Troubleshooting an FID

[karwaldo]

I use an Agilent 7890B to analyze alcohols in gasoline. I use N2 as a carrier and detection is via an FID.
The GC is not quite a year in operation, but in the last couple of months the FID output signal has started to fade. Up until lately the FID baseline has been around 10pA, now it's around 1.1. During a column back flush period, when no sample is going to the FID, the signal drops below 1.0 and the unit thinks the FID flame is out and tries to relight. The only other symptom is ethanol QC checks are coming in low on a fresh cal, (88-92% recovery). I think I may have a carrier leak? or flows aren't what the PCM think they are?
When the unit is not in service I manually shut down the flame but I keep N2 flow going and temps at normal. Good or bad?
The only config change is I replaced the Agilent block style gas filter system with in-line filters. The block filters were too short lived and they always leaked. Other than that it's been regular PM, liner, oring, gold-seal, and periodically I clean the FID jet.
Any suggestions?

[HPLC chemist]

Did you check the pressure in the Hydrogen tank and the regulator?

[rb6banjo]

I too turn off the support gases when not operating. Never had a problem operating that way (25+ years).

I would get out my manual flow meter and actually measure my support gas and makeup gas flows to see how well my measurement compares to what I dial into the EPC. If the flows don't check out, perhaps you have a faulty EPC module.

Your low recovery for EtOH on your standard could point to an inlet problem. The fact that you clean your detector regularly takes away one of my suggestions. Constant relighting of the detector has usually been solved by a good cleaning for me.

I'll keep thinking............

[Peter Apps]

Welcome to the forum.

In addition to what rb6banjo said about checking flows -

Is your method set to autozero before each run ? What happens if you autozero ?

What is the source of your hydrogen, air and makeup gas (if you use makeup) - cylinders, generators, compressor ?

You mention column backflush - what additional plumbing do you have to accomplish that ?

Peter

[tkubowicz]

Hello

Check "Lit Offset" value.

Do as below:

1.Click "config" button than "Front or Back Detector" button
2.Scroll down to "Lir Offset"
3.Default value is 2.0 - change to 1.0 and click "OK"
4.Try to run sample ro see if FID signal is stable

Note: sometimes if Makeup Gas flow is high it can blow FID flame during run (especially if column flow is changing during run)

Good luck

Regards

Tomasz Kubowicz

[dblux_]

...
3.Default value is 2.0 - change to 1.0 and click "OK"
...

Or even lower because karwaldo says "signal drops below 1.0".
I would check if capillary adapter (the one in the oven under insulating cup) is fastened sufficiently tight.

[tkubowicz]

Hello

I would check if capillary adapter (the one in the oven under insulating cup) is fastened sufficiently tight.

as I remember 7890B does not have capillary adapter for FID (it capillary optimized)

Regards

Tomasz Kubowicz

[dblux_]

7890B FID is available in two versions: capillary column optimized or adaptable for either packed or capillary columns. (Agilent 5991-1436EN)

[karwaldo]

Thanks everyone for the advice.
I also notice when the FID relites, the N2 flow is temporarily shut off. As as the N2 flow ramps back up the signal drops from around 5mA to 1.1, or lower, and sometimes it extinguishes the flame.

The back-flush is a column switching valve.

The LIR offset is set at 1. I even tried turning it down to .8 without success. I use a chrome plated surface to look for condensate. The flame really is going out.

Support gases are from UHP cylinders. I have a two-stage regulator with the output pressure set at 60psi. cylinders are next to the GC so only about 30" of 1/8" copper tubing. None of that has changed since the original install.

Looks like the next step is to verifying flows. I was going there next. Just trying to avoid any unnecessary disassembly. With this unit only being online for a year I would have suspected a faulty PCM last.

[rb6banjo]

What is the flow of your makeup gas? It doesn't need to be ridiculously high. You could try to back it down 1 mL/min at a time until it stays lit. Then, check your chromatography to see if it affected it adversely or not.

[dblux_]

What is:
column diameter
carrier flow
hydrogen flow
air flow
makeup flow (already asked by rb6banjo)

[Peter Apps]

Thanks everyone for the advice.
I also notice when the FID relites, the N2 flow is temporarily shut off. As as the N2 flow ramps back up the signal drops from around 5mA to 1.1, or lower, and sometimes it extinguishes the flame.

Looks like the next step is to verifying flows. I was going there next. Just trying to avoid any unnecessary disassembly. With this unit only being online for a year I would have suspected a faulty PCM last.

You do not need to disassemble anything to measure flows - measure them at the outlet of the FID.

It certainly sounds as if you have a problem with the nitrogen flow - it is normal for it to switch off to ignite the flame, but it should not extinguish the flame as it ramps back up again.

Peter

[dblux_]

Real, measured flows is further step.
For now we still don't know if he set sensible flow rates with EPC.
And we don't know if the jet orifice isn't substantially clogged, what may cause similar problems.

[karwaldo]

What is:
column diameter
carrier flow
hydrogen flow
air flow
makeup flow (already asked by rb6banjo)

Column Diam .53
Carrier in col 22.7
H2: 30
Air: 450
Constant fuel and M/U 27

[karwaldo]

Real, measured flows is further step.
For now we still don't know if he set sensible flow rates with EPC.
And we don't know if the jet orifice isn't substantially clogged, what may cause similar problems.

Jet orifice is clean. Also checked column ends, they are clear and clean as well.