GC/FID recovery dropped

[rv3]

I am running an old 5890 with an OI FID/PID combo attached. we're doing 8015 purge and trap method. here is the full set up

DPM 16 port autosampler
OI Analytical 4650 P&T
HP 5890 GC
OI 4540(?) FID/PID

we were having some computer issues after a power surge a few days ago, there was a port comm error from the GC, I fixed that but now my I can only get about 10% of the usual recoveries from my QC. I'm not used to this instrument, the usual user is out of town for the next month, so I'm not sure what the flows are normally but all the gauges seem to read where they are marked on the face. I couldn't find any leaks in the lines from the purge vessel to the column or around the detector. i am completely at a loss. I have removed the FID to see if it needed cleaning and it seemed fine, nothing discolored and no sediments on the electrode. Any ideas on what this could be?

[rb6banjo]

Is that uniformly across all of your analytes? if yes, it would seem that you have a leak somewhere in the sample path. They can be tough to find.

Is your GC a Series II or is it older than that? The really old 5890's have analog regulators to control the flows. These would be unaffected by a power outage. The Series II's have electronic pressure controls for the carrier gas. They might be affected by the power surge.

[Bigbear]

Can you do a direct injection at the GC? This would split your set up in half .

[rv3]

@rb6banjo It is consistent all throughout the chromatogram which made me initially think leak but i couldn't find anything. i am not very familiar with the DPM though so it might be somewhere in that quagmire of a valve. It is a Series II but i was lead to believe those did not have an electronic pressure control, just the manual carrier flow valve. I'll take a look at that.

@Bigbear I don't really have a way to do a direct inject. we have the transfer line from the 4560 P&T directly connected to the column without any injection port. I could maybe spike the trap and run from there to rule out the DPM at least.

[Peter Apps]

When you say that you get 10% recoveries compared to QC are you using historical QC data (i.e. from before the computer problem) or are you comparing current QC standards with current recovery tests ? If you are using current QC standards (and they are analysed in the same way as recovery tests) then there must be something wrong with your standards - anything wrong with the instrument would affect QC, standards and samples equally, so your calculated recoveries would be the same as they were before.

Peter

[rv3]

When you say that you get 10% recoveries compared to QC are you using historical QC data (i.e. from before the computer problem) or are you comparing current QC standards with current recovery tests ?

Peter

The 10% is based on the area counts/signal strength returned the day before the power trouble.

[rv3]

When you say that you get 10% recoveries compared to QC are you using historical QC data (i.e. from before the computer problem) or are you comparing current QC standards with current recovery tests ?

Peter

The 10% is based on the area counts/signal strength returned the day before the power trouble.

[Peter Apps]

When you say that you get 10% recoveries compared to QC are you using historical QC data (i.e. from before the computer problem) or are you comparing current QC standards with current recovery tests ?

Peter

The 10% is based on the area counts/signal strength returned the day before the power trouble.

If you have not run new QC standards then you have not determined recovery - your problem is more accurateley described as a 90% loss in response.

A software problem and a drop to 10% of previous peak areas makes me suspect a signal processing setting - check through the method as it is currently loaded compared to what it was before the software glitch. Have a look at the baseline - if it is 10 times cleaner than it was before you have a change in signal processing.

Peter

[rv3]

I know it's been a while but the PC associated with this instrument went down and i just got it back online. So new install of the software and method build and i'm still seeing the low signal. Is it possible the power surge damaged an electrical component, like the FID electronic board? does anyone know of a way to test it? i couldn't find any obvious testing posts or anything of the sort.

[rb6banjo]

On the Series II, there are buttons on the front panel labeled "Attn" and "Range". When you load your method and then press those keys, what does the display on the GC say?

[rv3]

On the Series II, there are buttons on the front panel labeled "Attn" and "Range". When you load your method and then press those keys, what does the display on the GC say?

For this i followed the recommendation in the manual and set the Attenuation at 2 and the Range at 5.

[rv3]

On the Series II, there are buttons on the front panel labeled "Attn" and "Range". When you load your method and then press those keys, what does the display on the GC say?

For this i followed the recommendation in the manual and set the Attenuation at 2 and the Range at 5.

[James_Ball]

If you set the Attn and Range differently than was set before the power problems it could be dividing your results by 10x. Try going up or down with Attn and see if that changes the results. If not try the Range.

[rb6banjo]

Those settings are certainly attenuating your signal. I run my Series II at 0 and 0 for those. I get a very quiet signal and very good detection limits. Mostly, I use headspace solid-phase microextraction as my mode of sampling.

[emar]

Hi I'm having a similar problem but I am not aware of any sort of power surge. I'm wondering if your issue ever got resolved and what you did to fix it.

I am running a 5890 GC-FID/PID system with an OI 4430PID with a 4410FID attached to it. I have been having some issues with little to no detectable response. I've gone from a PID response of 140000 counts to almost nothing and an FID response of 5000 counts to 2000 (the FID wasn't doing so well before either but we were able to get by). I noticed it first with the FID signal and then soon after, the PID went. I'm at a point where I can't even pass a calibration. We replaced the lamp last year and it seems to still be in pretty good shape. I recently replaced the jet (making sure that there was no left over graphite in the exhaust), ferrules, and electron capture rod in the FID, but still no improvement. I feel like I have disassembled and reassembled the detector many times over with no success. I have clipped both sides of the column and verified that there is flow going through the column, and that there is pressure at the inlet. When reinstalling the detector side of the column, I made sure to follow the instructions given in the PID manual very carefully and am very certain that it is at the correct length. I have done manual injections with a super concentrated standard. While I was able to see peaks with that manual injection, it wasn't nearly at the concentration it should be. This test allowed me to eliminate my purge and trap concentrator.
I'm sure there has to be a leak or restriction somewhere, but I am at a loss as to where. There is nothing obvious I can find. I want to believe that my inlet is leak free because the small peaks I actually am seeing are very sharp with minimal tailing or fronting.
Do you have any suggestions regarding other troubleshooting spots? I am running out of ideas. Thanks in advance for your help.

My conditions are as follows:
Encon Evolution purge and trap desorbs for 1min to transfer sample to the 5890 series II.
Split 50:1
Inlet pressure holds at about 15 (w/EPC).
Oven Program: Start at 40c for 2 minutes. Ramp 6C/min to 140C. Ramp 18C/min to 180C. Hold 9 minutes.
Detector is at 220. Inlet is at 180.
The FID and PID signals seem fairly strong (17 and 90 respectively).