Measuring back emf to determine correct diode size?

[cygnusv]

I got a lot of help last year from you lot when I was having nightmares with my Johnson 42PSI 5GPM water pump. The problem was with the pressure switch not the pump itself.

Following advice from here I fitted an external (Whale) pressure switch with a new microswitch inside, a 24 volt relay and put a diode across the pos / neg terminals (ringed end to positive. This has worked great for a year, but recently failed again.

I replaced the relay and micro switch a couple of days ago, and when the relay (Lucas) started getting warm I replaced the diode too. Now it seems to be working okay.

Can I measure the back emf that was killing the micro switch or if this is not easy, is it feasible to fit a 2nd diode?

Many thanks for any help forthcoming, Stu

 

[Bilgediver]

I got a lot of help last year from you lot when I was having nightmares with my Johnson 42PSI 5GPM water pump. The problem was with the pressure switch not the pump itself.

Following advice from here I fitted an external (Whale) pressure switch with a new microswitch inside, a 24 volt relay and put a diode across the pos / neg terminals (ringed end to positive. This has worked great for a year, but recently failed again.

I replaced the relay and micro switch a couple of days ago, and when the relay (Lucas) started getting warm I replaced the diode too. Now it seems to be working okay.

Can I measure the back emf that was killing the micro switch or if this is not easy, is it feasible to fit a 2nd diode?

Many thanks for any help forthcoming, Stu

I think you would need an oscilloscope to see the voltage spikes but could be hundreds of volts. Maybe your diode only rated to 24 volts????
Here is a discussion mentioning a second diode.

https://www.element14.com/community/thread/3942

 

[nigelmercier]

Is this to quench the back EMF on the relay? Use a 1N4001.

 

[ReefMagnet]

Rule of thumb is 10x the operating Voltage. This would make it a 1N4004 (400V) diode

 

[William_H]

Diodes are so cheap that if you think the diode is being damaged by the spikes then just fit a diode with higher ratings.
Interesting that discussion talks about both pos and neg spikes. I can't see how both would develop but apparently so.
The old post office relays used to come in a variety with a big copper slug at one end of the coil. This was mainly intended to slow the relay down both in operating and releasing. (half a second or so) it occurred to me that if you can gain access to the relay coil you might be able to fit a shorted turn of copper wire. This would I would think dampen the back EMF dramatically. good luck olewill

 

[ReefMagnet]

A diode is typically placed across the relay coil in parallel. The diode is connected "in reverse" - i.e. the end with the white stripe is connected to the +ve terminal of the relay coil. In this application the diode is known, amongst other descriptions, as a "freewheeling diode". Usually these are used to protect spike sensitive electronics on the relay coil circuit, not to stop arcing across a switch. I'd do a quick double check that the micro switch is rated for 24VDC and the relay circuit current also. Any less and it may open/close too slow or not have enough contact gap in the open position to resist arcing in operation.

 

[nimbusgb]

Bilgediver answered your question in post 2, everyone else just added noise not actually answering the question but that seems normal for forums.

To expand on Bd's answer a storage 'scope triggered on a falling edge set say 100% below 24v supply will capture the pulse of energy being reflected up the line. It could well be several hundred volts or more but with a 10:1 probe you should be able to see most of it.

View attachment 30909

In this diagram several traces and circuits are compared.

View attachment 30908

 

[nigelmercier]

Rule of thumb is 10x the operating Voltage. This would make it a 1N4004 (400V) diode

You are quite right, I was referring to the 1N400x series, but too busy at the time to look it up.

With respect, I think the zener diode is only needed in high power circuits.

 

[ReefMagnet]

You are quite right, I was referring to the 1N400x series, but too busy at the time to look it up.

With respect, I think the zener diode is only needed in high power circuits.

No, not a Zener. The diode is wired in "backwards" because the back EMF current flows in the opposite direction to the power supply current creating an inverse polarity. The diode conducts the inverse voltage, thereby shorting out the back EMF spike.

nimbusgb I appreciate what you are saying but...

I would hazard a guess that the OP is very unlikely to have a digital storage oscilliscope floating around in his (or her) tool box and is even less likely to know how to use one.

And besides, how else can us noobs hope to hit 4 figure + post counts if we don't contribute irrelevant and otherwise useless information to threads?

 

[nigelmercier]

No, not a Zener...

I was referring to the posts that advocated the use of two diodes, one of them a zener.

 

[ReefMagnet]

I was referring to the posts that advocated the use of two diodes, one of them a zener.

Ahh ok, didn't see that...

 

[cygnusv]

A diode is typically placed across the relay coil in parallel. The diode is connected "in reverse" - i.e. the end with the white stripe is connected to the +ve terminal of the relay coil. In this application the diode is known, amongst other descriptions, as a "freewheeling diode". Usually these are used to protect spike sensitive electronics on the relay coil circuit, not to stop arcing across a switch. I'd do a quick double check that the micro switch is rated for 24VDC and the relay circuit current also. Any less and it may open/close too slow or not have enough contact gap in the open position to resist arcing in operation.

Do they do a microswitch rated for 24 volt that is 10.3mm wide? I seem to remember they were rated for 250VAC, though I could be wrong. Where would be a good place to look for a pukka 24VDC switch. The relay is definitely a 24VDC Lucas.

Just got the diodes out and they appear to be n4007 rated at 1000Volts? does the fact that our pump has DC voltage make a difference?

 

[lw395]

The diodes suppress the sparks/back emf on the coil side of the relay.
The switch side of it is probably what is getting hot, due to the arcing from switching off the motor.
Also, the salt/wet air in a boat often causes relays to be unreliable in my experience, once you get a little corrosion on the contacts, they get hotter and corrode more until they fail.
A small capacitor across the contacts, or another diode may help, but a plentiful supply of free relays from scrap cars is an option?

