R3395 'Mallard' TTS, Suddenly No Power!

[Stephen in Kerry]

When I last ran my R3395 Mallard, about 2 weeks ago, it ran fine at all speeds, with sound and no issues at all. Today, it ran for about 3 minutes, juddered on some points (that may have been set the wrong way - but not certain) and lost power.

It got sound back once or twice, but didn't run at all after that, and now it's quite 'dead'. The connecting wires between loco and tender all seem to be intact.

Does anyone know what might be happening?

Hope you're all doing well - and Happy Christmas from the west of Ireland!

Stephen

[P-Henny]

It seems to me from your decription that your TTS Decoder could be damaged.

Let's assume for a moment that you did indeed run the loco against points set the wrong way. When these types of short circuits occur on DCC powered layouts, you can get what is termed 'ringing' on the DCC track voltage. Ringing is a form of positive runaway oscillation which can generate peak spike voltages that exceed two or three times the base DCC voltage. Hornby DCC is 28.8 volts peak to peak, so ringing could potentially create voltages up to 60 to 90 volts peak to peak. These high voltage spikes can damage capacitor components on DCC decoders. The risk of ringing damage can be reduced by fitting a snubber filter across your DCC track power. A snubber will not provide a repair to a decoder that is already affected, it is fitted as an insurance policy to protect DCC Decoders against future short circuit issues arising.

If you navigate to the 'Useful Links' sticky at the top of the 'General Discussion' forum section, then scroll down the sticky to the 'DCC Technical' sub-section, you will find a link to 'RC Filter / Snubber' theory.

[Stephen in Kerry]

@P-Henny Thanks for the reply, I was beginning to think the same thing about the decoder. Can you suggest a remedy, or do I just need a new one?

[P-Henny]

Once a decoder is proven to be damaged, the only solution is replacement. But do review my later edit to my earlier reply regarding future protection insurance policy (snubber).

Sometimes a return of functionality can be obtained by factory resetting the decoder (writing 8 to CV8) but I find that the decoder is still weakened, such that another instance of a short circuit will corrupt the decoder again. I once had to reset a damaged decoder everytime I used it, I had to end up replacing it. Now that I have subsequently fitted a snubber, I have never had another damaged decoder since.

[Brew Man]

@Stephen

Do you not have a spare decoder, or one you can take out of another loco just to check?

[Stephen in Kerry]

I put the loco on my programming track and it worked, at the original cv address. This made me think that the issue was with the track, but not the whole layout, as locos were working on other lines.

I've narrowed the problem down to one of two sets of points, which are next to each other on a straight line. I'm 90% sure that one of them is not creating a circuit, some of the time. At the moment, it's working, and I'll give them both a good clean tomorrow.

Thanks for your input guys.

[96RAF]

@stephen

There is an FAQ in general discussion sticky’s for testing dodgy points.

an overload test for TTS is to see if it will run with the sound off. Sometimes a marginal motor current draw will throw the decoder into self protect when the sound is on.

[ColinB]

I am a bit concerned about the description of "ringing" description on this thread. A short circuit is as it says, it shorts out the signal so it cannot ring. Now "ringing" generally occurs when you have a load which you suddenly remove generally caused by inductive effects. I suppose if you have a large layout you may get transmission line effect. I suspect something more simple like dodgy points as 96RAF says or an issue with your track. Sometimes it is the 4 pin connector not pushed in completely.

[P-Henny]

Ringing can occur when the short circuit is transient in nature and not firmly and directly made as a dead short. Read the theory link I provided before being critical of the comment. The article is written by a member of the DCC Standards commitee. Before I fitted my snubbers I have had transient short circuits where the ringing was distinctly audible in the audible frequency range. When I tested the effect of my snubbers (by applying a transient short before and after fitting) the ringing could be observed on my oscilloscope and reduced when the snubber was fitted.

[96RAF]

A DCC bus can be likened to a mains power transmission line - think electricity pylons carrying hi-voltage over many miles.

