Dapol Class 68 jerky on dcc

[phil Roberts]

It's my third class 68 but is very jerky mid range. I've tried to approach this problem methodically so here goes. It's runs problem free on DC. Have a Lenz and Dapol decoder spare and both behave the same. Slow and full speed are ok it's only mid range. I know it's something to do with back emf but not sure how to modify or sample this on either of these two decoders. I transferred an ESU with sound from my other 68 and that worked fine. So any advice on how to resolve EMF on either of these two chips would be welcome.

[Dagworth]

Has the 68 still got suppression capacitors in place?

 

Andi

[phil Roberts]

Andi,  no idea on that sorry. It does work ok with the borrowed ESU sound decoder though so I guess proving the loco is not at fault. 

[Dagworth]

Some loco/decoder combinations do work with caps in place, a different decoder i the same loco may cause issues. Probably worth snipping the caps out if you can, it may well solve the issue and certainly won't make it worse.

 

Andi

[34theletterbetweenB&D]

Have you tried the 'problem' decoders by installation on the 68 you removed the ESU sound decoder from?

 

I have seen similar effects over the years at decoder installation on all sorts of mechanisms. The root cause is a simple (and crude) tuned network of the motor, and any suppression components (capacitors, inductors) interacting with the circa 30kHz output of the decoder. The suppression components are very low spec. indeed and can cause the type of interaction you describe, seen it myself. Worst cases were mechanism and decoder combinations which at a critical speed just 'locked' uncontrollable, could only be stopped by removing track power. Placing an ear close to the moving loco a 'singing' sound could be heard, some kind of resonant effect. Chopping out the redundant suppression components was an instant fix  in every case. Many like myself do this routinely at decoder installation: decoders have much more effective suppression capability than what most RTR manufacturers install on their mechanisms.

[phil Roberts]

Thank you all for your comments. The general consensus seems to point towards the suppression components which I’m glad to admit is all new to my decoder knowledge so I’ve not missed the obvious. Being low spec I guess two cards by the same manufacturer could give different results. Now the big question (and I apologise if it seems I’m being spoon fed) does anybody know where these components are in a class 68?
I’ll also try the suggestion to install the suspect decoders in the good 68 just to prove the theory above.

thank you again

[phil Roberts]

Update....... with the above comments on suppression components In mind I fitted the suspect Lenz decoder into the existing old 68 and it was still jerky but no where near as bad and the Dapol decoder which had similar jerky characteristics in the New 68 actually came from the original 68 so we know it runs fine in that. It’s looking like the theory above is the answer, finding the right decoder for the right machine. Alternatively remove the suppression components as suggested. I’ve attached photo showing with purely a wild guess what might be the components in question. Can anyone confirm? 

 

 

 

[34theletterbetweenB&D]

That board is cryptic to the extent I wouldn't care to advise without looking at both sides and drawing the paths to see what those circled components are intended to do.

 

If you are confident with a soldering iron you could disconnect one terminal on each from the board, and then test for all functionality, knowing that an easy restoration is possible.

 

But see the two pale brown surface mount components between the components you have circled? Those are capacitors as they are C1 and C2, and they might equally well be suppression components, but I don't know as their circuit placement and function is unclear. They could be equally well be present as part of a circuit to provide a flickering firebox glow for all I know...

 

 

[97403_Ixion]

Hi Phil Roberts,

 

It's difficult to tell the exact shape of those black components but they look slightly rounded at the ends in which case, they look more like diodes to me (also see below).  Generally tantalum capacitors have flat end terminals but they look more like active devices (e.g. diodes) as the leads appear to extend out of the case... though as I say, it's a bit tricky to say from the photo.

 

If they are diodes, then my guess is they may perform some kind of rectification, perhaps along with the brown surface mount capacitor, to eliminate some flickering on the lights?

It is also possible they are what are called Zener-diodes and may be used to clamp the voltage of part of the circuit to a lower level - in which case removing them may prevent this from happening and could cause damage... perhaps if linked to the LED lights.

 

It is also possible they are mini-inductors (the right hand one looks like it may have 'L1' in white underneath it).  L1 would certainly indicate an inductor - i.e. a coil, in which case they may be the suppression coils mentioned previously.

 

Personally, with the possibility they are Zener diodes, I would be cautious about removing them, unless I could be sure they were inductors.

 

When I setup DCC in a Hornby 121, I found it was rather jerky.  The decoder had back-emf, so I tried modifying the settings for the feedback, after which I got it running well.  It's not an easy thing to do but tweaking values is probably not likely to cause damage.  I've not read the entire thread, so ignore that if you have already done so...

 

By all means get some more close up photos around the black components... I realise you added the arrows to highlight the components but if you have a non-annotated photo, it may be easier to tell from the PCB tracks what's going on.

