Noise filtering problem

[tomocar]

Hello all,

I've got a tricky one for you - does anyone have experience with solving this kind of problem:

I'm using a JLD633 as the main controller and a JLD7100 as a redundant high limiter. This is an electrically noisy environment - we are using a high-powered SCR to trim the voltage going to the heating elements so there's a lot of 120 Hz buzz on the power lines. Also our thermocouple wires are 50 feet long and they are the common type - not twisted or shielded.

I was hoping the digital filtering built in to the controllers would take care of any noise problem we might encounter. And it appears to be working well on the JLD7100 but not on the JLD633 (both are set on max filt at 3). The PV display on the 7100 holds steady as the power cycles on and off, but the PV on the 633 jumps around when the power cycles on. In fact it jumps down a few degrees, then back up to the correct reading as the power cycles off again. This is causing false PID settings when trying to run auto-tune resulting in sizable over-shoot. It sets a very short CT (1 second) and a very high SF (1963). I know those values can't be right.

Any ideas on how we can further filter or reduce the noise on the input? Or can I manually adjust certain PID parameters such that it will not be affected hardly at all by these brief dips in apparent PV caused by the noise?

But another question, for you, Marco, is: do you have any ideas why the 7100 filter seems to be filtering out our noise but the 633 is not?

Thanks to anyone willing to offer me some advice!

-Tom C in Eugene, OR (www.samuelartglass.com)

[tomocar]

correction: I meant JLD634

[richiem]

An AC line filter -- a combination of inductors and capacitors in a metal box -- for the controller should help a lot. Generally, these look like squarish bricks a few inches in size, with AC inputs and output terminals and a case ground -- hot and neutral to the AC in and out, the case connected to earth ground/power ground. They are readily available, and you don't need high current rating -- 1 Amp will be plenty. Check eBay for a start, or Allied Electronics, etc.

Of course, it will be much more effective if you put it and the controller in a metal box that is also at power ground.

[tomocar]

Good idea Rich - thanks.
I was guessing that the noise was coming in on the thermocouple wires and didn't consider that it might be coming in on the power supply, but you are right that it might well be.

I'll give it a try and let you know what happens.

Tom