Where to turn? Bringing a 5890/5970 GC/MS back on line

[Don Baker]

Ron,

Thanks for the info. The die is cast at this point. I already disposed of the old oil, so I have to go with the new oil.

I have been receiving help - all of it very much appreciated - literally from all over the world getting this thing up and running. I am about to attempt an autotune with PFTBA. I expect it to fail, but it should give an idea of the condition of the ion source and the electron multiplier.


Mariosapm - I failed to address your question earler. HP-UX is HP's version of the Unix operating system, including X-Windows. The version of X-Windows that I have can barely be called a GUI.

Onward and upward.

Don Baker

[Don Baker]

Well, I'm glad I dug into the roughing pump a bit.

Originally, my purpose was to clean the sight glass which was so covered with grime that it was impossible to see the oil level. I removed the oil case, which required only the removal of four screws. I was concerned that I would damage the gasket, so I loosened the case by pounding near the seal with a Dead Blow hammer. It worked, and the case and gasket were removed without damage.

The oil case and pump parts were coated with a layer of sludge, most of which I was able to remove with absorbent towels. Since my main goal was the cleaning of the glass, I cleaned up the case mostly and got what was easily accessible on the pump.

Behind the sight glass is a strip of metal that serves as a reflector to help see the oil level. For some reason, the engineers at Edwards apparently felt that it would never be necessary to remove the metal to clean the glass, because they chose to rivet the metal strip in place. I chose a drill that would serve as a tap drill for the #6 sheet metal screws that I planned to reattach the metal plate with and used a drill press to control the depth that the drill could penetrate. I was careful to leave at least 1/8" of metal between the tip of my drill and the exterior of the oil case.

With the rivets out, the metal strip came off easily and I was able to clean it, and the sight glass, with acetone. I tried to minimize the amount of solvent used because I wasn't sure what effect it might have on the seals or vanes of the pump.

The shortest screws I could find were 3/8" long and I had to grind a bit from the tip to shorten them so that they would seat in the short holes that had been drilled in the case. The screws only have to hold the thin strip of metal in place, so they don't need much holding power. Silicone glue (which is used for the sight glass) would have worked to hold the metal as well, but the screws will allow for easier removal should cleaning once again be required in the future.

Just before replacing the oil case I took another look at the pump. I went back over it with absorbent towels again to get as much of the gunk off as I could. In the process, I discovered the oil inlet screen which, up to this point, had been an unrecognizable protrusion near the bottom of the pump. The filter screen is held by one screw and a collar and came off easily. Not only was the outside of the screen covered with gunk, but there was also a buildup, perhaps a mm thick, of goop on the inside of the screen. I was very glad that I had taken the time to take another look around.

Since the filter screen is stainless steel and easily removable, it went to the ultrasonic cleaner. A short while later, it was good as new. The reassembly was uneventful.

It had been recommended that I circulate fresh oil to help clean out the rest of the pump. That seemed reasonable, so I ran the pump for a day with fresh oil, removed that, and replaced with fresh oil again. The oil that I removed showed a definite discoloration (the fresh oil is water clear), so the exercise was worth the effort.

Thanks again to everyone for the support. The pump has been chugging along happily for the past three days and (with the turbo pump on, of course) the pressure in the MSD is now down to 1.7 x 10 -6 torr. This is the lowest pressure that I have gotten to date (but it is also the longest that I have run the pumps).

I haven't loaded PFTBA into the vial yet. I removed the vial to do that and realized that the hole at the top is so small that there is no hope of simply pouring the PFTBA from the bottle into the vial. I'll need some sort of syringe. All the syringes I have are dirty, so I now have people out looking for someone who can get me a clean syringe. I'm hoping I can find a diabetic who won't mind parting with one.

Is there an easier way? Since this is a calibration standard, I assume that it shouldn't come into contact with anything but the cleanest tools. What is the usual procedure for loading PFTBA into that tiny vial?

That's it for today

Don Baker

[lgchrom]

hi Don,

interesting update.

if you find some time could you post some screenshots of your working environment (the software i mean) to see what it's like with this old software you use. Could you show some typical screens?


also, what parameters do you plan on analyzing.


thanks

[Peter Apps]

HI Don

Interesting stuff on the pump - it sounds as if it has worked hard for its living ! If you can avoid it, do not use plastic syringes for anything that has to be chemically pure - they are all lubricated, and the plastic and rubber seals tend to leach undesirable stuff. Sooner or later you will need some glass-bodies micro-syringes for sample injections and dilutions of standard solutions. Now might be a good time to get them, then use one of them for the tuning fluid. Alternatively, heat the glass bulb and then put the tip of the bulb under the surface of the fluid, as the air in the bulb cools again it will suck fluid into the bulb.