 

Happy to help, if I can!

Cheers,

Ixion.

[markw]

The suppression components are the two black inductors (L1 and L2) highlighted and the capacitor C1.

If you remove the inductors you need to replace with a wire link in their place.

Capacitor C2 is part of the lighting circuit.

[phil Roberts]

Hi Ixion. Have attached a cleaned up copy of the cct board As requested. As you say guessing which components do what is not really best way forward. I’ve spent some considerable time unsuccessfully searching the internet to see if anyone else has experienced this and also if anyone knows which are the capacitors in question. I’ve already ordered a new Dapol 21 pin in the hope we get a match, if not I have a choice of playing around with BEMF parameters Of either the Dapol or Lenz neither of which fill me with hope as both have limited information available. Thank you

[phil Roberts]

Hi Mark, posts must have passed mid flight . Thank you for confirming which components are which especially when mine was a guess. At least I have a fall back plan now should the new Dapol 21 pin (expected Tuesday) fail to match. What I have gained from these discussions is the realisation that removal of the suppression capacitors is common practice. Had no idea. 

[Dagworth]

I would expect that if you take the circuit board out you will find capacitors soldered around the motor itself as well as what's on the board.

 

Andi

[Izzy]

I’m puzzled by the fact the problem decoders work fine in the original 68, but not in the latest. Does this mean from the suggestions to touch components on the pcb that it is presumed the specs or parts have changed? I am also a bit surprised that another Dapol decoder is being obtained. I don’t see how it will be any different and would have thought getting a Zimo would be a better and no dearer option (MX638 @£20). But then I am biased towards using them for their superior motor control.

 

Izzy

[markw]

2 hours ago, Dagworth said:

I would expect that if you take the circuit board out you will find capacitors soldered around the motor itself as well as what's on the board.

 

Andi

No, the capacitor C1 on the top of circuit board is the only one fitted.

[34theletterbetweenB&D]

12 hours ago, Izzy said:

I’m puzzled by the fact the problem decoders work fine in the original 68, but not in the latest. Does this mean from the suggestions to touch components on the pcb that it is presumed the specs or parts have changed? 

Probably no deliberate change in spec or design, just the effect of the particular mix of components each model happened to get. That's just my opinion, but the reasoning is as follows. The electronic components used for the suppression circuit will be very cheap, and vary widely from their nominal values. (Do manufacturers actually review whether the suppression circuit design is a good match to the motor design used on that particular model, or do they just do what they have always done?)

 

Now, that is normal: there's variation in all manufactured product, and we can specify tight control of this at a price, or 'near enough will do' for a saving..

 

The motor ; I have had plenty of multiple examples of good motors in RTR OO which have shown a range of performance: all of it acceptable interms of how it makes the loco move on track, but easily seen to not be uniform with each other.

 

And there will be variation in decoder performance, when multiple examples of the same design are compared: this however really has to be looked for, these are manufactured under good control to achieve very uniform performance.

 

Personally, when I run into such troubles on a mechanism when a Lenz or Zimo decoder is installed (I have a high degree of confidence in these brands' manufacturing process control) I don't worry about it, provided the motor has run well on my acceptance test using 'vanilla DC' from an old resistance type controller. Just rip out the board, remove any components on the motor, and hardwire the decoder between pick ups and motor terminals. Never yet had a performance problem on track after that. A good decoder will outperform the RTR manufacturer's installed board every time thus far in my 15 years experience of DCC on a good range of OO and HO mechanisms.

[Izzy]

47 minutes ago, 34theletterbetweenB&D said:

Probably no deliberate change in spec or design, just the effect of the particular mix of components each model happened to get. That's just my opinion, but the reasoning is as follows. The electronic components used for the suppression circuit will be very cheap, and vary widely from their nominal values. (Do manufacturers actually review whether the suppression circuit design is a good match to the motor design used on that particular model, or do they just do what they have always done?)

 

Now, that is normal: there's variation in all manufactured product, and we can specify tight control of this at a price, or 'near enough will do' for a saving..

 

The motor ; I have had plenty of multiple examples of good motors in RTR OO which have shown a range of performance: all of it acceptable interms of how it makes the loco move on track, but easily seen to not be uniform with each other.

 

And there will be variation in decoder performance, when multiple examples of the same design are compared: this however really has to be looked for, these are manufactured under good control to achieve very uniform performance.

 

Personally, when I run into such troubles on a mechanism when a Lenz or Zimo decoder is installed (I have a high degree of confidence in these brands' manufacturing process control) I don't worry about it, provided the motor has run well on my acceptance test using 'vanilla DC' from an old resistance type controller. Just rip out the board, remove any components on the motor, and hardwire the decoder between pick ups and motor terminals. Never yet had a performance problem on track after that. A good decoder will outperform the RTR manufacturer's installed board every time thus far in my 15 years experience of DCC on a good range of OO and HO mechanisms.

