This comes up on many threads over and over so here's a simple explanation all consolidated into a single thread.

A TVS goes in the system to prevent over-voltage stress from damaging electronics. This is different than a fuse or circuit breaker that protects from over-current stress. Unlike current at low frequency (60 Hz for most layouts) the transient voltage events that result from derailments, collisions, and general sparking are very broadband transient events, usually in the range of 1ns to 100ns.

That means they are a wave function, like a pulse of voltage bouncing up and down the wiring that has a fixed length, not like a continuous current flowing through it. At 1ns (so 1 GHz bandwidth) the wavelength in air is 11.8 inches and in RF we usually say 1/4 wave is where the wave effects start to matter.... so about 3 inches in air. However cables are not air and have a slower wave velocity depending on their dielectric constant (a property of the insulator) and their physical geometry (diameter, and how far apart the + and - are). All in all ... you can think of transient pulses as about a 1 inch wide event.

So if you put a TVS into a circuit and that TVS is more than an inch away from the thing you're trying to protect, chances are you're not actually protecting it because of the wave effects.

See the handy diagram below that explains what's going on. The bird is your electronics and the guy waving the rope is the train, making angry waves. If you pinch the cable in the wrong place (that's what a TVS does, clamp the cable voltage) the bird still gets bounced. The only way to avoid this is to do the pinching near the bird, in this case within the 1 inch.

and that's why the TVS needs to go in the locomotive right by the board, or in the TIU right by the drivers.

If you want a much more academic understanding.... Take a look at the telegrapher's equation.

#### Attachments

Images (1)
Original Post

Very well said. Just sitting here waiting for all those that believe their fast-acting breakers will take care of the issue to chime in...

I am currently in the throes of doing this.  I'm a newbie, so what's a TIU?  So this should go before the board, not after (like just before the motor).

A TIU (Track Interface Unit) is a control system used by MTH (Mikes Train House) to control DCS (Digital Command System) equipped engines. It's a box that sits between your power source and the track- hence why they want to keep voltage spikes from killing it.

@texgeekboy posted:

So this should go before the board, not after (like just before the motor).

Correct, if putting inside a loco, the idea is the power entry into the board from the track pickups. In theory, a good control board already accounts for anything at the motor that it controls so not putting one between the motor and board.

I have always been skeptical of how much good a TVS will do when installed at a track feeder, as I have done. If your engine is the cause of the transient spike and your engine is between two feeders it will take a hit. Some claim the TVS is doing something in this situation and I don't have a reason to doubt that too much. The absolute best place for a TVS is in circuit (edit to clarify I didn't mean in series) with the engine pickups. Someday I'll work up the courage to install one in each of my engines and remove the 6 or seven I have installed on the layout.

Last edited by turkey_hollow_rr
@texgeekboy posted:

I am currently in the throes of doing this.  I'm a newbie, so what's a TIU?  So this should go before the board, not after (like just before the motor).

A TIU is a coveted rare and very valuable device that will become the new gold standard this year.  It is part of a Digital Crugerrand System [sic].

John

For my newish layout I have the ZW-L and did not install any TVS(s)... the ZW-L cuts power if it even 'thinks' there will be a derailment.  The breakers on it have never tripped.  On the Christmas layout I use the ZW with 5amp breakers (apparently to slow & a very, very sad situation) and TVS(s) at the feeders (apparently, a waste of time).  In either case I've not blown a board... although, I did fry a motor due to a quartering issue that I didn't catch in time many, many years ago.

Very well said. Just sitting here waiting for all those that believe their fast-acting breakers will take care of the issue to chime in...

Rob,

Those folks don't have their beliefs and stick to them because they want to be bull-headed.

They just don't know any better at this point, and they aren't listening to the conversation.  There's no one to argue with what you're saying.  They're just flat-out ignoring it.

I thank @Adrian! for adding a deeper technical analysis that clarifies a little better what many of us, including you, have been saying, for quite some time.

Keep up the good work.  At some point the message may finally get through.

Mike

Adrian, Wow!  It's been too long since I had to think this through.  I read the Telegrapher's Equation link, and it reminded me of college.  I graduated in 1976!!  Actually, The bird diagram caught my eye before I read your text or the link.  As soon as I saw the diagram, I realized putting the TVS where the track feeder connects to the track doesn't take into account the rail, engine pickup roller, and wheels.  Yes I have PDXs on my TIU outputs.  Now I need to put the TVSs in the engines.  I'm rotating through my engines, lubing them, adding BCRs where needed, and giving them a good run, so I'll dig out the extra TVSs too.

Rob,

Those folks don't have their beliefs and stick to them because they want to be bull-headed.

They just don't know any better at this point, and they aren't listening to the conversation.  There's no one to argue with what you're saying.  They're just flat-out ignoring it.

I thank @Adrian! for adding a deeper technical analysis that clarifies a little better what many of us, including you, have been saying, for quite some time.

