EMI interference with PID controllers

[ggltd]

On my Brutus I use controllers from Auber to control my natural gas valves, which are salvaged 24 V units from home heating. About 40 minutes into mash time, the readouts on the controller flash three times while turning the gas valve on and off at the same time. This continues randomly until I shut off the power to the gas valves. Plays havoc with mash temps. Tech support at Auber suggests this is due to EMI coming from the solenoids on the gas valves which makes sense. Anyone have this problem, and any suggestions to fix? I am thinking to first try and put some magnet cores on the 24 V wires at the gas valves, and maybe also in the box around the temp probes going to the controllers.

[bigdan]

You can add a diode as a rectifier for the solenoid and it should fix the problem. I would guess that a 1N4001 would be sized big enough. Just wire the diode parallel with the solenoid. Hope this helps

[ggltd]

Thanks BigDan. I will try this as it seems easier than other ideas I have.

[ajdelange]

If you want to try DC I would recommend a bridge rather than just 1 diode. With just one diode the coil is only energized for half the AC cycle and may chatter or buzz. If you do go to full wave also put a back biased diode across the coil. When denergized the collapsing field will be discharged through this diode preventing spikes on the lines. You could also source the power for the PID controllers through EMI filtering power modules. This may be enough by itself to solve the problem.

[bigdan]

I guess I said rectifier, but should have said biased diode across the solenoid. I haven't had problems with small solenoids and a single back biased diode across the coil. I am guessing that your valve is AC and you don't need to use any kind of rectifier.

[ggltd]

Yes this would be 24 V AC to the gas valves. I stopped at Radio Shack last night and they had the part number that you mentioned. I will solder this into the circuit and test it out. I also got some magnets with holes and I think I will put these on the wires going into the back of the PID controllers and see if that also helps. I am still trying to figure out if this is coming back through the 24 V AC lines out of the PID controller, or this is finding it's way somehow through the wire from the K type sealed thermocouple back to the PID. The wire is really small gauge with the mesh braid covering.

[ajdelange]

I'm not understanding something here. A diode across a coil fed with AC is almost a dead short across the source during 1 half cycle. This can't be good for the 24V source or the diode. Also no current would flow through the coil on those half cycles when the diode is forward biased (except very briefly as the field discharges).

What I suggested is a bridge to convert the 24 VAC to 24 VDC and then a diode across the coil back biased to that 24 VDC. It would only conduct briefly after the coil is de-energized as the field collapsed dissipating the stored energy in the diode rather than in a spark across relay contacts.
Ferrites might work too.

Annother approach might be to switch power to the solenoids through SSR's controlled by the TTL levels out of your controllers. Not much noise generated in an arrangement like that.

[codewritinfool]

AJ is giving you good advice, and is suggesting this:



Sorry for the quick MSPaint artwork.