Zener diodes in Parallel .... Not your typical project

Yes, I know, before someone says you can't do that because one Zener will hog the system, hear me out.


For a large project I am working on, imagine a long 500 foot 4 conductor power rail (+48V DC, GND, and two signal lines).

Each signal line is pulled up to the 48V through two 5 Watt 5k resistors at each end of the rail (Only FOUR pullup resistors on the entire system).

Every 10 feet you have a carriage that rides (<-- it can move along the rail) and picks up data and power from the 4-conductor rail via a brush track.

Each carriage has two 12V 500mW Watt Zener diodes. One connected between each Signal line and Ground.

There would be a maximum of 50 Carriages but at times there may only be 30 Carriages essentially paralleling the Zener diodes between carriages.


Does anyone see a potential issue with this configuration? Susceptible noise issues, etc.?

Note: A single carriage can handle the load from the 48V and 5k resistor (effective 2.5k because of one 5k on each end of the track) on each signal line Zener. The additional carriages would be distributive across the system.

how about a schematic▌of your proposed configuration.

A schematic would be great - are there current limiting resistors on each carriage signal zener? E the zeners 500W or 500mW?

is this a toy Lionel train or a person-carrying size train?

Sure, it's basically what is described in the text. (See "Test setup.pdf" for a basic the schematic)

Sorry, the Zeners are 500mW

More like a "person-carrying size train"... each Carriage weighs about 400 pounds

Note: The Carriages will also grab the full 48V DC, but it's not shown in the schematic for the purpose of this question

I can't see a problem. The lowest voltage zener might hog all the current, but it doesn't matter. With two 5 k resistors in parallel, it works out at about 170 mW to be dissipated in the 500 mW zeners.

The objection to paralleling zeners is that doing so doesn't usefully increase the power, due to one hogging all the current, but this is done to protect the electronics each carriage, I think, and the zeners have a large enough rating to cope individually.

It seems odd to have the pull-up to a voltage so much higher than the running voltage. Any capacitance on the signal lines would mean that if the system were powered before the first carriage was attached, the zeners in that carriage would take a current spike.

Is there any reason why there couldn't be a lower voltage limit? Perhaps a 15 or 18 V zener between the signal rails and ground on the fixed part of the installation?

"It seems odd to have the pull-up to a voltage so much higher than the running voltage " - I wanted a specific current level on a multi drop communication scheme to help overcome any noise issues. Each carriage is capable of pulling (sinking) the signal line to ground for communications purposes.

"if the system were powered before the first carriage was attached" - there will always be at least 30 Carriages in the system.

"Is there any reason why there couldn't be a lower voltage limit? Perhaps a 15 or 18 V zener between the signal rails and ground on the fixed part of the installation? " ... That's a good idea, I will implement that

I'd not pull the track supply lower than 24V - That gives a good wetting voltage for the moving contacts.

And consider increasing the individual carriage zener voltage, to 24V as well.

As things age and wear, 12V may be marginal for contact "wetting"; I've seen regular problems on 12V systems as they get older, with visually-perfect switch or relay contacts going high resistance.

If any carriage has good contact, the track voltage is held to 12V with the setup as described, and that could cause problems with another carriage that has slightly less than perfect contact.

24V (or higher) is enough to break the oxide or tarnish layer and keep working reliably.

If the only in line contacts are the track brushes, so the whole system is getting continuously scraped clean, it may be OK - but just something to be aware of..

I maintain a machine that uses very similar track/wipers.
But it doesnt send data just proxy signals.
They are not that reliable, a small amount of crud makes trouble.