DCC Layout - Transient Voltage Protection

[NFWEM57]

I am planning to fit Transient Voltage Suppression (TVS) diodes across each power district on my layout but wondered what value to use.  I was considering using 600V TVS diodes with a minimum breakdown voltage of circa 17 volts and a maximum clamp of circa 25V.

 

Does anybody have any experience of using TVS diodes on a DC / DCC layout and can offer advice?

 

Patrick

Edited February 23, 2021 by NFWEM57
typo

[Crosland]

I have to ask why? This is not normal practice for a DCC layout. The DCC track voltage can be greater than 17V, depending what system you are using.

 

It is more usual to fit an RC snubber circuit to remove high frequency noise.

[NFWEM57]

2 hours ago, Crosland said:

I have to ask why? This is not normal practice for a DCC layout. The DCC track voltage can be greater than 17V, depending what system you are using.

 

It is more usual to fit an RC snubber circuit to remove high frequency noise.

Hi,

 

Thank you for responding.  On the why, because it is a simple, cheaply available solution which is not frequency dependent and draws no current; 27p a diode..!  An RC solution always draws current which will be a problem for current based block detection.  The TVS will absolutely resolve the issue of overvoltage, and so protect expensive decoders, but the problem with the transmission line issues, reflections and null points, needs other solutions, if indeed they occurs on a specific layout.

 

Patrick

[WIMorrison]

An RC 'snubber' does cause any issue to current based occupancy detection as they are fitted to the bus not the track feeds.

 

I have 80 current based occupancy feedbacks and I can issue you that the 'snubber' hasn't affected the performance of a single feedback (or anything else for that matter)

[NFWEM57]

1 hour ago, WIMorrison said:

An RC 'snubber' does cause any issue to current based occupancy detection as they are fitted to the bus not the track feeds.

 

I have 80 current based occupancy feedbacks and I can issue you that the 'snubber' hasn't affected the performance of a single feedback (or anything else for that matter)

Thanks for responding.

 

But an RC will not stop an overvoltage of significant duration relative to the time constant , so it is a combination of the two that will give true overvoltage protection and transmission line damping.

 

Very interesting that the US community often use them but because of their relative obscurity here they are seemingly ignored or even frowned upon.

 

Patrick

[Harlequin]

Some interesting discussion of voltage spikes on DCC systems here:

https://sites.google.com/site/markgurries/home/dcc-problems/decoder-problems/decoders-blowing-up

 

The author suggests that it’s an issue on busses longer than about 10m during short circuit events and is caused by inductance in the track bus.

 

He suggests twisting the track bus wiring (and explains why and when to do this in some detail) and/or fitting an RC snubber.

[Crosland]

15 hours ago, NFWEM57 said:

But an RC will not stop an overvoltage of significant duration relative to the time constant , so it is a combination of the two that will give true overvoltage protection and transmission line damping.

 

What's generating the overvoltage "of siginificant duration"? The problem with DCC is high frequency spikes due to poor wiring (another poster mentions the inductance), which an RC snubber will deal with. I wouldn't say they are obscure in the UK. The topic comes up reasonably often. The general consensus is that they do work.

 

Decoders will sometimes (often? always? I'm not sure) have a zener clamp to protect themselves.

 

You can't treat a DCC layout as a transmission line. It would be  futile exercise, the main problem being that the loads have a habit of moving around.