Can't eliminate random signal spiking from HP 5890 (GC/FID)

[wgehli1]

I am getting random peak spikes in my signal (both positive and negative spikes), and I am out of ideas to troubleshoot the problem. Is there anyone out there that can give me some more ideas to fix my signal?

My setup includes an HP 5890 GC equipped with an FID. I have a cable attached to the signal channel out port at one end and attached to a data logger at the other end (minilab 1008). I am running the data logger in strip chart mode using the appropriate settings and connections (+/- 10V, single-ended configuration, ground pin connected to ground and detector+ pin connected to channel in on the board).

I have eliminated many possible sources of the problem. For example(s), I have eliminated the data logger by making a jumper from the channel in to the ground and there were no spikes in the signal. It just gave the normal zero signal. I attached the cable to the data logger (without it being attached to the FID) and saw no signal spikes. Again, it was just the normal zero signal. It is only when I connect the cable to the FID that I start to to see the spikes, even when the instrument is off and unplugged. I have tried both signal channel out connections, and I even changed the main board (there was an extra one lying around in the lab), thinking it might be both signal out connections that were loose, but I still see the signal spikes. The only thing I have yet to do is replace the FID itself. Is that the only solution, or am I missing something more obvious?

The funny part is this just started happening this week. Before this week, everything was working normally, and I initially thought this was just the result from a dirty detector. So, I completely cleaned the FID (hardly anything there), inlet, everything, in addition to baking out anything in the column (which should not be the case since I am only using low molecular weight gas products). It wasn't until after I had done the cleaning that I finally started to check the electronics and still currently believe that to be the problem.

Thanks for taking the time to at least hear my plea. This is my first time working on older equipment, so I can't exactly call tech help. Any help whatsoever would be greatly appreciated.

[Peter Apps]

It is only when I connect the cable to the FID that I start to to see the spikes, even when the instrument is off and unplugged.
The funny part is this just started happening this week.

Welcome to the forum.

Spikes from an instrument that is unplugged are indeed a puzzle. That one thing eliminates a whole host of sensible possibilities, so we can jump straight to the highly improbable.

What happens if you disconnect all the (presumably metal) gas supply lines form the GC ? Do any of the gas lines pass near anything that could generate electrical interference ?, and are any of them earthed anywhere ?

Peter

[wgehli1]

Peter, you solved it!

I unplugged everything electrical in the vicinity, and the signal spikes were completely eliminated! Now, it's just a matter of figuring out exactly which one caused the interference. I will go back and double check all the electrical and plumbing to see where anything might possibly be touching or not properly grounded. wow....just wow. Thank you so much!

[AICMM]

Just a bit to add since Peter has Id'd your problem. Sounds to me like you have a ground loop going. The computer or the board are on a different ground than the instrument. You might try grounding the data board to the top of the GC with a little wire and an alligator clip. You would also want to verify the computer and the data board have the same ground.

Some things to think about.

Best regards,

AICMM

[James_Ball]

Also try testing the voltage from your neutral to you ground prongs at your plugs. If you have a bad ground no the building electrical you will have a voltage across those two prongs of the plug. Just put the ground end of a volt meter into the ground prong(hole) of the outlet and test the other two. One should give you the supply voltage and the other should give zero volts, if not have an electrician check out the grounding in your service panels and electrical meter outside.

[wgehli1]

Thanks James and AICMM for your suggestions.

The cause of this was a short in a wire running from a reactor heater I am currently using in a specialized setup I have prior to the GC that was giving me problems. I ended up replacing the wires for that particular heater, made sure they were not touching anything, and I now have no problems whatsoever.

[Peter Apps]

Result ! Thanks for the feedback.

Peter