Keep up the good work.  At some point the message may finally get through.

Mike

Mike, you are so right.  Even many of us who have worked in electronics for 40++ years don't think things through.  No wonder they pushed me out with an early retirement.  Now the question is, why did they take me right back as a contract engineer for 3 more years before they finally had me hang up my multimeter?

Rob, Adrian's drawing is what did it for me.  About 10 years ago while working in engineering at the telecom company, I happened upon a video that explained how the Internet works in very simple, every day terms.  I shared it with our technicians, and they said it worked wonders in helping explain to business customers why they were having certain issues.

Last edited by Mark Boyce

Very well said. Just sitting here waiting for all those that believe their fast-acting breakers will take care of the issue to chime in...

Those folks don't have their beliefs and stick to them because they want to be bull-headed.

They just don't know any better at this point, and they aren't listening to the conversation.  There's no one to argue with what you're saying.  They're just flat-out ignoring it.

I thank @Adrian! for adding a deeper technical analysis that clarifies a little better what many of us, including you, have been saying, for quite some time.

The instant breakers aren’t meant to protect against voltage spikes.  They’re meant to open the circuit as quickly as possible to minimize the time a derailment is sparking, arcing, etc.

Adrian’s explanation of TVS diodes is wonderful, and explains why placing them as close as possible to the electronics is the best way to install them.

Personally, I recommend both TVS diodes and quality circuit breakers for all layouts that have Protosound, DCS, TMCC, or Legacy locomotives running on them.

@texgeekboy posted:

If the diode has to do its job, is it easy to tell if it did and needs replaced?

They don't "expire" or become consumed with a "use". The typical failure mode is shorted, so yes, it will be easy to tell if they fail, impossible to tell if they are doing their job without an oscilloscope.

Rob, Adrian's drawing is what did it for me.

A forum search will show that I've advocated for placing the TVS diodes as close to the protected component as possible... it's most often just not practical.

For my newish layout I have the ZW-L and did not install any TVS(s)... the ZW-L cuts power if it even 'thinks' there will be a derailment.  The breakers on it have never tripped.  On the Christmas layout I use the ZW with 5amp breakers (apparently to slow & a very, very sad situation) and TVS(s) at the feeders (apparently, a waste of time).  In either case I've not blown a board... although, I did fry a motor due to a quartering issue that I didn't catch in time many, many years ago.

The transients discussed here will not usually trip a breaker. You can have these waves of sporadic voltages wash over your electronics and you would never know, unless it takes out a component. The TVS is pretty inexpensive so it is a form of cheap insurance for our digital command electronics. I know I poked a finger at TVSs installed at the track feeders but I can't say for sure they don't do anything good so I wouldn't be quick to call it a waste of time.

Dan,  I don't plan on changing anything.  The Christmas layout already has the TVS(s) built into the feeders by me as is the mini rack of terrible, awful incredibly slow 5amp breakers.  At the very least the TVS(s) will hopefully remind me that I should move them closer to the internal electronics one of these days... and that is probably not any better than a 50/50 proposition as I've yet to blow a board in the last 20+ years.  Science is fun... and, hats off to Adrian for providing a lucid explanation with demonstrable graphics for knuckle heads  like me!

Last edited by Dennis-LaRock

Dennis, it is a small victory but I admit I sleep a little better with the TVSs in place.

Dennis,

People nowadays enjoy digging in their heels.  Stubbornness is a sign of the times.

My version is similar. I have an American car that is 13 years old and has gone 250,000 without a major repair (knock on wood).

By your logic every American car should also go 250,000 miles without a major repair because mine did.  As much as I vigorously support the American car makers, because I'm employed by them, even I can't believe that.

Statistics are not in your favor.

Something will die, sooner rather than later.

Mike

Last edited by Mellow Hudson Mike

Mike,  I do get it.  Probability is not in my favor.  Here is a question... when, I do blow a board will it definitely and only be because of transient spike?

Here is a question... when, I do blow a board will it definitely and only be because of transient spike?

Of course not!  There are tons of reasons for a board dying, and there is no concrete way to verify if a dead board was the result of a voltage spike.  If every dead board was the result of a voltage transient, I'd cover my locomotives with TVS protection diodes and never have to work on another electronic failure!

Adrian, excellent use of a metaphor. I understand the benefits that a TVS diode can provide and have purchased them.

Do any train manufacturers see value in adding a TVS prior to their electronic boards during the manufacturing process?

Thank you John.  My point is that there are 'many' variables... and, at my age I'm  more likely to have a catastrophic failure due to dropping a train on the floor ...so, should I apply adhesive to my hands before handling the trains?  I'm absolutely 'not' saying don't install the TVS(s) to further protection... by all means do so!  I'm just not as concerned... I have more pressing issues such as did I enter 9 or 90 when addressing my MU(s)???