Peter

[Don Baker]

I figured out how to get the PFTBA into the machine's vial. I found some glass capillary tubes. I took one and flamed it to the point that it started to bend. I let it cool momentarily, then dipped the tube into the PFTBA, capped the end with my finger and transferred to the vial. No drips, no spills, and hopefully, no contamination. I don't know exactly how much PFTBA I got into the vial, but I did four transfers and called it good.

Pressure-wise, I had gotten down into the 10^-7 range, so I felt it would be OK to move on. I first turned the electronics on to boil off whatever had collected there during the instrument's long time in storage. I had turned on the electronics for a short time during the last pump down, but this time I left the power on for several hours.

Shortly after turning on the electronics, the pressure started to rise. I didn't keep careful track, but it went well into the 10^-5 range. But after about four hours, the pressure had dropped back into the 10^-6 zone. I hadn't yet opened the tune valve and I knew that it would release a bubble of air into the system, so I just went ahead and opened the valve to get that out of the way.

The ion gauge didn't like it at all and shut down immediately. And I swear I heard the turbo pump complain momentarily. But by this morning the pressure was back down to 2.6 x 10^-6. All is well.

I'll give the autotune procedure another try. I expect it to fail again, but a technician I have been speaking with says, "You never know, you may get lucky." I'll take some photos of the screen as I go through the procedure and post some, as mariosapm requested.

Is there a forum guideline on the maximum image size for posting?

Don Baker

[Don Baker]

> Is there a forum guideline on the maximum image size for posting?

Never mind. I found it
I don't know that screen shots will be legible at 400 pix by 400 pix, but I'll keep the size as small as possible while maintaining legibility.

Don

[Don Baker]

The autotune procedure failed, as expected. The good news is that the machine went through the motions and really did try to find the three main mass peaks it was looking for. It was partially successful, as I will show below.

As usual, I have some questions, which I will present at the end.

Here is what the screen looks like at the beginning of the autotune procedure. I didn't take a photo* of the main screen, as you can still see it behind the Autotune screen. There is nothing else visible on the main screen other than the menu choices that you see at the top.



As the procedure progressed, the system attempted to find three key peaks at 69, 219 and 502 m/z. As is evident in the following image, the quality of the peaks was not good. The EM voltage was also high. This shot shows the EM at 2600 volts, but I saw it go as high as 2800 without any improvement.



Ultimately, the system gave up and produced an error as shown below:



Additional attempts also ended in failure, but often the reported error was different (note that the system was actually able to locate the 502 peak this time):




QUESTIONS:
1) If I am interpreting these results properly, the EM does need to be replaced. Is that what you would conclude as well?

2) I was not able to ascertain anything regarding the state of the ion source from the test. Should I be able to?

3) I am assuming that if I have to replace the EM, I might as well take the time to clean the ion source. When I checked the maintenance manual for the appropriate procedure, I found that there are at least two special purpose items needed, identified in the manual only as "Bench Block, HP part number 05990-20254" and "Assembly Fixture, HP part number 5995-60098". I have not found these part numbers on any website. Does anyone know where I can find them?

4) If I have to break into the high vacuum area, is there anything else I should be looking at while I'm in there?

I still have the system under vacuum so that if anyone suggests a test I can do it quickly. Vacuum is holding steady at 5.5 x 10^-7 torr. I will keep the pumps running until I determine that there is no more useful information to be gained from running instrument tests.

As usual, all comments and suggestions are very welcome

Don Baker

* The images are all photographs. There is no way to do a screen capture with this computer.

[Peter Apps]

HI Don

I have no idea whether it will help in this case, but with the Agilent / HP mass specs that I used to work with the autotune alwys started off with the current tune settings and worked from there. Sometimes it helped to manually set all the values to either the defaults, or something close to the minimum, and then run the autotune.

Peter

[Don Baker]

Now we come to one of the more difficult areas. I have to order some real chemicals. Of course I'm interested in quality and a decent price, but I also don't want the DEA kicking in my door in the middle of the night simply because I ordered some dichloromethane (a member of my microscopy group recently had his lab ransacked and many items taken in such a raid. Though the charges have been dropped, he is left with his lab destroyed and his equipment still being held by the police as "evidence" even though there will be no trial).