There is a standing waveform transiting the power lines which upon reaching the far end becomes reflective and bounces back along the lines in opposition to the regular waveform precession. A suitably designed snubber damps out this reflection. For the general science see here http://ibiblio.org/kuphaldt/socratic/model/mod_tline.pdf

There is a simpler to understand noddy guide animation on line that better demonstrates these effects, See here

[ColinB]

I know 3rd year of my degree in electronics. Effectively you match the load with the source, more applicable to microwaves. That is why I said it probably occurs on large layouts, but if it is sectioned less so. I suppose I could go and get my old notes and work out what it is at 50 kilohertz which I think is the DCC frequency. I know the limit of CAN is about 6 feet. Also DCC is not a fixed frequency. I am sure the snubbers probably do something not sure you really need them if you have decent wiring, then you have the issue of them combined with the capacitance of the ballasted track rounding off the DCC signal. As I said in this case the issue with the loco is probably a simpler fix. My layout is 16 foot by 10 foot in DCC sections and no I don't have any snubbers and virtually all the decoders that die have been due to overheating (Hornby decoders are poor at heat transfer).

On 50 hertz it is miles but they do switch in different lines to match the load, the bigger issue is having no signal in the middle.

[ColinB]

I went and watched the video, thanks 96RAF and it all came back to me, termination resistors. On a CAN link you have to put a termination resistor at the end of the chain so that is what the snubber is probably doing plus a bit more. The thing is, it is not as simple as in the video, as the loco could be anywhere on the track so the transmission line is not the same length all the same time. You can get "ringing" of the square wave signals as the drivers switch on and off, but that is what the capacitors on the decoder are supposed to stop.

Anyway it is a red herring to this query so I suggest it is best ignored.

[rpjallan]

For the price, I don’t know why anyone would not fit “snubbers” to the ends of their bus wires. Just in case! Same as twisting your bus wires. It is a simple thing that will help prevent running problems on your layout & could save decoders.

[Stephen in Kerry]

Nice to see that my post 'sparked' (pun intended) an in-depth debate on advanced electronics.

Unfortunately, my degrees (Master's, Honours and Post-Graduate diplomas) are all in unrelated subjects, so I'm quite lost - and sticking to the basics.

I'm going to assume that the issue I have is caused by the switch rails on the points not making full contact with main rails on the track. Possibly due to debris.

I'll give them a clean today and let you know. :)

[Fishmanoz]

Stephen, please reset your decoder as recommended above as well as confirming your track is ok. Before doing so you might check if it will run on address 03 as sometimes shorts cause decoders to reset and, as P-henny says, they may become more susceptible to this after a short.

And Colin, while it has been a little while since the last of these discussions - transients being under-damped ringing arising from imperfect shorts, solved by fitting snubbers - I’m surprised you’ve missed a number of them since you joined. And the regular reports by people with improved layout performance after fitting them. Snubbers are very cheap insurance against potential decoder damage.

[Stephen in Kerry]

@Fishmanoz and others...The loco is working on it's own address and at address 3 (default).

I think I've narrowed down the issue (even further) to this diamond crossing, which is just after the 2 points I mentioned. The A4 goes dead as it his the plastic area (see photos).

All thoughts on the cause and (mainly) the solution are very welcome.

[ColinB]

Just to keep you happy Fishmanoz I will buy one and analyse what it does, although I must admit reading the description on the Rails website I am not really sure I need one.

[Stephen in Kerry]

I've worked it out, finally. The inner (far) switch rail needs to be pushed right up against the inner main rail, somewhere in the middle of the switch rail, in order for it to make contact.

I'm not sure why that is, and it's a bit of a pain, as use that point a lot.

I seem to need to push it with my finger every time I change the point and need to put it back.

If there's a solution, I'd like to hear it. :)

[96RAF]

I see no evidence of a motor actuating that point so is it man-draulic.

If it has a motor then is it perfectly central or can it be biased a bit more towards that side.

[ColinB]

It is either the spring on the cross member has fell out or the little clips that move with the blade on the point are not making contact. My points occasionally do it.

[Aussie Fred]

Fishy, am considering fitting snubbers on my new extended track. Is there any "negatives" to them. Tried researching it on Mr Google and got completely lost on all the technicalities.

[rpjallan]

No there isn’t any negatives to using them. It’s like removing capacitors when fitting a decoder. Your loco won’t run worse…

[Fishmanoz]

They use a little of your power, not much.