The background here is that we have a control board in production which, among many other things, switches relays on and off to control 1/3HP motors. We get all sorts of noise from this switching and I've been learning all I can about it lately. We made huge improvements in the switching noise with improved grounding and adding snubbers to the motors, but there's still a mysterious burst of noise when the motor stops. I thought the motor snubber would do the trick, but later realized it could be the relay's EMF we're catching. When I looked at the triac driver circuit which controls the relay, I was surprised to find no indication of snubbing, filtering, etc. other than a varistor preventing over-voltage situations.

The interesting thing is that there's a capacitor in the schematic, C13, which seems like it could possibly be there for noise reduction, and which is not included in our boards. The BOM makes no mention of it. I'm assuming the "DNI" next to the specs is for "Do Not Install?" And I'm assuming 1000p is 1000pF (1nF.) Would this capacitor, if installed, serve as a snubber, working in conjunction with R7 perhaps? Or would it serve some other purpose? Any guess as to why it would be drawn into the schematic and then later abandoned? I've tried to find info on how to spec snubbers, what values to choose in different situations, but I'm having a hard time finding anything that seems relevant to motors, relay coils, etc. Would C13 be an appropriate value to be snubbing the relay?

I'm really tempted to pick up a few capacitors and try soldering them in on a few boards to see if it reduces relay switching noise (there are pads on the board and everything - just not populated!) But if this has nothing to do with relay noise, or if there's some reason it might be harmful to the circuit, I definitely don't want to mess anything up.

The input to the opto is 5VDC. The relay power is 200-240VAC, with one leg always connected to the coil, and the other leg switched through the triac circuit in question. The schematic for this part of the circuit is below, along with a few images to help identify the relays in question:

 

C13 is in position to be a snubber but R7 is usually less resistance than 180 ohms. The value of R7 is related to instantaneous current through the triac when it first comes on and discharges the capacitor, whatever voltage the sine wave is, at that instant. (Theoretically, the capacitor current approaches infinity when the triac first comes on.)

Anyway, you have the holes in the board and I say 1000pf won't hurt anything. It's a rather small amount of capacitance, but a relay is a rather small load current. The RC value of R7 and C13 is in the nanosecond range, so I don't expect that to cause a significant delay in firing the triac.

Why not install it?
1) to save a penny.
2) maybe it doesn't fix anything.

Watch your scope. One experimental result is worth a dozen theories.

 

Looks a conventional set-up. As I understand it, R7/C13 are there to limit the rate of change of voltage applied to the triac gate, to reduce the chance of false triggering.

 

Yes, it is a snubber, but to keep the triac from turning on with noise or power on from a switch. Not likely your problem. Many modern triacs don't need them.
http://www.onsemi.com/pub_link/Collateral/AN1048-D.PDF

 

Well, my boss managed to get in touch with the original designer, who confirmed what Alec_t and ronv said. At the same time, I tried a different experiment that I should've thought of sooner, and determined that the relay coil isn't the noise source anyway. (Disconnected motor leads from relay, then operated machine to see what noise relay coil generated as opposed to what noise was from motor, and found no significant noise from relay.)

Thank you all for your help, and sorry I wasted your time on this one. I'll be back shortly with a new post that more directly addresses where we're actually seeing the effects of the noise problems, which is in a handheld interface with a 4x20 character LCD.

 

I tried a different experiment that I should've thought of sooner, and determined that the relay coil isn't the noise source anyway.

Good job! I'm glad your experiment provided useful results.

 

Thanks! I'm a fan of your approach, preferring experimentation to guesswork whenever possible. I do check in first sometimes when I fear my experiments run the risk of frying hundreds of dollars worth of gear at a time. Generally though, I like to dive right in.