[Richard Johnson]

QUOTE (Geoff Booth @ 21 Jul 2008, 06:41) <{POST_SNAPBACK}>Hi Tony,

I will give it another check over but haven't put the body back on as this started once the body was off; that came off routinely without any problems.

Regards

Geoff

Hello Geoff

Reset decoder set CV30 to 2

Check that both the decoder and the controller are set to 28/128 speed steps. The light problem + lurching can often be related to a mismatch of speed steps - usually the controller has accidentally been changed to 14 steps when this happens, but it can be decoder side too....

also check:

100% of suppression parts (capacitors and chokes) are removed.

Reset decoder CV30 to 2

re-check and let us know

regards

Richard

 

[Richard Johnson]

QUOTE (Sol @ 21 Jul 2008, 14:59) <{POST_SNAPBACK}>Afternoon,Richard, I just noticed that you recommend not only removing capacitors but chokes ( what I called inductors) as well.
After my recent experiences with NCE decoders, I must agree - I had upto now only removed capacitors but will be getting rid of the chokes as well.

Hi Sol

Yes, I used to simply remove caps but recently Bachmann add so many odd bits including more than one capacitor - sometimes with them hidden under the motor, that I now just remove everything from the brushes and hard wire almost all loco's... its worth it to not hit the odd problem loco, and I can always guarantee a perfect runner for my customers - and myself too!

Interesting that in an experiment recently, Martin71 tested with an without suppression/caps for both ESU and TCS, and found the effect on the ESU was much greater than on the TCS.... I'd have expected different, but its yet another reason to avoid them!

Not easy for many modellers with increasing use of surface mount components though

Richard
DCCconcepts

 

[Richard Johnson]

QUOTE (Geoff Booth @ 21 Jul 2008, 16:00) <{POST_SNAPBACK}>Hi Richard

Done another reset, problem remains. Checked that the speed step is at 28 on CV29 and the hand controller shows a max of 28 so that appears to match. No suppression parts on the circuit board, they have been removed.

I got it to stop doing this at one point when one of the bulbs was shifted in its housing but the underlying problem remains. BTW I am using a LZV100 and LH100 both version 3.6.

Regards

Geoff

Hi Geoff

Hmmm.

If its only 1 loco then I really don't think is a command station issue.

OK... I'm presuming being N scale the bulbs are using white/yellow but no blue wire/grounding thru the chassis. I have no easy answer so lets eliminate what we can....

As an experiment, take out both bulbs.... does this fix it running-wise?

If it does then it may be a bulb with a semi faulty filament creating an occasional intermittent resistance from rail voltage to fuction wire - that'll be a "back door" into the decoder as the Fn wire is after all rectification - so if its broken away and contacting occasionally it will be creating all sort of interesting interference as its impedance rapidly changes as it heats/cools and lets go with vibration/movement.

Either way, bulbs can bea nuisance later on so I'd replace both with LEDs using both blue and function wire if U have wiring room.... if not, use white LEDs via function wire with the + leads to chassis and 500ohm to 1k resistors.

There's not much else it can be if the wiring is right....

** recheck wiring just in case for the tiniest strand of copper can create havoc.
** motor - but thats very very unlikely as the symptoms aren't right
**mechanism - ditto - doesn't match symptoms

So

If its not fixed by removing the bulbs it may well be a decoder fault.... try another perhaps

fingers crossed

Richard

 

[Richard Johnson]

QUOTE (TonyDaly @ 21 Jul 2008, 19:14) <{POST_SNAPBACK}>Hi Richard.
Well done. Would the bulbs be causing too much of a drain on the decoder ? or is it simply a bulb fault as Geoff mentioned that with only one bulb working at the start it ran ok. It was when he connected up the second bulb that the problem appeared.

Hi Tony - as always its a process of elimination really.

No, its not strain on the decoder.... Arnolds bulbs ae generally good qquality and quite consistent.

What was probably happening was that one bulb was on the way out, and had an intermittent internal contact at the filament. Because its made of very super thin and floppy metal, the merest vibration may make an intermittent contact then lose it again... so its hard to identify.

Because the filament is really a "resistive coil", it has a little inductance too, especially when intermittent as a coil that turned on and off randomly will generate a small field which will in turn possibly generate a bit of hash as it starts and collapses.... certainly just eh random make/break with a few Ma load via the coil will create interference.

The function wires are connected inside the decoder after all rectification and filtering, so any hash that feeds back will directly affect the smoothness of things in the very core of the decoder, causing all sorts of false readings by the processor - basically, it doesn't know how to interpret them so does silly things.

All in all most install or decoder fixes are just a process of patient elimination - take away the variables and the problem is often then much clearer.

Geoff - I'm very pleased I was able to help, as its a nice wee loco and its very frustrating when such strange things happen!

kind regards

Richard
DCCconcepts

 

[Richard Johnson]

QUOTE (Marek Klimczyk @ 5 Aug 2008, 23:50) <{POST_SNAPBACK}>I did, no change I'm affraid.

Marek

*** The older Fl motors are often lumpy on DCC at slow speeds, and take quite some fiddling to get smooth. They are not very good motors really.

Try these things and see what happens

1/ Set both CV3 and CV4 to 15 for a start....

2/ then set PWM frequency low
CV49 to 1 or 17 (both OK)

3/ then set back EMF
CV53 to about 35, CV54 to about 50 (do one at a time and then check result - if some improvement is seen when you do the CV 53 and 54 changes, go back to them individually and raise/lower by 5 to see what happens then too)

4/ then set forward trim
CV66 to 6

Alternately....

you could also try a simpler approach - setting CV 3 and 4 as above, then also setting the Vmax and Vmid.... that is, set CV 5 to say 150, CV 6 to 90. This will compress the speed steps and smooth out the cogging of the 3 pole a bit, plus give a quite small step change at slow speed - top speed will also be slowed.

No guarantees, but these are the two most likely setting sequences to succeed with.... exact values will vary a bit depending on the loco / decoder - I'm presuming the decoder has full ESU capability by the way.
regards

Richard
DCCconcepts