 

Thanks, that's very interesting to read. With apologies to the OP if this takes things a bit 'off-track' but I have found over recent years that pcb's in DCC ready locos with no lighting circuits need removing completely otherwise decent running just isn't possible, but strangely, if any lighting circuits are involved then things are just fine. This covers both 4mm and N gauge. I have absolutely no idea why this should be.

 

Izzy

[martin_l_jones]

The only caution I would express is modifying the PCB is that it would invalidate the warranty. I have quite a few 68s and fortunately have not suffered any erratic running. Mine all have Dapol factory fitted decoders or I have fitted Imperium decoders which were preprogrammed for the 68’s....

[34theletterbetweenB&D]

7 minutes ago, Izzy said:

...I have found over recent years that pcb's in DCC ready locos with no lighting circuits need removing completely otherwise decent running just isn't possible, but strangely, if any lighting circuits are involved then things are just fine...

About 50/50 for me, RTR OO steam models in the main so no lighting. Some lose their PCB and socket because the manufacturer has positioned it ineptly, some because the running performance isn't as good as the DC suggested should be possible; and sure enough with the decoder doing its work 'unassisted' the expected refined running is produced. Especially noticeable at transition in and out of movement and running at dead slow, a hint of what I think of as 'granularity' is eliminated.

 

Happily, my sole Dapol loco, a class 21 with a Zimo, is creamy smooth and no work was required to make it so.

 

3 minutes ago, martin_l_jones said:

The only caution I would express is modifying the PCB is that it would invalidate the warranty...

I give all my new purchases an immediate extended run test to find out any infant mortality which might require exercise of the warranty. After that they are MINE, and get whatever treatment is required to make them work right...

[martin_l_jones]

22 hours ago, 34theletterbetweenB&D said:

I give all my new purchases an immediate extended run test to find out any infant mortality which might require exercise of the warranty. After that they are MINE, and get whatever treatment is required to make them work right...

 

I'm not questioning anything you may want to do. I was making him aware of the fact that should any issues arise making modifications will invalidate the warranty. In reality if it's a poor runner the general consensus on this forum is to send it back.

[phil Roberts]

Update..... the new Dapol 21 Pin decoder arrived today and was duly installed with.... 99% success. Mid range is now a lot smoother so apparently there is a difference between two decoders of the same make. Both v4. However I say 99% because it is still slightly jerky and I mean slightly, so decision time. Either break the warranty and remove the suppression components or live with it. I’ve  decided to live with it for the time being as it’s not so noticeable now with a five coach loading. Knowing me though I won’t be happy until it’s perfect. 
thank you all for sharing your knowledge and experience, it’s been an enjoyable learning curve.
Phil

[boxbrownie]

On 24/05/2020 at 20:33, Dagworth said:

I would expect that if you take the circuit board out you will find capacitors soldered around the motor itself as well as what's on the board.

 

Andi

It is interesting you say that as I had a similar issue with a Hornby Loco and posted a picture of the circuit board to ask which of the components were suppression caps, no one seemed to know exactly but I did get an answer from Hornby technical after a week or so and they state the board is the controlling board, the motor has its own suppression capacitors fitted, which up in closer inspection proved to be correct, and removing those helped considerably.

 

I guess from reading above Dapol do it differently.

[TomScrut]

Have you turned back EMF off on the decoders in question? I presume they work OK then?

 

If so then either do the cap mod or tweaks some CVs. All my stuff has ESU decoders fitted with caps still in place and I haven't ever had issues like this, although on stuff I am going to keep I am going to at some point take the caps out.

[Les1952]

Related to this I have a few Continental locos that have taken a dislike to decoders, so it isn't just a Dapol problem.  In one case the only decoder that will physically go in the model will not perform, yet performs well enough in others.  In another case no decoder of any type will work in the model.

 

 With each separate component being made to a tolerance the variation becomes wider and wider as you put individual components together.  If you are unlucky enough to have all the variations running the same way the final board can be a VERY long way off tolerance.

 

As for testing I asked a manufacturer of circuit boards a long time ago why they didn't supply a bare bones board.  The answer was that to do that they would have to mount a set of components on it, test it, then unsolder the components again.  Even where labour is cheap that is just not going to happen.

 

Les

[TomScrut]

9 minutes ago, Les1952 said:

why they didn't supply a bare bones board

 

I have often wondered that, I'd be keen to have a 21 pin board, with tracks to pads for each pin. On each track 2 holes to fit a resistor, which if you want to do that run a knife through the track.

 

Would mean one could whip the board out of their locos to put back in in case of resale, and also give an easier way of wiring a highly functioned decoder into a loco.