Noise in A/W Check, Tune Failure, and more

[James_Ball]

Thanks for sharing your experience, Ben! I think you might be onto something about the autotune setting the EM high. Since I started here the EM has been tuning way higher than I'd expect for it's age, even the new one installed by the engineer last week. Something is going on there, but I don't know enough about the tuning process to know what could be affecting the EM voltage this way.

Unfortunately, I could not find a "reset default" type setting in this software (I agree this software is a misfortune haha). However, I found a way to set the maximum final EM voltage. I tried setting this to 1350, but still the tune would not complete. Received the same error of "Unable to set appropriate values for low/high mass resolution." I can manually tweak the tune to get ok values, but it very much worries me that the MS cannot complete an autotune.

Yesterday I tried using older tune files (even deleting the newer "bad" ones to avoid the software using those pathways). No change.

An engineer is coming today to install a new main board, so we'll see if that has an effect.

Thanks again!

Mass resolution error could be related to peak widths not being where they should be or not being stable, which is either dirty quads or the electronics that run them. If you have a tune that can scan, watch the peak widths. If you scan for an hour or two and they get wider or narrower then that is where the problem is. Of course that would be with the cap in place of the column, since a lot of air can cause drifting peak widths also.

[Kacie]

Sorry for the radio silence everyone. I hope you all had a great holiday!

An update: Last week a technician came in, inspected the source, and ordered an entirely new source assembly. New source arrived Thursday, and I installed it-- no change, would still not complete an autotune.

Monday: The technician installed the new main board. Still no change.

Tuesday: A new technician stopped by. He made a series of adjustments to the potentiometer (a control on the main board he described as a coarse tuning of mass resolution). The MS was capped at the time. On initial tests, the MS would complete an autotune! I was beyond excited and was 100% sure my problems were behind me.

Wednesday: I vented the MS and reinstalled the column. Super routine, nothing strange. Pumped down for >2 hours, checked a/w. H2O= 11%, NO2= 30%, and O2= 11% I did a mini bake-out for a couple hours. A/W was basically unchanged, but I tried an autotune anyway. It failed to complete. With my hopes and dreams shattered, I was happy to have the 4th off.

Friday (today): Upon the technician's suggestion, I am attempting to achieve a passing BFB with an untuned instrument. Here's the chromatogram of the first blank I ran, with a window of representative spectra (spoiler: it's not good): https://alleghenycounty-my.sharepoint.c ... Q?e=EOuOzf

I am running a few more blanks and will see what happens.

Do any of you have any ideas? Thanks!

[James_Ball]

Was the BFB attempt made using the tune file that passed when it was capped?

If not it may have been using settings from the failing tune later and if the resolution is way out of spec you are just getting noise.

I would cap the detector, maybe let it set over the weekend, and tune again and see if it will complete. If it does, vent, reinstall the column, let pump over night then inject a BFB using the passing tune(if it passed capped) and see what the results look like.

It really seems there is some air getting into the system somewhere in the flow path. There has to be some area or valve that just can't be easily seen that has a leak. I had one in the electronic flow module of the GC once on a 6890 that was very difficult to find. If it tunes when capped, the MS probably is ok, but just not working with a large leak coming in from the rest of the system.

[Kacie]

Yes, the BFB attempt was made using the passing tune file. Thanks for the idea though!

I agree there is probably still a leak somewhere. After all that's what started this whole thing in the first place. But I feel like I have to fix this tuning issue before I can accurately address the leak.

Yesterday, a colleague suggested the column could be installed too far or not far enough. There are 3 methods of column installation in the manual. I've always been using the measurement method. So we tried the "optical" method, but didn't have any luck getting autotune to complete.

I played around with tuning/mass calibration this morning to give the PE technician some info. Here's what I sent him. And here's a link to the referenced pictures and video https://alleghenycounty-my.sharepoint.c ... g?e=lspzb3

-Pic A: Immediately after two hour pump down, using settings from our successful tune last week
-Pic B: Ran an autotune, it failed. I reset the settings to the “good” tune, and this is what I saw.
-Pic C: Did a “maintenance” autotune instead of a “full” autotune, in an effort to run the autotune based on the “good” settings (on full autotune it adjusts the start point settings iteself). These are the results
-Pic D: Ran a calibration—it completed! But did not show 502.
-Video E: Ran an autotune and it completed! But is atrocious.
-Ran another calibration, it failed to complete and did not show any mz 502.
-Ran an “maintenance” autotune WITHOUT mz 502 (only 69, 131, and 219). Completed.
-Pic F: Immediately ran a “maintenance” autotune WITH 502. Completed, but did not show mz 502.
-Pic G: Ran a calibration. Completed, shows noise, has 502 but it comes and goes.

[James_Ball]

Those peak shapes for 131 and 219 really don't look good at all. The 69 has a nice sharp peak, while the others are broad and have shoulders on both sides.

