Virgin TTS Becomes Dead When Swapping Controllers

[RDS]

I am currently investigating a 'situation' where one of my Virgin EC TTS units 'loses control' - becoming dead - but when I swap from my normal Roco Multimaus to a Roco Maus2 as 'Master' - the loco is okay, and can also be programmed (a slightly different issue) -  BUT when left overnight - and the Multimaus is used again, the loco responds correctly!  This has suggestions of being a borderline timing issue and reminds me of problems with charged capacitors - and waiting for them to leak away ... but it will take me several days to check through all combinations of my TTS locos and controller types and look at the waveforms on a scope to see if I can identify the cause.   IT may not have been helped by placing on the track when power was on - as temporary shorts may have 'glitched the decoder' - but this doesn't appear to explain why it immediatley has no problem if I change which controller is 'Master' (ie provides the timing). So after rechecking all the pickups are okay and reaching the decoder, try leaving it for a day or two, before re-trying?

[96RAF]

Do you have access to another DCC controller and if so to what effect on the offending loco.

Rob

[Phil_Spiegel.]

As promised last week, I said I would report back on having tested the locos, and other TTS fitted locos, on more of my range of controllers. In addition to the multiple versions of both the Roco Multimaus (s/w 1.02 and v1.04), and Roco Maus2 using different examples of the Roco Amplifier 10764 (The ouput power stage), and The Roco MultiCentralePro [which is its own 'Master'] used previously, - testing has now included the Roco Z21 (with Railcon cutout), and even a Hornby Select 1.03 with v3 hardware  I still have 2 other makes I can test them all on!.

On all the frequency / waveform period of the NmraDCC signal was at the specified centre value of 8.62kHz / 0.116ms and all the Roco waveforms displayed 'perfect' waveforms with anticipated slope, and a minimal overshoot.  From the Select waveform, with no load on the track, it is surprising that many decoders manage to read the signal all ! ... the overshoot ringing is almost 50% in initial amplitude - thankfully dying away within about 0.050ms - and when a load is present on the track, the ringing is noticeably reduced - and with just  the 2 x TTS Sound decoders (2 HST Power cars) is only about 10%... becoming quite acceptable.  [ Which reinforces why I do not believe in adding 'snubbers' or other components, as when the track is operated normally it has moving loads, which make the situation quite dynamic. ]**

When testing today, all 3 Virgin HSTs  worked correctly on all systems - although there was a possiblility that when speed steps were set to 14 or 28, on the Select and Multimaus, that the decoder(s) became more reluctant to respond to function commands - but this was not consisitant - in repeated testing this seemed to have no effect with the Multimaus. 

The complete and repeatable 'latchout' experienced last week (and a previous ocasion in February) which was clearly matching with the Multimaus, but never a problem with  the Maus2, has not yet reappeared. 

I will endeavour to test with both the Sprog2 and Dimax controllers later this week, if time is available: but for the moment it appears as if the problem was a 'glitch' resulting in a latch-up which persists for several days before recovering naturally ... suggesting capactitor charging related. [Past experience with the hornby 'Basic' decoder has shown it to lose CV programming with a 'glitch' on track].  I will also attach the NCE dcc reader to monitor ALL of the data being cycled at the time to see if repeat rates etc are part of the problem  [Differing controllers have different numbers of locos in their 'stack' which are cycled through on the tracl/dcc waveform - cycle duration, and where a controller inserts function commands etc might be an unknown variable here - but until the problem chooses to resurface, I cannot follow it up further 8-(.

**30+ years ago, with my Zero-1, and 30 accessory decoders; I did have to modify the last few accessory decoders to maintain the waveform shape .... but then it was probably only designed for a maximum of 25 8-) 

This DCC scenario is well within 'spec': a test track of 700mm, and optionally an accessory decoder connected (currently disconnected). A 100MHz  'scope monitoring the waveform,and a Voltage Module monitoring the Trackdcc value.  All controllers running from their own SMPS. In february it was 2 simple loops of track on a pair-of-doors.

Now its time to go outside and vacuum the garden to find the railway track (G Scale) out there..... 8-)

[96RAF]

Just spent 20 minutes answering your post Phil and it got lost as I was logged out by the system AGAIN. Unfortunately I did not copy it before posting so it is lost as I can neither recall the detail nor have the motivation to attempt to repost.

 

It is good data though Phil and confirms Hornby claim that their kit does ring off load as it is tuned to meet on load values including European EN55014 EMC requirements in addition to NMRA regs. On load means having locos on track, not necessarily running. 

 

As we know with any off load readings they can be unrealistic.

Rob