 which was, with the benefit of your input, the wrong one. Presumably this why JMRI offerred the choice of speed table setting? Should I have forced it to the Bachmann choice or use Lokpilot 1? I note that there seemed to be no choice of different types of Bachmann, just EZ command.

 

 

I don't think there is any difference between the LokPilot1 and the Bachmann files in JMRI (there shouldn't be!!). You'll find the Bachmann chip under "ESU - Bachmann" on the expanding list of decoder types.

 

Yes, using the LokPilot v3 will introduce a number of additional CV options which will not do anything on the version 1 of the chip (almost certainly do nothing, rather than actually doing something else!), and that would include the speed table. Those LokPilot v3 features are, of course, included in the JMRI file for that chip.

 

 

 

ESU are one of the most difficult to identify correctly by JMRI; the maker is found easily but narrowing further is difficult due to lack of reliable CV to question for this information.

In contrast, if you have a Soundtraxx or a Zimo chip, the indentication is much more precise as those makers have additional identification CV values which allow the software to narrow to a very small list of options (in Zimo's case the decoder can usually be precisely identified and the options should be the wiring plugs on the end of the decoder ).

 

 

 

To be honest at this moment I am starting to consider replacing this 21 pin decoder as the experience with the 37 and G2a seem to indicate I will not get the flexibility and control I want. The 36-554 is probably a good value chip and will suit many applications but does not appear to offer me the control I prefer. It is basically "pay more to get more". Luckily I have no more waiting to be fitted.

 

More features are usually found on newer chip designs, along with improvements in the motor control circuits. The current LokPilot v4 is a good chip, as are many others.

 

 

 

- Nigel

[royallancer]

if your lights dont work on f0 after resetting the dcc id cv01 - try this

 

Ive got the answer to fix the lights not working on Bachmann 21 pin chips - when the id is changed using e-link

 

when changing the loco id CV01, e-link changes CV29 as well (a configuration CV) - this is not visible on the e-link screen since it is just over the page

 

sometimes when changing/writing Cv29 this doesnt work properly on a Bachmann 21 pin chip - it either fails or write the wrong id

 

Here is how to fix it

 

Change the loco back to factory defaults - which I think involves rewriting CV08 to 08 in many chips (you should check yours)

 

Not what CV29 is

 

Reset your loco id CV01 using e-link

 

check CV29 - if it has changed then rwrite the original value to it

 

Your lights should now work on F0

 

It worked a treat on my 3 locos

 

Thankyou Hornby Railmaster for the advice.

 

 

Kevin

[Andy7]

Really! 3.5 years!

[royallancer]

Apologies for my typo below - it should say Note CV29

 

 

if your lights dont work on f0 after resetting the dcc id cv01 - try this

Ive got the answer to fix the lights not working on Bachmann 21 pin chips - when the id is changed using e-link

when changing the loco id CV01, e-link changes CV29 as well (a configuration CV) - this is not visible on the e-link screen since it is just over the page

sometimes when changing/writing Cv29 this doesnt work properly on a Bachmann 21 pin chip - it either fails or write the wrong id

Here is how to fix it

Change the loco back to factory defaults - which I think involves rewriting CV08 to 08 in many chips (you should check yours)

Not what CV29 is

Reset your loco id CV01 using e-link

check CV29 - if it has changed then rwrite the original value to it

Your lights should now work on F0

It worked a treat on my 3 locos

Thankyou Hornby Railmaster for the advice.

Kevin