5971 Tune Drifting

[benhutcherson]

Two weeks ago, I dug a 5890/5971a out of storage and set it up. I cleaned the source and gave the MSD a once-over the day I pulled it out(the 11th), then I installed a column and started pumping it down on the 12th. I also refilled the cal gas vial that day as it was showing low 502 abundance and the vial was close to empty. It tuned fine the next morning with the EMV at ~2000V.

Per the log book that was stored with the instrument, it was last used in late March of 2010, so had been out of service right at years. The last recorded tune value was 1647V on the EM. I fought a high water back-ground on the instrument for about a week, but it had mostly gone away by the time I left work on Friday afternoon(I understand this isn't terribly uncommon on instruments that have been open to the atmosphere for several years, especially since the 5971 doesn't have have separate source and quad heaters to allow for bake-out like the 5973 and later).

In any case, even though the background has settled down quite a bit, the tune seems to still be all over the place. I might get it to tune one morning with good peak widths, good 502 abundance, and have it go to ~1700V. That afternoon a manual tune might show 502 at ~3amu wide, and hitting autotune might have it fail by maxing out the AMU gain narrowing the peaks or EMV getting the 502 abundance up.

I did FINALLY hit "reset defaults" on Friday morning, and even with the EM at 1500V it actually looked quite good across the board. I had good abundances across the board, only had to set the AMU gain to the low 500s to get 502 down to ~.55 or so, and the AMU offset only needed a slight tweak to have them all in that range. Quicktune only made slight adjustments to the manual tune parameters I'd set, and I think ended up with the EMV at ~1550 to get the abundances it wanted(and I think used AMU offset rather than gain to get peak widths to ~.52). When I checked it before leaving, 502 had widened out to ~.65 or so, but everything else looked good and again I brought it back in line by adjusting the AMU gain again.

Still, though, it concerns me that it drifts that much after a few hours. I'll see what it looks like when I get to work tomorrow(I won't be in until late in the day), but it's still drifting more than I'd like if Friday was any indication. Granted I've been spoiled by a 5975 over the past few years-the only 5971s I work with are ones I service independent of my job, and when I was using one regularly back in my undergrad days I didn't even understand the concept of tuning. This is the first time I've been in the position of doing BOTH analysis and maintenance on a 5971(I've only ever done one or the other). In any case, if I pull up the manual tune on a 5971 even a month after a full autotune my peak widths will still be well under .6.

Also, as general information-I didn't get a rough pump with this instrument, but have a massive Varian unit I scavenged from a Saturn 2000 that needs to be sent to surplus. I've used this particular pump for the past year or better with the 5975, and it gave better foreline pressures than I'm getting even with my newly rebuilt Edwards pump that shipped with the 5975. I swapped them so that I could put the Varian on the 5971, and I've seen foreline pressures below 20mTorr at .5mL/min He and ~30mTorr at 1mL/min(this 5890 does have an EPC-it's a bit weird to me to not have a gauge on the front a 5890). I'm waiting on my Ebay-special vac gauge controller to arrive, as that got lost in storage.

Any thoughts on if the tune drifting is something I need to be concerned about, or if it just needs more time to settle down after storage?

One last thing-when I asked about cleaning the quads in the 5975, my predecessor in my job said "Brenda(local Agilent FSE) would probably kill me if I even touched them." From what I understand, the 5973 and have gold-plated rods fused to the ceramic that are basically not ever to be disturbed. The 5971 manual DISCOURAGES cleaning the quads, and says that they require no periodic maintenance and not to disturb their orientation, but that they can be rinsed with acetone, hexane, or freon. I've also seen reference on this forum to the front part of the 5971 quads being susceptible to "burning" much as what happens in the source. Knowing that I could easily turn a working instrument into a pile of junk, is it worth at least flushing the quads with a bit of acetone or should I leave well enough alone?

[James_Ball]

The quads on the 5971 and 5972 were a type of silver tape attached to the inside of the fused silica quad. This tape is not the most robust thing ever made, which is why they switched to the gold inside the fused silica of the 5973 and later models. I have cleaned quads on a 5971 before, but once you are at that point it is more a try at salvaging one that is beyond use, unless you just had something like a diffusion pump back stream that needs to be cleaned off.