I am thinking there is still a lot of air coming through and giving the mass filter problems maintaining good peak shapes.

The first pic is using the tune made when capped but with the column installed correct?

[Kacie]

Agreed, it doesn't look good at all.

Yes, column is in but using the tune that was successful when the MS was capped.

I'll be away for the rest of the week so the system will just be pumping down untouched. Thanks!

[Kacie]

An Update:

We've had a couple of PE service engineers here for two days each. Here's a summary of what they've tried. At this point, no one has any idea what could possibly be causing the weird connection and signal issues we're having.

1. Removed and reinstalled TurboMass software, updated firmware, and hooked MS up to his own laptop. No effect.

2. Checked voltages on everything they could think of (two engineers did this separately). All voltages normal.

3. Installed 2nd new main board in case there is an underlying problem killing main boards. No change.

4. Noticed that MS will sometimes go into "pump down" on its own upon turning it on. Had done this to me once. Happened to engineer about 4 times in a day. Super weird.

5. Replaced head amp board, no change.

6. Replaced RF generator. Actually made the tune noise WORSE.

7. Inspected and cleaned HED with methanol. No change.

8. Replaced power distribution board. Saw PFTBA peaks, even 502 which has been rare. Tried autotune, but it failed to complete. Then mz 131 and 502 disappeared. After a few mins 131 came back but 502 never did. Heated up source to 200 and all peaks went back to their original bad state with noise.

8a. Tested the "source hot= bad tune" theory and it wasn't super consistent. Definitely an intermittent issue. All readbacks for the source are normal.

9. Noticed that the "vacuum ok" reading would often flash back and forth to "vented" when the filament was on. A previous engineer told me this was normal. This engineer said it was definitely NOT normal. Thought it pointed to a communication issue.

10. Ran a blank. Got a bunch of "peaks" interspersed with dead space. Within each "peak" the spectra had a similar pattern. Patches of grass (every mz) with dead spots of no signal. So odd.

11. Replaced OSISCR board. No change.

12. Replaced connection cables to OSISCR board and RF generator. Now see no peaks at all, only mild noise (for PFTBA peaks and a/w peaks). Autotune completed but gave 0 signal for everything. Not even noise anymore.

13. Tested power supply to instrument. Normal.

14. Installed new style piece which has both the EM and head amp board as one. Supposed to be much better design. No signal at all at normal EM levels. Can see very minor noise if EM is cranked to at least 3000. Tried different tune files and parameters with no luck.

15. There's one last board they haven't replaced, but it's back ordered. PE is trying to come up with our next step...

This is the most ludicrous problem. Nothing about it makes any sense

[James_Ball]

Is there a wire that looks like a spark plug wire connecting to the feedthrough for the HED?

I had that wire go bad on my 5975 Agilent and had similar problems, actually it was the pin on the feedthrough broke off and was stuck in the end of the wire, causing intermittent contact.

I also once had a 5971 that when the plate to the vacuum chamber was built the weld has spatter on the ceramic feedthrough and it would cause current to pass between two pins and the filament would not come on, but when checked with low voltage ohm meter it showed no problem. Sanding between the pins solved the problem.

It sounds more like something inside the vacuum chamber isn't getting the voltages it needs, but not sure what part. May want to double check everything related to the source since it seems to be worse when heated.

[Kacie]

Great news, it's fixed!!

After the new style EM was installed, the following happened:

1. New back plane board was installed

1a. While installing the back plane board, it was discovered that a prong of the new RF generator cable was messed up! The previous cable was reinstalled.

2. Capped the MS, and everything was NORMAL! The tuning peaks were nice, no noise. Autotune completed normally. After reinstalling the column, it was still all good.

Conclusion: It could have been the new style EM (but the bad cable was obscuring it), the OSISCR board cable, or the back plane board that was the issue. It's frustrating that we'll never know exactly what it was, but honestly after 1.5 months of this everyone is just happy it's fixed haha. Now onto the tough part of method development haha.

Thank you all so much for your suggestions and support!

[James_Ball]

Great news, it's fixed!!

After the new style EM was installed, the following happened:

1. New back plane board was installed

1a. While installing the back plane board, it was discovered that a prong of the new RF generator cable was messed up! The previous cable was reinstalled.

2. Capped the MS, and everything was NORMAL! The tuning peaks were nice, no noise. Autotune completed normally. After reinstalling the column, it was still all good.

Conclusion: It could have been the new style EM (but the bad cable was obscuring it), the OSISCR board cable, or the back plane board that was the issue. It's frustrating that we'll never know exactly what it was, but honestly after 1.5 months of this everyone is just happy it's fixed haha. Now onto the tough part of method development haha.

Thank you all so much for your suggestions and support!

Congratulations!

At least method development is much more fun than chasing a moving error in the equipment

[Kacie]

Thanks James

Yeah I suppose you're right haha