Last edited by Dennis-LaRock

Adrian, excellent use of a metaphor. I understand the benefits that a TVS diode can provide and have purchased them.

Do any train manufacturers see value in adding a TVS prior to their electronic boards during the manufacturing process?

Not so much the Train manufacturer's, but when I was designing avionics in a previous life, we used them to protect electronics.  MTH has them in transformers and the TIU.  It seems odd that neither MTH or Lionel has added a TVS to the inputs to their locomotive electronics, it would be a cheap addition that would likely yield big reliability benefits.

Last edited by gunrunnerjohn

It seems odd that neither MTH or Lionel has added a TVS to the inputs to their locomotive electronics, it would be a cheap addition that would likely yield big reliability benefits.

CHEAP!   Going from xx95.95 to xx96.95 (parts & labor (China)) is for all practical purposes, inconsequential to say the least.  Just imagine... this thread and the inherent anxiety wouldn't even exist.

CHEAP!   Going from xx95.95 to xx96.95 (parts & labor (China)) is for all practical purposes, inconsequential to say the least.  Just imagine... this thread and the inherent anxiety wouldn't even exist.

So what would we have to complain about!

Since I only purchased my first locomotive with an IC board a month ago, and seeing all the discussions here on that, I've wondered that as well.  Surely it wouldn't be that expensive or time consuming during the building of these things to add that protection.

Again, since I'm relatively new to this I have to ask if you know of any units that were damaged that could have been protected by TVS?

@texgeekboy posted:
Again, since I'm relatively new to this I have to ask if you know of any units that were damaged that could have been protected by TVS?

There's really no concrete way to know if when a board stopped working that it was due to a transient event that could have been prevented with a TVS.

There's really no concrete way to know if when a board stopped working that it was due to a transient event that could have been prevented with a TVS.

A good destructive analysis lab with x-rays and SEM and everything can do this pretty easily actually. It's super duper painful and super expensive and only happens when I fail horribly as a designer!

A good destructive analysis lab with x-rays and SEM and everything can do this pretty easily actually. It's super duper painful and super expensive and only happens when I fail horribly as a designer!

Yep, I was talking practical ways.   I have heard of doing stuff like this, but I never had occasion to have such an analysis done.  I'm talking about back here in the real world.

Adrian... Have TVS(s) been installed in all of your (or, the clubs for that matter) locomotives which require circuit boards for proper operation?

Adrian... Have TVS(s) been installed in all of your (or, the clubs for that matter) locomotives which require circuit boards for proper operation?

We have them in the TIUs, and I have them in my locomotives. I don't know what the rest of the demolition derby is doing...

We have them in the TIUs, and I have them in my locomotives. I don't know what the rest of the demolition derby is doing...

Awesome!!!!

Well, I do sincerely apologize for my lack of anxiety... and, for not blowing a board since my first TMCC purchase around 20 years ago.  I know my days are numbered!

One of the worst offenders for HV transients I've seen is when people use an insulated section of the outer rail to control a electro-mechanical relay for an accessory.

If you don't remember to put the kickback diode on the relay coil it annihilates the little lionchief trains pretty handily.

One of the worst offenders for HV transients I've seen is when people use an insulated section of the outer rail to control a electro-mechanical relay for an accessory.

If you don't remember to put the kickback diode on the relay coil it annihilates the little lionchief trains pretty handily.

Got that problem licked! Insulated Track Signal Driver, Rev. 2

#### Attachments

Images (1)

I've asked this before in other threads, but I'd like to hear your answer.  If I trace the wire coming from the wheels, and the wire coming from the center pickup, I should place a TVS between those 2 wires just before the molex connector that goes into the circuit board, correct?

You can actually usually solder it to the underside of the board power input connector.

@texgeekboy posted:

I've asked this before in other threads, but I'd like to hear your answer.  If I trace the wire coming from the wheels, and the wire coming from the center pickup, I should place a TVS between those 2 wires just before the molex connector that goes into the circuit board, correct?

Okay so this goes into the weeds a bit.

Inside the TIU where I put the TVS diodes (at the ACT244) there is no power signal, just the +/-5V logic signal, so if you put a 9V-limit TVS that's good.

Out at the train pickups there is also power, not just signal, which at least on our layout is 18V RMS, so if you put a 9V clamp there, it'll clamp the power supply and blow up. So if you want to put it right there, you need to pick a clamping voltage that won't clamp the power, and also have a bit of margin, say 3-4V, So you'd be looking for a 25V range TVS. Then yes it can go right across the inputs. Also you need to make sure it's bi-directional and not a one way one. If all you can find is one-way ones, you can just put two in parallel pointing opposite directions.

I just bought these on eBay in anticipation of doing this.  Would they be okay?  From what I can tell, the specs are pretty close to yours.

Lionel ZW Transformer Circuit Protection 1.5KE36CA TVS Diodes 30.8V 1500W