These were also prone to having the entrance lens wear away the tape at the ends of the quads from being inserted and touching it slightly. You can see it if you remove the quad and it will appear as "mouse bites" (as Brenda would call it) in the ends of the tape. A few tiny ones are ok, but once it digs in deep the quads are pretty much done for. Removing them is not difficult, just have to remove the two RF contacts on the side of the radiator, which are the brown vespel parts with braided wires on them, then you can remove the detector and the white ceramic end fitting which on the inside has a circular ridge that holds the quads in the proper place, and gently slide them out. When reinstalling I normally would slide a long cotton swab in from the front end with the source removed, then slide the quad in from the rear and pick it up by inserting the swab inside and guiding it into the circular ridge on the front insulator, then do the same on the rear. Just mark the orientation so the same side it up and forward when you reinstall.

Being at atmosphere so long, it could be it is still just equilibrating and all the moisture being drawn out, those ceramics hold on to the moisture for a long time. 502 response is very sensitive to the amount of moisture in the system on the older models. Another engineer once told me you can set the mass to 500 as you would when DIPing the quads with the set RFPA menu item and let it set for a while, the higher energy of the high mass setting can warm the quads some and help dry them more quickly.

[benhutcherson]

Thanks James for the details response.

Reading a bit more into your response, it sounds like I'm best to let it be for now and see if it settles down.

I don't REALLY want to vent it at the moment. I have someone bugging me about analyzing aqueous solutions, which I'm not wild about but may vent it in a week or two so that I can fit a CW-20M column. Of course, I don't know how smart putting an aqueous solution-regardless of the column type-in a 5971 is-I'd probably rather do it in the 5975, but at the same time I'd have too many people after my head if it had anything other than a 30mx.25mm DB-5(or eq.) column in it since there are way too many degrees riding on repeatable results from that instrument at the moment.

[James_Ball]

If the peak widths continue to get wider, especially going out to over 1.0 within a weeks time, then you may also want to look at the top board. I had a 71 once that had a dying resistor or capacitor on it with that problem, once it completely blew out I finally knew what it was.

71/72s aren't great for aqueous samples. What you really need for those is a 5970. The metal quads are better for aqueous in my experience.

I don't know if they still make them or not, but I think AlphaOmega used to make a set of metal replacement quads to retrofit into the 5971, if you ever need to replace one.

[benhutcherson]

Well, I came in today and it had widened out to .75ish at 502, which is a lot better than what I have been seeing. There again, I brought it back with manual tune.

It does indeed seem to be getting better the longer it's heated/under vacuum, so perhaps it will settle down.

I'm a bit anxious about routinely doing aqueous samples. I'd actually really like to have a 5970, but it's only going to happen if I find one in storage somewhere or otherwise get one for free...if anyone has one lying around that they'd like to donate to an aspiring chromatographer at a public university, I'm all ears . I'd really like to have an instrument that I can do CI in also(I currently don't have one) and a 5970 or even a 5889 seems ideal for that. With that said, data systems for MSs that old do seem to be a bit of a problem.

[benhutcherson]

Here's another head scratcher on this one...

I came in today and pulled up manual tune, and saw no abundance for 219 or 502. I tried doing a scan over a ~20 amu range around those masses, and it wouldn't do it because it couldn't find any peaks there. I let it sit for a while, and nothing changed.

Finally, I switched over to filament 2, and the tune was almost perfect with my only complaint being that the peak widths were a bit too narrow. I manually brought them close, and Quicktune got them there with the EM at 1588V.

This isn't the first time I've had issues with filament 1, although it seems to at least be working somewhat since even today I was getting some abundance on 69.

Also, I seem to still have a fair bit of air(28) in the system. I've of course re-tightened the ferrules at both ends(I'm using graphite/vespal). I checked with an unlit propane torch and can't find any obvious leaks. The carrier would SEEM an obvious source of contamination, but at the same increasing the carrier flow rate pretty dramatically decreases the abundance of 28 relative to 69...

[James_Ball]

We had a 5971 that had problems with a filament and the engineer(Jim) found that when they welded in the ceramic pass through for the filament pins there was what looked like ion burn, which was probably some spatter or just metal vapor deposit, from the welding process and when high current was placed on the pins it would conduct between them. When using filament 2 you just have a ground potential on filament 1 to draw the electrons through the ion chamber of the source so if it works on filament 2 I would go with that for now.

I never ran a 5889, but did use two of the 5995 models, which were a precursor to the 5970, as they had the same analyzer setup. But don't go looking for one of those or Brenda will never talk to you again as they were a genuine pain to keep going.

