Will short circuits damage a DCC decoder?

[drgw0579]

Besides my own, I sometimes operate regularly on two other layouts.  The other two layouts seem to experience more decoder failures and the owners blame short circuits caused by simple operator inattention like running through a switch.   Doesn't matter what type/brand of decoder.  The three layouts all have different brands of DCC boosters.  One has been wired by an Electronic professional, and he is installing DCC Specialties circuit breakers in the many power districts.  He also maintains a circuit breaker will save the booster and not the decoder.

I don't want to get into a brand X vs brand Y discussion, but do you all sometimes experience decoder failures that you feel was caused by a simple short circuit?  I am not talking about some sort of catastrophic problem such as accidentally having high voltage come into contact with your track bus.  Just ordinary operating.  Modern boosters seem to be able to detect short circuits and cut the power immediately.  Why would that damage a decoder?

Bill Kepner

[RBrodzinsky]

The "normal" short circuits caused by running a switch, misaligning wheels, etc., should never cause a properly installed decoder to fail, as long as circuit breakers or command station react to it. The only shorts that normally let the magic smoke out are issues within the loco, such as having a short between the track power and the motor leads.  Of course, without proper circuit protection on the system, and too much current, if your loco is causing the short (bridging rails) and that current continues to flow without something else stopping/breaking it, you can do more damage to your loco than just blowing the decoder.

[peteski]

Are the decoders truly fried or do they just "lose their mind"  and will work again after factory reset and reprogramming?

I agree that decoders shouldn't be damaged (physically burned components) by any problems with the track voltage assuming that it doesn't grossly  exceed the decoders ratings.

But (depending on many factors like the brand of booster or circuit breaker and the electrical characteristics of the DCC bus, including the track), a voltage spike might get generated either when the track power is turned of or reapplied,  It seems that this voltage spike might affect the decoder's programming (like I said, scramble its settings).  But that problem is usually reversible (as I mentioned earlier).

It is best to avoid any shorts on DCC tracks. Decoders are complex mobile computers and while they are robust, they can be affected by conditions which are out of their normal operating rating.  DC-based models did not have these problems because all the model had was a motor, light bulbs and maybe couple of diodes.  Simple, robust, and no-frills design. But no bells-n-whistles.   

[sp org div]

Have only experienced one decoder total failure (out of fleet of 40), attribution unknown...  We are almost done with a replacement program of all the Dgtrx PM42's with PSX cards for power distribution. The layout utilizes Dgtrx dcs / bstr. Decoders have never had any noticeable reaction to locos that short against improperly thrown turnouts (or other shorts), other than sound units briefly cutting out in unrelated Power Districts, when there was a short trigger of another PD section (the reason for the recent change).  The worst annomalies seen with decoders to date, have been a loss of consisting (Universal Consisting), only at start of ops after all locos have gone through wheel cleaning procedure (less than one loco per session).  Some layouts utilize additional snubber protection at the ends of feeders to every block, but with a near 40 block sections and some over 30 feet away, we have not yet seen an urgent need to go there...  Preliminary testing of the PSX cards has been favorable, so anticipate a much more stable sound environment here.

Jeff
SP Oregon Division

[C855B]

... or do they just "lose their mind"  and will work again after factory reset and reprogramming? ...

This. Usual sloppy operating conditions with our N-Trak layout at a show yesterday. By the end of the day we had to reset and reprogram close to a dozen units after one mishap or another.

[RBrodzinsky]

I only have one loco that sometimes loses its mind - my Kato GS4 with a Tsunami 750.  For whatever reason, that unit seems the most sensitive to spikes.   I keep its DecoderPro profile handy, for quick restoration.

[sp org div]

This. Usual sloppy operating conditions with our N-Trak layout at a show yesterday. By the end of the day we had to reset and reprogram close to a dozen units after one mishap or another.

Interested to know why on some layouts and not others. Perhaps lack of additional power management protection... brand of DCS... or other?  I use a full mix of decoder brands and do not see this at all.

[C855B]

I know I'm not going to get into diagnosing it on our layout. Modules range from relatively new and finely-crafted to 30-year-old weatherbeaten wonders. The range of member's locomotives is equally wide, plus we have one member in particular, Joe Btfsplk Jr., who curses the layout every time he shows up trying to run any number of basket case locos. We see him coming and I pretty much station myself at the electronics cabinet waiting to deal with the fallout from his carelessness.

[crusader27529]

A short condition could definitely damage a decoder......

When a short happens, the electronics of the booster tries to supply the correct voltage by increasing the current (ohm's law). And when the short is removed, there is likely to be a voltage spike because the electronics are still trying to supply the correct voltage, but the voltage overshoots what is correct. Depending on the overshoot level, it may damage the decoder or just get it confused. The spike is worst close to where the physical short is located, so the loco causing the issue is the most likely to experience the overshoot, but other decoders also can be affected.

So, depending on how fast the booster responds to the short, or the severity of the short, and the design of the decoder and how it protects itself from overvoltage determines if damage happens.

On my relatively small club layout some members were complaining/experiencing decoder problems and/or failures. I installed multiple 18v zener diode pairs on the layout, and all problems went away.

The zener diodes were configured as 2 in series with one being reverse polarity to the other. The way it works is that the one of the zener diodes acts like a zener shoul, while the other acts as a blocking diode......so the blocking diode blocks reverse current through the zener and when the voltage exceeds the zerer rating(plus one diode drop), the zener shunts any excess voltage, which chops off JUST the overshoot. On the opposite polarity of the DCC waveform, the functions of the diode pair are reversed.

Configuring the diodes this way allows for current to flow through them ONLY when an overshoot occurs. Think about it.

The diodes I used are rated at 18v, so the clipping happens only when the voltage exceeds 18v + 0.7v. The clipping is VERY fast, in the sub-microsecond area.

