Flow instability after GC6890 move

[Johnny Rod]

We moved our 6890 with 5793 and Markes TD system to our new lab. We paid Agilent to do this, though too much time passed between the move and discovering a problem, so now I'm trying to sort it out.

I noticed instability in the total inlet flow, it should have been about 14.6ml/min (the TD uses 10ml/min) but was randomly varying by 0.1-0.2ml/min each way. I disconnected the TD part and it is now a standalone GC again, and the problem is still there. The setup is as follows:

6890 with EPC inlet, column out to Deans splitter at constant 4psi (auxillary EPC), from there to FID and MSD
RT Q-bond PLOT 30m x 320um with particle trap
1.6ml/min constant flow, which is 12psi at 40C (start temp), splitless inlet setting
Total inlet flow 4.6ml/min (1.6 plus 3 for septum purge, unadjustable)

I leak checked the inlet and zeroed the flow and pressure sensors. I tried constant pressure mode, and split (10:1), the former was a little better but not much. I tried this on another 6890 and there was no problem. I also had some pre-move data where I logged flows and this looked fine. I disconnected the column tail from the splitter so it was at a constant pressure (atmospheric) but it made no difference.

It looks like poor EPC control but I don't know if there is anything I've missed or could try. I could try swapping the EPCs from the other 6890 but it is in use so I'd really need to want to. As always I'd very much appreciate your thoughts!
John

[Peter Apps]

I don't think that you can have a total inlet flow of 4.6 ml/min with all of it going to either the column or the septum purge. There has to be some flow out through the split for the back-pressure regulator to work with. If you set a split flow of 10:1 the total flow would have been 14.6 ml/min, and that might be why you saw some improvement. Try increasing the total flow - this makes no difference to the split ratio in splitless mode, because the extra gas does not pass through the liner anyway.

Muck in the spsit vent line and filter seems to be a common problem with Agilents, it might be worth checking that.

Peter

[Johnny Rod]

Column flow is 1.6ml/min, septum purge is 3ml/min s the total is 4.6, in splitless mode, so the split vent should be shut off. It passed the leak check with no cap on the split vent (only on the septum purge and column fitting) so I presume that's actually shut off properly. See page 4:
http://www.ecs.umass.edu/eve/facilities ... 0inlet.pdf



When the TD is attached, the purge and trap has its own split flow of 10ml/min to prevent back-diffusion of air into the system and so the MS, so total flow is 14.6ml/min; I disconnected the TD though to simplify things and found the problem remains.

[Peter Apps]

That schematic (210-2) shows a front pressure regulated system (PVI is controlled by feedback from the pressure sensor downstream) while in split made it has back pressure regulation at PV2 via feedback from a pressure sensor upstream. If PV1 is faulty then the problem should go away completely in split mode, and if PV2 is faulty there would not be a problem in splitless.

Are you sure that the input pressure is stable - did moving the GC involve connecting it to a different gas line ?

What is the time scale of the pressure fluctuations (and how are you measuring them ?), aad is it a continual wandering, a regular cycle, or periods of stability punctuated by fluctuations ?

Peter

[Johnny Rod]

Some progress! Thanks for the info Peter, I went back to split mode and it seems fine - I'm now second-guessing myself from the first time I tried that but I've tried it several times today and each time the total flow line is nice and flat. Switch it to splitless and it's back as it was. It's a noisy trace with the occasional upset where it seems to go off and try to correct itself, but badly.

I swapped PV1 for another from the aux EPC on an unused channel. This had the same part number but I had to guess that the valve in the EPC assembly (for the front inlet) nearest the gas inlet (from the line supply), as all three had different part numbers although the split on/off valve was labelled as a digital valve (the other two as analogue). It sits under the pressure sensor too so I'm hoping it was the right one. Anyway it made no difference at all.

I am presuming the flow sensor is fine otherwise it wouldn't be able to produce a flat line (minimal noise) in any mode. As far as I am aware there were no changes made to whatever control parameters when it was moved. It was moved from one site to another, along with a number of other GCs of various makes. One of these is a Varian 430 which has turned out to be quite sensitive to the line regulator used, but this is fine on the line helium, which is shared by the GCMS and others. I obtained the same trace on another 6890 in splitless mode and the total flow was a flat line.

So, maybe it's the pressure sensor? I zeroed this again today, made no odds. Is this replaceable, or is it a whole new EPC unit?

Many thanks

John

[Peter Apps]

Hi John

My guess would be that you would need a whole new EPC - most instrument companies these days only sell complete assemblies. It would be worth checking that the pressure really is fluctuating before you spend your money - hook up a Borden pressure gauge to a syringe needle and stick the needle through the septum for a direct readout independent of the GC EPC. With front pressure regulation during splitless a fluctuating leak could make the pressure fluctuate, check whether the flow out through the septum purge is constant (this is a bit of a chicken and egg though - fuctuating pressure will cause fluctuating flow and vice versa). As far as I can see from the schematic the septum purge has its own controller, so if it has a problem it will be there in both split and splitless.

Peter

[Johnny Rod]

I think you could be right, am waiting on Agilent to call me back. Thanks for the help though. I capped off the septum purge vent and this made no difference so I think the regulator is fine. I'm a little unsure myself, it could still be a case of controlling parameters, but it looks like these are factory set and I don't see why moving would affect anything like that. As for the other components, I think with your help we've eliminate theother things, unless I've managed to swap one faulty valve for another, which seems unlikely. Will post an update when it's fixed.

[Johnny Rod]

Well this isn't getting any easier. I swapped the EPC module from another 6890 and the problem stayed with the GC not the module. I re-plumbed the helium. I also went back and checked the pre-move data (one of the problem-solving rules: double check) and on closer inspection with more cynicism than previously employed, it doesn't look that great, in fact much like I have now. So this isn't a new problem.

I also noticed it jumped around more when touching bits of tubing, but this wasn't consistent. I turned the oven off, left the door open, fitted a different column with the tail end loose, went all over it with a helium sniffer, turned off various nearby bits of equipment, and nothing seemed to help. I also tried fitting the inlet to the rear instead (just the cabling) but I couldn't seem to make this work.

I might just have to live with it, but out of interest, where are the control parameters located? In other words, what factors does it use to control based on the sensor readings? Or is there something else I should be looking at?

[Peter Apps]

Hi John

That all makes me wonder if the problem is with the readout rather than with the actual pressure ? Do you see any variation in the column flow rate as measured by dead time (tricky to do with your setup I know) ? Try the independent pressure gauge with a needle through the septum - the fluctuations are pretty small so you will need something quite sensitive. If you put a flow meter on the septum purge do you see fluctuations ?

Peter

[AIM]

Johnny,

Can I please buy or barter for a copy of the Markes Unity software? our computer crashed and I cand find the disk and Markes does not sell it anymore.

Alex

[Jake]

I'm puzzled by this. I believe it was an engineer who told us that flow stability below a total flow of 10 mL/min on 6890s/7890s was poor. This came up when we were trying to save gas during our sleep method and using a low total flow, maintaining the desireable column flow. We were having all kinds of flow-related errors. We bumped total flow to 20 mL/min and haven't had any problems.