Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Back EMF protection and an H bridge.

Status
Not open for further replies.

T i m

New Member
Hi and a HNY to all,

I am still doings bits on an Arduino driven model train project for my BIL and last night I had a loco shuttling up a test mainline track and back via a little runaround loop (with two sets of points working in parallel).

I was testing driving the 'solenoid' type points / turnouts using a 2A (3A peak) 'Dual H bridge' (very cheap on eBay):

https://preview.tinyurl.com/lktzzp3

and a pair of diodes and a capacitor to ensure the solenoid coils aren't burned out. The circuit looks like this:

**broken link removed**

My question is please, should the back-EMF diodes that are on the bridge module itself be sufficient (to protect the bridge IC), or would it also be prudent / possible to fit some additional diodes, like this? Or maybe because of how the circuit works they aren't needed?

**broken link removed**

If it will still work with those additional diodes in place, I was thinking of making a 4 - way PCB that would go between the H bridge and the points and I could fit all the components on there (in case I use an H bridge that doesn't have them fitted), rather than nearer the points themselves, as I (or another) might (accidentally) wire the points a different way round and thefore the diodes could end up being fwd biased.

All the best for 2015 and thank you for your time in any case. ;-)

Cheers, T i m
 
Extra diodes not needed. There is no current through the switch (therefore no voltage spike or arc) at the instant that the switch is thrown.
 
Extra diodes not needed. There is no current through the switch (therefore no voltage spike or arc) at the instant that the switch is thrown.

Hi Mike and thanks for the prompt reply.

Good point, (and good news for me and driving the points electronically especially) and when you say 'thrown' I assume you mean (specifically, as that is when back EMF is an issue ... and just to clarify), the switch contact is 'broken'.

As an aside I did an experiment where I limited the current from my bench PSU to 1A and the points didn't fully throw at the same voltage (15 in that case), but without putting a scope on it (or knowing the maths) I don't know what the current (pulse) value actually is. Doubling the value of the switching cap and adding a cap directly across the PSU input of the H bridge allowed the points to work reliably again at 15V.

However, I think I understand it may be better to simply raise the voltage to around 20, as that should result in a lower (or no worse) current pulse than if I increased the value of the switching capacitor?

Cheers, T i m
 
I was assuming that the switch is moved to the "other" position, and left there until the next time it is moved to the "original" position again. It isn't a center-off, three position switch, is it?
 
I was assuming that the switch is moved to the "other" position, and left there until the next time it is moved to the "original" position again. It isn't a center-off, three position switch, is it?

Yes Mike, you were spot on in that it's a straight 'changeover switch' but it's not actually a electro-mechanical switch as such but one side of an 'H Bridge'.

**broken link removed**

I looked at the original mechanical switch and saw it provided a voltage and when switched, a path to earth. I looked at the H Bridge (I initially bought for track PWM power and direction control) and thought that one 'side' of the bridge did the same thing, took a point from Vcc to Gnd. I wired it up as per the mechanical switch (with the cap that prevents coil burnout) and it worked as I hoped it would. ;-)

This means a very cheap (dual) H bridge can drive 4 sets of points and ... without wanting to re-invent the (model railway) wheel, thought it might be a practical, readily available and value_for_money solution?

My next thought (re modularising and simplifying) is to have a PCB made to hold the two diodes and the cap (x4) but if the additional back EMF diodes aren't required I might just add the basic components directly 'in-line' as such.

All the best and thanks again.

T i m
 
If the load is a long way from the bridge then back emf protection at the load is a good idea to prevent the wiring radiating rf, for motors 3 x 100nf caps work well, one from each terminal to the case, and one across the terminals.
 
If the load is a long way from the bridge then back emf protection at the load is a good idea to prevent the wiring radiating rf, for motors 3 x 100nf caps work well, one from each terminal to the case, and one across the terminals.

Hi,

On some of the points it could be quite a long way from the bridge to the points, unless, I locate the bridge near the (group of (max 4)) points and just extend the data cables from the Arduino to the H Bridge(s)?

My BIL has fitted some 'Pico' branded point 'motors' (solenoids) on the test track he provided for me and they come with a short length of cable (~25cm), so I could put any suppresion devices there at the closest. However, if I remote the H Bridges ...

I think the OO (HO) gauge locos come with some supression components and I'm not sure how such will interract with PWM power delivery?

Cheers, T i m
 
250mm is ok, over a meter is when rf might become an issue.

Might be worth having a look inside the loco's to see what suppresion they are using, these days it probably has been designed with pwm in mind, however if for some reason theres a big electrolytic cap in there it could well cause issues with pwm.
 
250mm is ok, over a meter is when rf might become an issue.

Ok thanks. OOI, if we do have to run some longer leads to the points (say 5m) would it be worth doing so in some twin core screened cable (with the screen earthed of course)?

Might be worth having a look inside the loco's to see what suppresion they are using, these days it probably has been designed with pwm in mind, however if for some reason theres a big electrolytic cap in there it could well cause issues with pwm.

Ok, I'll get my BIL to do that.

I have been working on the project today (with remote assistance from my mate on the programming) and we now have the shuttle routine working pretty reliably. In fact, the only real issue has been the train derailing on one of the points now and again but I've not noticed it was because it wasn't fully switched or anything.

I also managed to drive the H Bridge directly from the Arduino today. Previously I was driving it via a pair of SPDT relays as the program I used only had one output for the direction and the bridge required two inputs (the inverse of each other. I could have done it electronically but I wanted to leave it separate as then I can kill the output from the bridge completely).

Cheers, T i m
 
Screened probably isnt necessary, besides it tends to be less capable current wise, unless you use something expensive like sy.
100nf caps across the coils ought to suffice, if there is a lot of metal close to the coil a 100nf to it from each terminal will quiet things down.
Its better to kill the noise at source than shield it in the cable.
 
Screened probably isnt necessary, besides it tends to be less capable current wise, unless you use something expensive like sy.
100nf caps across the coils ought to suffice, if there is a lot of metal close to the coil a 100nf to it from each terminal will quiet things down.

Unfortunately there is not. Even the housing for the points is plastic.

Its better to kill the noise at source than shield it in the cable.

Understood. I can remember (back in the day) fitting an earthed steel can over the distributor (along with bonding the bonnet to the body) of my car so it minimised the noise on my PMR (and CB).

Cheers, T i m
 
Not having metal around the point motor isnt much a problem, it would be if there was and it wasnt connected, if any kind of voltage found its way to it through capacitive coupling then it would radiate it.

When my wifes landy had a petrol engine it had a collar around the dizzy, I thought it was for water protection till I saw it was connected to ground.
 
Not having metal around the point motor isnt much a problem, it would be if there was and it wasnt connected, if any kind of voltage found its way to it through capacitive coupling then it would radiate it.

Understood, thanks.

When my wifes landy had a petrol engine it had a collar around the dizzy, I thought it was for water protection till I saw it was connected to ground.

;-)

Cheers, T i m
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top