Design Challenge : relay spike protection

Mike,

The problem with just a diode is you only suppress half of the transient wave form.

I have tested relays that were diode protected, and I still get intense interference. In fact, the signals were so strong that it would disconnect my mouse from the computer it was connected to.

Cabe

Hi Cabe,

Could you give us more details as to how you are actuating this relay? Mike's diode recomendation is the textbook way to handle this problem and it's usually not serious you just have to make sure the diode can handle the current of the coil.

Perhaps, there is some detail that we are not aware that would make this solution fail, how are you measuring the waveform? Is the relay AC or DC actuated?

If you let us know we may be able to figure this out.

Best Regards,

Jorge Garcia

For the sake of example, I'll use the one I noticed this problem on:

24 VDC control voltage

I watched the transient wave form on an oscilloscope.

Cabe

Hi Cabe,

This is interesting. The textbook solution to this problem is to place a diode reverse parallel just like Mike mentioned and I know it works because I've used it myself. The waveform used has to be taken with a grain of salt for a few reasons. If you place a diode like mike has mentioned you won't get any ringing(maybe a little but nothing like the waveforms your seeing) because the diode will provide a path for the inductor current to dissipate hence the inductor voltage isn't going to sky rocket trying to maintain current flow.

I can't think of a way to test it, but I think that the waveforms your obtaining have something to do with your scope. Nothing inherently broken with your scope but just thinking about the front end of the scope. We generally model an ADC as a parallel resistor capacitor combination, when you connect this in parallel with the inductor you effectively obtain a resonant circuit which I think accounts for the waveforms your seeing.

I've included a PDF which I think should clear things up a bit, however the only way to know is to test the hypothesis and I can't think of any way to do that.

For sure, the solution to your problem is a diode in the arrangement previously mentioned it does work, none of your components will overheat or suffer any un due stress.  We just need to find a way to explain the results your getting.

hth,

Best Regards,

Jorge Garcia

Attachments:

Response for Cabe.pdf

One possible solution would be to alter the drive you are using to power the relay coil.

I can't see why the 'industry solution' of the reverse diode protection doesn't work for you, but hey ho.

I know you've seen the voltage oscillation on the oscilloscope, and i think that the diode will protect against the reverse current and the forward current will be so small (current, not voltage) as to be a non issue.

However, you could drive the relay using something like a ULN2003 (for more than one relay) or a darlington plus reverse diode for single relays - the darlington will sniff it's nose at any risidual current.

Finally, make sure you are not exceeding the rating of the diode switch contacts, and therefore setting up more current, and therefore a bigger magnetic field, than the relay is designed for.

One possible solution would be to alter the drive you are using to power the relay coil.

I can't see why the 'industry solution' of the reverse diode protection doesn't work for you, but hey ho.

I know you've seen the voltage oscillation on the oscilloscope, and i think that the diode will protect against the reverse current and the forward current will be so small (current, not voltage) as to be a non issue.

However, you could drive the relay using something like a ULN2003 (for more than one relay) or a darlington plus reverse diode for single relays - the darlington will sniff it's nose at any risidual current.

Finally, make sure you are not exceeding the rating of the diode switch contacts, and therefore setting up more current, and therefore a bigger magnetic field, than the relay is designed for.

Kenny,

When you disengage the relay, no matter what you drive it with, the coil deenergizes causing a massive spike larger than anything you planned for it. See the reply post I did on the subject. The documents attached explain everything.

Cabe

Well, yes, thats a whole load of theory there.

In practice, one needs to find a balance between cost and complexity - and find a solution which is cost effective, whilst at the same time providing a measured approach to protection.

I'm not denying that the research in these docs is valid and true, but one needs to design to a level which provides an acceptable level of protection, not necessarily FULL protection.

Hi Cabe,

Thanks for the articles, they were really informative. I do think some clarification is in order, as you can see the oscilloscope plots look nothing like the one you obtained so like enrico mentioned it's likely your probes were picking up some sort of interference.

As you can see in the oscilloscope plots and the articles provided both the reverse rectifier diode and reverse rectfier diode with zener suppress the transient voltage spike successfully. The disadvantage of using just the single diode is that it takes a long time for the CONTACTS not the coil to de-latch which can cause the tack soldering effect. The addition of the Zener diode forces the coil inductance to build up a greater voltage that wil increase the current flow once the diodes breakdown. This translates to a faster response from the contacts hence no welding effect.

I think the zener/rectifier solution is still a pretty cheap solution so I don't think it would be a big issue to use it as the standard suppresion method.

Thanks again for the articles.

Jorge Garcia

A friend of mine had a problem with relays a long time ago, and it was all due to the coil discharging. Adding a single diode helps, but only block half of the wave form. Since it ocillates in both the positive and negative. I am thinking this zener solution might at least clamp the issue.

Or just go with a transorb. The bidirectional one acts as two opposing zener diodes and block current in both directions. Where the text book solution of one diode only block one direction.

Picking a transorb with a low breakdown voltage might do the trick best.

Cabe