Agilent FID detector flame out

[paddypower]

Hi All
I hope you can help.My problem is with an Agilent FID detector.It lites perfectly and at the same point circa 2.5 mins during a run it flames out and then reignites and carries on to complete the run.Out of 20 samples it could flame out on 4 occasions.
I am using a DB 624 30 M column ,split injection 2;1 and a G1888 headspace sampler.I initially thought it was the electrometer or the board but i have eliminated those.
I have capped off the detector and it runs perfectly no flameout during a 20 sample run
The headspace and inlet set up are typical of what i have always been using so i ned to ideas as to what it can be
All help greatly appreciated

[rb6banjo]

Have you made sure that your H2, air and makeup gas flows are correct? Don't rely on EPC if you have it. Measure them with a flow meter. What is the nature of your samples? Are they aqueous? Perhaps the water coming through in your headspace injections is enough to blow it out? I've seen that happen when doing liquid injections of water-based samples.

[dblux_]

... All help greatly appreciated

H2 flow = ?
Air flow = ?
make-up flow = ?
carrier gas flow = ?
lit off for detector = ?
to start with

[Peter Apps]

What are you analysing, what are the headspace conditions, and what is the column temperature programme ?.

Let me guess; aqueous matrix, temperature near the BP of water, column starts at around 40C.

You are putting 1/3 ml (split is 2:1) of steam onto the column, it condenses to liquid water which blows along the column until it flashes back to steam at the hot detector and blows the flame out.

Peter

[Consumer Products Guy]

I'm retired now so don't know exactly what it was termed, but on the FID window of the Chemstation there was a certain threshold that if the signal dropped below this, the detector would wig out and start relighting during a run. The default setting/level was 2.0, and we changed it to 0.5 for all methods on that GC (6890) and saved the methods, and then there was never again such problem.

[dblux_]

I'm retired now so don't know exactly what it was termed, but on the FID window of the Chemstation there was a certain threshold that if the signal dropped below this, the detector would wig out and start relighting during a run. The default setting/level was 2.0, and we changed it to 0.5 for all methods on that GC (6890) and saved the methods, and then there was never again such problem.

Didn't I already ask for "lit off" ?
----------------
Edit: In the Lab again, it's called "lit offset".