[benhutcherson]

Duly noted

I actually don't have that much contact with Brenda, and have only met her a couple of times in passing. I know she is quite revered by anyone who has ever worked with her, though.

I still wouldn't mind having a 5970 . There's one on Ebay now for a few hundred dollars, but I'm afraid it's a bit of a pig in a poke, and assuming I can manage to land an interview with Agilent(and end up with an FSE job with them) I'll be relegated to running crazy experiments with playing around with this stuff in my garage

At least at my current job I can legitimately make a case for a 5970, esp. since we don't have anything that can do CI and you mentioned it being better for aqueous samples. The guy wanting to do aqueous samples, though, has moved on to looking at doing headspace sampling-I've offered to either improvise the Agilent unit that the paper he found references, or to try it with SPME...

[benhutcherson]

Just to report on this-I was out of town for the second half of last week, but before I left I vented and went digging in the mass analyzer.

First of all, there was what seemed to be some diffusion pump oil splattered toward the back end of the mass analyzer. It was around the general area of the electron multiplier, but I didn't see any evidence of it on the EM itself. I used a Kimwipe and methanol to clean that up, along with the flanges on both the analyzer assembly and the bottom half of the vacuum chamber. The gasket got a good wipe-down too.

I removed the EM and dropped out the source, but was afraid to dig too much into the quad. I pulled the ceramic insulator off the detector end, but didn't want to disturb it any more than this. In any case, I had never really looked at one of the fused silica quads up close before, and wasn't entirely sure of what I was looking at. I didn't really even see any areas that looked like they were conductive. Since it had at least been working, I decided to just put it all back together and let it pump back down and stabilize.

I revisited it on Tuesday, and of course the tune was a mess. I hit "reset defaults", and amazingly enough almost immediately saw good abundances of all three ions with decent peak shape. The high masses were a bit wide and 502 was a bit low, but altering between quick tune and manual tuning I ended up getting all three to .52 with the appropriate abundances and ~1700V. My AMU gain and offset are at mid-range values.

I've tweaked the manual tune parameters a bit since(specifically gain and offset), but didn't mess with them at all today and still had peak widths .52-.55 and appropriate abundances.

I do think just letting it sit after being open to the atmosphere is proving to be healthy for this instrument. I didn't run any samples this week, but I have no reason to think I'd have trouble if I did.

Also, I did get a vac gauge controller. With a flow rate of .5mL/min He and the cal gas valve open, I'm seeing ~7.8 x10^-6 torr(and around 7x10^-6 with the cal gas valve closed). This seems awfully low to me for a diffusion pump instrument-in particular I've never seen my 5975 that low with the same flow rate. At the same time, though, the manual cautions that you need to multiply the readings by 6 for helium, which taking that into account DOES give similar numbers to what I see on the micro ion gauge on the 5975.

[James_Ball]

Not uncommon for a little oil to get up around the bottom of the chamber, I normally see droplets on the baffles and sometimes a little around the opening. If someone who shut it down the last time vented it at the rough pump instead of at the column interface, that would explain so much up near the multiplier and bottom of the radiator. 5971/72 you have to either remove the column at the interface nut or slide it away from the GC and vent at the clamp where the interface goes into the side of the source, to prevent backstreaming of the diff pump oil.

Sounds like the excess air and water are finally desorbing from the internals, it does get deep into all the items even the metal when allowed to sit open for such a long time, and without direct heating of the source and quads it takes the older models longer to fully degas.

[benhutcherson]

Thanks again-I'm glad to know that the small amount of diffusion pump splatter I saw is probably not a cause for concern.

I do have two 5971s under my care now, so have gotten in the habit of venting them through the transfer line even though it seems to forever. It's amazing to me that a broken column even a foot away from the nut will pretty quickly have the rough pump gurgling and splitting. Physically pulling the ferrule back from the transfer line opening, though, seems to take a couple of minutes to get to atmospheric pressure. The last time I opened this one, I THOUGHT it had stopped hissing, but still had a nice rush of air when I pulled the transfer line from the source.

My water background IS much lower now, so I think you're right that it just needed to sit hot under vacuum for a while after nearly 10 years of being open. With the transfer line at 280º, the MS runs in the 170-175º range. I miss being able to hit "bake out MSD" to make most issues go away overnight, but I guess that's the nature of the beast.