Right now, I just need some solvents for cleaning the ion source (after abrasive cleaning), but eventually I'll need chemicals for extraction processes too. I can order through my (oilfield related) company which should help in setting up an account. Specifically, I am looking for 1-1-1 trichrloroethane, methanol and acetone. The trichloroethane is what is recommended by HP for cleaning. I understand that this chemical is being phased out (though it looks as though VWR still has it). What would be a suitable replacement if I have difficulty getting it?

I would like to find a supplier that has a good all-around stock so that I don't have to go here for one chemical and there for a different one. Any recommendations for companies that have good stock, acceptable prices, and minimal hassles?

Don Baker

[Ron]

If the multiplier was in the instrument when you got it I would assume that it needs replacement, especially since the voltage is so high during tune. The rule of thumb I always used with HP mas specs was when the multiplier voltage hit 2000 with a clean ion source the end was near. It will take longer, but maybe go ahead and clean the ion source and then try tuning again. I'm not real optimistic that it will tune, but at least you haven't spent the money for a new multiplier. You will have to clean the source anyway before you put the instrument back in service.

I think that you are going to get this system up and runniing again. Vacuum system works, electronics appear to be good, all signs look positive so far.

[Don Baker]

Ron,

Actually, I have assumed, since before this all started, that I would need a new EM, so I already set aside money for the purpose. And if I'm going to crack open the vacuum system, I'll take care of both the EM and cleaning the ion source as well so that I don't have to open it again later, which I might have to do if I choose to test one part and not the other.

Speaking of cleaning the ion source, HP specifies two tools, a "bench block" and an "assembly fixture" that I do not see listed on the Agilent site. They appear to be simple devices, but there is no way I can get a machine shop to make them for me without detailed specs.

Does anyone know if there is a replacement or if there is any way to get a sufficiently detailed drawing/spec that would allow me to get one built?

Don

[Schmitty]

Ron,

Actually, I have assumed, since before this all started, that I would need a new EM, so I already set aside money for the purpose. And if I'm going to crack open the vacuum system, I'll take care of both the EM and cleaning the ion source as well so that I don't have to open it again later, which I might have to do if I choose to test one part and not the other.

Speaking of cleaning the ion source, HP specifies two tools, a "bench block" and an "assembly fixture" that I do not see listed on the Agilent site. They appear to be simple devices, but there is no way I can get a machine shop to make them for me without detailed specs.

Does anyone know if there is a replacement or if there is any way to get a sufficiently detailed drawing/spec that would allow me to get one built?

Don

Hey Don,
I'm not familiar with the 5970, just the 72, 73 and 75. But, if I can make a suggestion about the purpose of the bench block and assembly fixture; they are probably unnecessary and are just used to help assemble the source when taken apart for cleaning. Personally, I just lay the source on a Kim-wipe or a glove when assembling. I could be wrong, but I just thought I'd say something, since no one else chimed in.

[AICMM]

Don,

Take lots of pictures. The 5970 source is not nearly as intuitive as the 71's and beyond. I think it is your EM based on what I see but considering how worked over this instrument was, cleaning the source is quite appropriate. Also, I would recommend dichloromethane rather than TCA. Easy to get, used by lots of environmental folks, heavier than water, good solvent, good for cleaning sources.

Best regards.

[Don Baker]

AICMM - Good suggestion. I have the hardware manual and it has picture, but they're rather muddy and some fresh photos would probably be quite helpful.

Electron multipliers:
It appears that there are at least two configurations and several manufacturers of EMs for the 5970.
1) Does anyone have any comment regarding the "channel" type or "discrete dynode" type of EM?
2) For any given detector type there are always claims of better sensitivity and lower noise. Is there really much practical difference or should I just get the best price and go with it?
3) Some EM suppliers are quoting a maximum bake time for their unit. I was not aware that I had to bake my EM. What am I missing?

Some possibly relevant information:
The compounds being measured should usually be present in significant quantity (perhaps 5-10% of the sample)
I don't expect the machine to be heavily used, so EM lifetime is an important factor
While not the primary goal, I may have to look for pesticides as well and I don't know if that affects how sensitive I may need to be.

Thanks,

Don Baker

[Ron]

I prefer the discrete dynode multipliers. Not only do they have higher gain and lower noise than the channeltron type, the lifetime tends to be longer. I don't know if SGE makes an ETP multiplier for this mass spectrometer, but in my experience these have the best combination of lifetime and sensitivity. Not the cheapest, but they last for years.