Found Source of Current On Empty Layout

[BCR751]

I recently purchased a RRAMP meter to monitor my partially completed layout.  I hooked it up yesterday and noticed a current draw of about 0.550 amps on the layout which didn't have anything that could draw current.  At least I didn't think so until I dug a little deeper.  My 6-turn helix has snubber circuits at the ends of each turn.  Each of those little R/C circuits draws about 0.092 amps which totals about 0.55 amps for the whole helix.  Learn something new every day 

Doug

[fcwilt]

I recently purchased a RRAMP meter to monitor my partially completed layout.  I hooked it up yesterday and noticed a current draw of about 0.550 amps on the layout which didn't have anything that could draw current.  At least I didn't think so until I dug a little deeper.  My 6-turn helix has snubber circuits at the ends of each turn.  Each of those little R/C circuits draws about 0.092 amps which totals about 0.55 amps for the whole helix.  Learn something new every day 

Doug

Why at the end of each turn?

Usually a snubber is placed at the end of the power bus or each bus in the case of multiple buses.

Frederick

[peteski]

Yeah, what Frederick said.  Besides, do you need snubbers at all?  Did the DCC system behave strangely without them?

Snubbers do consume current because they shunt some of the DCC current.  0.55A is like running couple of locos at full throttle under load.  That is a needless waste of current unless you really need the snubbers. If your DCC bus is noisy maybe you could rearrange it to minimize the problem.

What are the values of the cap and resistor you used in the snubbers?

[BCR751]

When I was preparing to build the helix, I did some research regarding how to wire it including asking many questions, most of them here on TRW.  Many of the answers said that each turn should be fed separately and that the open ends should be terminated with an R/C circuit, as should any open ends anywhere on the layout.  So, that's what I did and it seems to be working well, with the exception of the current draw.  By the way, the caps are 0.10mfd and the resistors are 100ohm, again as per the suggestions provided by the responders here.  I was reluctant to construct the helix as one complete wired unit and find out later it required individual wired turns with snubber circuits and have to re-build it.  So, it was built with individual turns and snubbers from the start and I don't know how the DCC system would have reacted had I done it differently.

[peteski]

I have feeling that you have been misdirected.  Can you find the specific threads which stated to feed each turn separately and add snubbers?  I would like to see the reasoning behind that method.

Helix is nothing than a simple straight track with its DCC bus underneath, except that it is spirally wrapped around an imaginary cylinder.  Noting there seems to me to indicate that each turn needs to be gapped and fed on one end of the turn and have snubber on the other end,  That seems silly.

I'm pretty sure that if you simply removed those subbers your layout would work just fine. It might be worth a try as you shouldn't have to modify any existing wiring (escape for disconnecting one lead of each snubber.

I don't have a layout but my friends do (DCC) and none of them as I recall use any snubbers anywhere.  I also help in setting up the NTRAK layouts (say about 20 x 20 feet average) at local model shows and we also do not use snubbers on the DCC loops.

[C855B]

"Snubbers", technically known as line terminators, were promoted a few years ago as a way to clean-up DCC waveforms on very long runs - 80-100 ft or more. And they did - a little.

Being a former RF guy, I understand the principle, which is to reduce the effect of standing waves which might reflect off the end of the network pair. The problem I have with this theory is the low frequency of DCC - around 10KHz. If it was >100KHz, maybe, >1MHz, for sure, but 10KHz is not a challenge, electronically. You might be able to see the effects of DCC standing waves on a o'scope if you looked really hard (slight rounding of the square-wave signal), but barring direct instrumented readout indicating the issue, I wouldn't mess with terminators.

Besides, even just one load on the bus (load = locomotive) will provide more than enough termination. I suspect - cannot prove, of course - that the whole "snubber" issue was promoted by somebody who actually saw waveform improvement, but there were other issues on their particular sample that were helped, not just network length alone.

FWIW.

[peteski]

I'm not an RF specialist (I do have digital and analog electronic background having worked in the field) so I don't know about standing waves (I did learn about them but that was long time ago). But one thing for sure is that DCC signal is a relatively low-frequency square wave and snubbers do dampen the ringing on the leading and trailing edges of the signal. But as you mentioned, they are an overkill in most installations. I would only install them as needed if the DCC system seemed to act weird  Overall, DCC is a fairly robust design which will work even if the signal is far from a perfect square wave.