 

[savageseadog]

One point to bear in mind is that it does matter where the suppression is fitted even though fitting across the switch or load might look the same. It's better to have the suppression nearest the motor or switched inductance because the spike current passes through the minimum length of wire and reduces the amount radiated in the way of EM interference.

 

[VicMallows]

Rule of thumb is 10x the operating Voltage. This would make it a 1N4004 (400V) diode

Can you still get any of the 1N400x series other than the ubiquitous 1N4007 ? (yes, I remember when they cost so much you purposefully chose the lowest voltage rating you could get away with).

 

[JohnGC]

I got a lot of help last year from you lot when I was having nightmares with my Johnson 42PSI 5GPM water pump. The problem was with the pressure switch not the pump itself.

Following advice from here I fitted an external (Whale) pressure switch with a new microswitch inside, a 24 volt relay and put a diode across the pos / neg terminals (ringed end to positive. This has worked great for a year, but recently failed again.

I replaced the relay and micro switch a couple of days ago, and when the relay (Lucas) started getting warm I replaced the diode too. Now it seems to be working okay.

Can I measure the back emf that was killing the micro switch or if this is not easy, is it feasible to fit a 2nd diode?

Many thanks for any help forthcoming, Stu

I'm can't quite decide from your description if the diode is protecting the relay contacts or the switch contacts. You should protect both. You don't say what sort of diode you fitted but it is more likely that it failed from trying to dissipate too much power than anything to do with overvoltage.

The best location for the diodes is close to the motor connections and close to the relay coil connections. That way the circulating currents (which can be several amps) flow around the shortest circuit and radiate a lot less.

The reverse voltage withstand value (EG 700V for a 1N4007) need not be very high. When the switch is closed the diode is reversed biased with 24V applied from your batteries. That nominal voltage will increase during charging and there will be some small spikes but a 100V device should be adequate. With a diode in place the high voltages are not generated because when the switch opens the voltage across the diode reverses, so it is forward biased; the voltage is limited to the diodes forward value which will be a couple of volts only. That is the point of fitting the freewheel diode.

Assuming the diode you fitted had failed, then I think it is likely that it failed due to over current not over voltage. The pump in particular will generate a lot of energy because the motor continues to spin after you remove the supply plus the usual effect of a collapsing magnetic field. The diode must conduct that current until all of the residual energy has been dissipated. A 1N4007 is rated at 30A for 8.3ms once every few seconds; you may think that is plenty but it's not.

Relay contacts and micro-switches bounce. IE the contacts close, then rebound and then open then close again and so on for several 10ms and several 10 times. During each closure the magnetic field builds up in the coil and more energy has to dissipated the next time it opens. So the diode has little chance to recover (which actually means time to cool down).

So, you can probably protect your micro-switch from the relay coil with a 1N4002 or better, but you probably need a diode with a higher current handling (and power dissipation) rating for the motor. Ideally the diodes should be fast recovery types although that's not as important as the power rating.

Specifying freewheel diodes isn't easy as the calculations are quite complex and mechanical switches make it harder still as there is so much variation in the way they operate. You should definitely fit a diode across the relay coil and the motor. They should probably both be capable of handling more current. Reverse voltage is important but don't over look the current.

Measuring back EMF with a scope isn't simple either; you can't just connect a standard scope lead and see what is actually happening because the scope ground lead is also inductive and has a surprisingly big effect.

 

[nigelmercier]

Is this to quench the back EMF on the relay? ...

The diodes suppress the sparks/back emf on the coil side of the relay.
The switch side of it is probably what is getting hot, due to the arcing from switching off the motor...

Which is why I asked the above question right at the beginning. It wasn't clear in the OP which pair of contacts were burning up.

 

[William_H]

And besides, how else can us noobs hope to hit 4 figure + post counts if we don't contribute irrelevant and otherwise useless information to threads?

Do you really want to aspire to 4 figure post counts? It might seem something to be proud of but on the other hand it might just reflect an old git with no more satisfaction in life than to tell people things which may or may not be true or relevant. ie a way to relive the past glories of sailing or anything else. Not that there is anything wrong with that or for that matter growing old.
But I would rather be a noobe at sailing or life in general. And besides irrelevance is not such a bad thing. Often it is not just the OP who reads answers which may be right out of the box. good luck ole ole ole will. Yes another pointless post and count...

 

[cygnusv]

Which is why I asked the above question right at the beginning. It wasn't clear in the OP which pair of contacts were burning up.

Hi Nigel

I'm a bit in the dark myself. When I first started this 'experiment' I fitted the external Whale adjustable switch. The micro switches (2) failed very rapidly. Someone then suggested a relay. Now this made sense to me. Low current across the micro switch with the relay carrying the real load. When the relay failed fairly quickly it was then suggested a diode across the relay coil should do the trick. And actually it has worked for about a year.

It all went wrong again last week so I ordered the bits, including what I believed to be a better quality 20/30 amp Lucas relay, and changed everything apart from the diode. I noticed the relay getting warm, not hot, so changed the diode. The arrangement is still sticking on occasionally and blowing the 3BAR PRV

Do I understand correctly that I should also use a diode across the pump? Presumably I can't do much with the micro switch as it's only breaking a positive wire.

Hope this sounds reasonably clear. Thanks, Stu

 

[nigelmercier]

I'd put a similar diode across the relay contacts, or the pump. However, this should not be necessary if the relay contacts are rated for DC and it has the correct coil voltage. Do you have a specification for it?