Hi, quick question about relays.

I have some 24V, 1.8k relays with a LED in series with 2k2 across the coils.
There are no snubbers on the relays. The LEDs seem to fail intermittently, and almost always on the same relays. Is it reasonable to think that it's the kickback EMF that kills the LEDs? How would you dimension an RC snubber for this application (i'd prefer that to a diode)?

Thanks!

 

Whatever capacitors you do use, it has to be non-polarized, and ceramic dielectric or better, and have a voltage rating far beyond your circuit operating voltages (probably 100V for your 24V circuit). Since your snubber is not a very "serious" snubber, you can probably get away with using ceramic capacitors.

SIzing is kind of a black art...see PDFs. In summary, a large capacitor will snub better, but will require a higher power resistor to dissipate the energy. Yours is a relay application so it's not really high speed like most circuits that need snubbers so power dissipation is probably not a very big problem for you (and that's where most of the tradeoff for snubbers comes in aside from cost). Just try regular values...0.1uF x7R 100V capacitor with a 1 carbon composition ohm resistor.

Inductive spikes usually harm the switch trying to disconnect the current and for relay. For a relay, this usually means the switch driving the relay coil, not the primary contact since it is basically a piece of wire...not much inductance there. I don't see how sparking across the relay contacts would damage the LED though. The voltage spike appears across the switch, not anywhere else. THe parts are cheap and common though so it's worth a try.

Attached Files

	Snubber Design.pdf (250.3 KB, 30 views)
	Snubber Theory, Design, & Application.pdf (1.93 MB, 27 views)

__________________
Tanaka Sensei (avatar) says: Please spell it "ridiculous" correctly! Not "rediculous". ^^

 

Thank you for your reply!

My reasoning is that when the relay is disengaged, a large reverse voltage appears across the coil. Since I have a LED across the coil (not on any of the relay's contacts) this spike could somehow damage the LED.

Do you think it's reasonable to blame the relay/EMF for the destruction of the LEDs? I can't see much else that could damage the LEDs.

Of course, the snubber would have the added advantage of protecting the contacts of the relay.

Is there a difference in using plastic vs ceramic capacitors? Do you recommend a resistor as low as 1 ohm?

 

Wait, is the LED in series with the primary contacts of the relay? Or the coil? I was under the impression the relay was being used to switch with LEDs (which I found a bit odd). But even then, the voltage spike generated by the inductance appears across the switch, not the LED. THat's not to say that I am not missing something though. I think a schematic is in order.

I think plastic is just the higher end polyester or polypropylene capacitors. YOu don't need them though. As far as this is concerned it's not a very demanding snubber application.

Your power level is really low because it is not high frequency, which is why I recommended the smallest resistor you can get and the largest capacitor you can get. Trial and error from there.

__________________
Tanaka Sensei (avatar) says: Please spell it "ridiculous" correctly! Not "rediculous". ^^

 

I attached a quick schematic to show you how it could be hooked up. Basically, it's relays driving other relays in various constellations.
There are LEDs across many relays to display their state. The 24V supplies are of course hooked together, so in some sense you could say that the LEDs are in series with the relay contacts, as well as being in parallel with the coils.

Upon seeing this, may I ask you for your reaction and where you feel the snubbers should go?

Attached Thumbnails

 

 

Hmm, I guess I have been misreading your descriptions of your circuit the entire time. The LED is in parallel with the coil, not in series with the coil or the primary relay contacts.

Yes, that would definately do it. The voltage spike appears across the inductor...and what's in parallel with the inductor? The freakin' LEDs! As we all know, things in parallel all have the same voltage. The inductive kickback produces a voltage spike across the inductor. This voltage spike also appears across the the LED-resistor (because they are parallel with the inductor) and applies a reverse voltage to the LED causing it to breakdown.

I doubt an RC snubber in parallel with the relay coil will be enough (it only dampens and slows the spike it does not clamp it). What you need is clamping because the LED is so sensitive to reverse breakdown. Use a fast diode (like a schottky) in parallel with the coil- it will clamp the reverse voltage across the LED to less than 1V which should be enough for the LED to block.

Or a better method that makes it so the LED does not have to block any voltage would be to actually place a diode (that is fast with a reverse breakdown voltage high enough to survive the voltage spike) in series with the LED. It will do most of the blocking when the inductive spike applies the reverse voltage so that the LED does not have to.

__________________
Tanaka Sensei (avatar) says: Please spell it "ridiculous" correctly! Not "rediculous". ^^

 

Yes, you need to stick a reverse connected diode directly across the relay coil - the high back-EMF is destroying the LED's.

This is a VERY standard requirement, and should be fitted to all relays involved with any electronics.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

OK, seems free-wheeling schottky diodes is the way to go here

What reverse-voltage should I go for? Do I need the diode to cope with several times the 24V in the reverse direction?

Here is one I found that's reasonably priced, could it work?
11DQ06 schottky diode 1.1A 60V DO41

 

I don't see as you need schottky at all, simple 1N4001 or similar should be perfectly fine.

Don't you have any rectifiers knocking about?.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Don't I need something that reacts faster than the LED (e.g. a schottky)? Would a normal diode be able to suppress the flyback before the LED is damaged?

 

Yes - the 'speed' of a schottky is basically when switching OFF, not switching ON. That's why they are called 'fast recovery' devices.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

OK, thanks for the feedback! 1N400x would certainly be easier to find

I spent some time reading Schottky datasheets here, and it seems many of them are recommended for flyback use. What is the characteristic that makes them suitable for that, if speed is not the issue? Also, it seems that Schottkys rarely come with reverse-voltage capacity of over 60V or so, and I guess for a 24V relay you would want something capable of at least 100-200V (again making 400x suitable from that standpoint)?

 

Like I said above, it's the 'recovery' speed of the diode that makes it 'fast', and it's that fast recovery that makes it suitable for flyback purposes.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

x400 operating voltage is suitable. It also depends on the current level just before shutting down. There are other diodes that have higher breakdown ratings that are not schottky that are still fast. Schottky just has some nice characteristics too. Just in case you can't find the schottky diode you need.

You could probably just use a 1N4448.

__________________
Tanaka Sensei (avatar) says: Please spell it "ridiculous" correctly! Not "rediculous". ^^

 

lower forward drop means less power dissipation and a higher regulator efficiency

the diode shouldn't see anything more than 24V - or perhaps double that if there is some inductive resonance on the input rail somewhere. It'll see 24V when the relay is on and just the forward drop when the relay is shut off and it is free wheeling.