Like I mentioned earlier, I am fairly certain that Doug's layout would work just fine even with no snubbers.  This is a good article for Doug to read about snubbers (lengthy, but lots of useful info and explanations - worth reading):
https://sites.google.com/site/markgurries/home/dcc-general-best-practices/wiring-planing/snubbers-rc-filter
And another, right from a DCC manufacturer:
https://ncedcc.zendesk.com/hc/en-us/articles/204128735-Snubber-RC-Filter
And a little less authortiative one from Tony's Trains:
https://tonystrains.com/snubbers/

If you Google for DCC snubbers, there is alots of info out there.  But some is laughable. Some people swear that snubbers cured mysterious slow-downs of their trains on their layouts. That makes no sense.  Ringing DCC signal (which snubbers cure) does not change the average voltage on the DCC bus (which the decoders see and use to energize the motor). Ringing (noisy) DCC signal also cannot somehow generate speed step control packets which would control the train speed.  Since the speed commands from the control station are not changing when their train slows down, the things I mentioned would be the only things which could affect the speed.  And I don't see how that could happen and how a snubber would cure it.

[C855B]

That first link is really good stuff, so thanks! He is talking about it as a technique for general filtering, not just simple network termination. He is also addressing (and agrees with) my recommendation for twisted-pair bus wiring, which I usually don't bring up around here due to getting flak about it, "But the rails aren't twisted!", etc., etc.

As he says (and I also agree), every situation is different, and that "snubbers" are a band-aid. I would actually avoid snubbers as much as possible. Lacking any indication of a problem and lacking confirmation with a 'scope, they actually introduce a problem by severely rounding-off the leading and trailing edges of the DCC square wave. Suppress ringing, yes, but I would experiment with my own setup. The current recommendation of 0.1uf and 100 ohms is a cutoff of anything above 15KHz. You're nearly turning the DCC signal into a sine wave, and digital logic isn't particularly happy with that. I would start at 0.03uf (~50KHz) and work my way down.

The Tony's Trains info is mostly unedu-guess BS. The observation about using speaker wire is sort of backing into a poor-man's twisted pair, two closely proximate conductors creating a balanced transmission line whose electromagnetic fields cancel each other out and provide some protection against noise.

I'm amazed at how much of this stuff I've forgotten, and had to look up. I used to write filter software for my little EE department at NASA/FRC. That was nearly 45 years ago.

[BCR751]

Well, my work-in-progress layout was started sometime ago when the snubber thing was all the rage and it seemed like everyone was doing it.  Not knowing a helluva lot about DCC at the time, I went along with it.  There has been a lot of excellent knowledge about DCC put forth since then.  The first link by Mark Gurries explains it very well.  I think what I'll do is remove the snubber terminations from the ends of the helix turns as Peteski has suggested.  If I notice any difference in train performance, I'll mention it here.  Also, I searched for the threads I read way back when but I can't find them.  By the way, my main bus runs are indeed twisted so I got that part right 

Thanks for all the input.  It's much appreciated.

Doug

[fcwilt]

While it does make sense to provide power feeders to each turn I don't think the turns needed to be isolated, unless you were doing some sort of current based occupancy detection.

I have snubbers on my layout and they did help BUT they are ONLY at the ends of the power buses.

I would remove the snubbers from each turn - they are not needed.

Frederick

[peteski]

I have snubbers on my layout and they did help BUT they are ONLY at the ends of the power buses.

I'm curious: what kind of problems did you have which were cured by installing those snubbers?  How large is your layout and length of the DCC bus?

[fcwilt]

I'm curious: what kind of problems did you have which were cured by installing those snubbers?  How large is your layout and length of the DCC bus?

The buses were up to 20 feet. Due to some unique wiring requirements I could not twist the bus wires to any significant degree.

The problem was trying to get a good reading with a RRampMeter with nothing on the track.

A scope showed a good deal of overshoot - enough that I was concerned about damage to decoders.

The snubbers fixed the problem and most of the overshoot was gone.

[peteski]

The buses were up to 20 feet. Due to some unique wiring requirements I could not twist the bus wires to any significant degree.

The problem was trying to get a good reading with a RRampMeter with nothing on the track.

A scope showed a good deal of overshoot - enough that I was concerned about damage to decoders.

The snubbers fixed the problem and most of the overshoot was gone.

Ah, that makes sense.  Like Mike said, snubber acts as a simple RC low-pass filter, so it smooths out (shorts out) the high-frequency components in the overshoot.

[BCR751]

Ah, that makes sense.  Like Mike said, snubber acts as a simple RC low-pass filter, so it smooths out (shorts out) the high-frequency components in the overshoot.

I have no idea what any of that means but if I start frying decoders, I'll re-install the snubbers 

Doug

[Lemosteam]

OK so is this type of wiring wrong (total length 20 feet):