[C855B]

A short condition could definitely damage a decoder......

...Think about it. ...

I do and did, and I disagree. If this was the case, then MRR'ers everywhere would be howling about fried decoders and the decoder manufacturers would quickly be out of business doing warranty replacements.

Shorts across the rails are SOP on even the best run and maintained model railroad. To assert "...definitely damage..." flies in the face of the empirical evidence otherwise. Moderate that to "...may damage..." and then we can talk about proving the theory behind adding zeners or other voltage limiters, which also presumes that fast-response voltage limiting is not already designed into decoders, control stations or boosters, which it is.

> ...when the short is removed, there is likely to be a voltage spike because the electronics are still trying to supply the correct voltage, but the voltage overshoots what is correct. ...

Sounds good on the surface, but I'd like to see before/after oscilloscope traces backing this up.

Given the outright careless operation by the clueless on our club layout, I have seen no "fried" decoders, just scrambled ones put back into operation after a quick reset. These are fully attributable to momentary shorts screwing up the data signals which in turn trigger non-robust firmware into unanticipated states. BSOD, decoder style.

[peteski]

Interested to know why on some layouts and not others. Perhaps lack of additional power management protection... brand of DCS... or other?  I use a full mix of decoder brands and do not see this at all.

The most likely of the problem (of the killer voltage spikes) is not the electronic components (boosters, circuit breakers) bu the DCC bus itself (the wiring and the track too).  From the electrical standpoint it has resistance, capacitance, and inductance. Because the square-wave shape of the DCC power signal as fast rise and fall times, the these characteristics of the DCC bus  will affect the DCC signal. If you were to hook up an oscilloscope to the DCC bus, instead of seeing a clean square-wave, you will see lots of noise and ringing at the leading and trailing edges of the waveform. THe amplitude of that ringing can get quite large (again, depending on the bus' electrical characteristics).  Then as mentioned, when a short occurs, even higher amplitude ringing can be generated.  This is not your fathers DC layout. 

A snubber circuit is a simple way to reduce the spikes and ringing.  Adding a double Zener diode should also work for trimming excessive-amplitude voltages.  Many decoders have a Zener diode installed right on the output of the bridge rectifier which should also clip the incoming  voltage spikes. But DCC decoders also  have the microprocessor inputs connected to both sides of the raw DCC track voltage (through resistors) so the mircoprocessor can monitor the bus and receive the data packets.  I suppose that it might be possible that the damaging spikes enter the decoder through those inputs (although most microcontrollers incorporate input pin protection internally).

The bottom line is that DCC is nowhere as electrically simple as DC and problems like what is described here will happen.   There are ways to deal with them (don't create shorts, install snubbers or Zener circuits).  This is the price we have to pay for the extra functionality DCC gives us over DC.  But the problems described here are an exception than a rule. They occur on relatively few DCC-equipped layouts.

[crusader27529]

I do and did, and I disagree. If this was the case, then MRR'ers everywhere would be howling about fried decoders and the decoder manufacturers would quickly be out of business doing warranty replacements.

Shorts across the rails are SOP on even the best run and maintained model railroad. To assert "...definitely damage..." flies in the face of the empirical evidence otherwise. Moderate that to "...may damage..." and then we can talk about proving the theory behind adding zeners or other voltage limiters, which also presumes that fast-response voltage limiting is not already designed into decoders, control stations or boosters, which it is.

> ...when the short is removed, there is likely to be a voltage spike because the electronics are still trying to supply the correct voltage, but the voltage overshoots what is correct. ...

Sounds good on the surface, but I'd like to see before/after oscilloscope traces backing this up.

Given the outright careless operation by the clueless on our club layout, I have seen no "fried" decoders, just scrambled ones put back into operation after a quick reset. These are fully attributable to momentary shorts screwing up the data signals which in turn trigger non-robust firmware into unanticipated states. BSOD, decoder style.

You missed the most important word in my response:

COULD definitely damage a decoder.....

[peteski]

A short condition could definitely damage a decoder......

LOL!  You would make a good  politician.   You could have even wrote "could possibly, maybe, certainly, definitely damage a decoder" to cover all the bases.     If pressed hard, would you say that it could or that it will definitely?

I think that the definitely Trumps could in your sentence,

[crusader27529]

The problem with certainty is that the electronics in question may be damaged and another time, the same hw and overvoltage pulse doesn't damage it.....

I have experience where decoders were scrambled by an overvoltage pulse (I assume, because it fit into the short causing a problem bucket) but as stated, the decoders aren't rediculously fragile, so observations far from a lab with test equipment are just educated guesses.

I can tell you that since I installed my zener diode protection on the club layout, we haven't seen any failures.....granted, that's not definite proof, but it's all I'm going to be able to add to the discussion.

So, yes, I've definitely seen a decoder killed when a short circuit condition was involved, and YES, it doesn't happen all the time. And, I despise PC commentary, so I don't do it.

I respect your comments, but I'm not trying to be polital at all.

[peteski]

The problem with certainty is that the electronics in question may be damaged and another time, the same hw and overvoltage pulse doesn't damage it.....

I have experience where decoders were scrambled by an overvoltage pulse (I assume, because it fit into the short causing a problem bucket) but as stated, the decoders aren't rediculously fragile, so observations far from a lab with test equipment are just educated guesses.

I can tell you that since I installed my zener diode protection on the club layout, we haven't seen any failures.....granted, that's not definite proof, but it's all I'm going to be able to add to the discussion.

So, yes, I've definitely seen a decoder killed when a short circuit condition was involved, and YES, it doesn't happen all the time.

A snubber would likely take care of the problem, but there is nothing wrong with using Zener diodes to clamp the DCC bus voltage).