Bringing a 6890N back into operation

[Gaston]

Dear colleagues,

I am currently trying to put back into operation a 6890N GC that we have inherited from an old research group at our university. I intend to use this GC for the on-line analysis of the outlet gas of an electrochemical reactor - more specifically, I want to determine the amount of CO and H2 in a CO2 stream.

Now, in order to be as cost-effective as possible, I was instructed to look into whether this "abandoned" 6890 can be used for that. As far as I can see, this GC is fitted with a TCD and an FID, and also has a six-way sampling valve. So, from my very limited understanding, it looks that it could be appropriate for my intended application.

The thing is that this GC has been lying in a room for some years now and no one knows what was it used for before and whether it even works or not. I have been tinkering with it lately, but as I am a complete newbie, I have not tried anything too bold, as I am afraid to cause any damage. I have been following the Installation Manual, but I was wondering is if you guys could point me to some kind of checklist I could follow to assess the status of this 6890 at minimal expense.

After connecting the GC to a helium line (and not removing the CarbonPLOT column which was already installed), I know so far that:

- The self-test diagnostics run OK.
- The heating of the oven, the inlet, the TCD, the FID and the valve box seem to work fine.
- Reference and make-up gas flows to the TCD seem operational. I have not turned the filament on, however (afraid of inadvertently ruining it!)
- When pressurizing the inlet, the pressure setpoint is reached and maintained quickly (although at first there seemed to be a leak, but this problem went away after tightening the connections of the six-way valve).
- The valve actuator seems to work fine.

I suppose that at this point, it would make sense to try to inject something and obtain a chromatogram. The problem is that the original ChemStation license has been lost and we would have to buy a new one. If possible, I would try to be almost certain that the GC would need no major repairs before splurging on the new software.

(Would it be it posssible to temporarily connect this 6890 to a computer with OpenLAB which is connected to another 6890? Not both at the same time, of course.)

Thanks a lot for your help.

[rb6banjo]

Do you have at least one split/splitless inlet on this GC?

If I am you, I would probably hook up the flame (you need a hydrogen tank and an air tank) and check it all out using that with the split/splitless inlet. You really can't mess that detector up. If I get good precision for an FID test mix on that, I'd move on to the TCD.

So far, it looks to me like you're on the right path. From your description, the GC seems to be behaving well. Always make sure that there is flow on any column that is installed in the GC (whether through the rotary valve or the split/splitless inlet).

[Gaston]

Dear rb6banjo,

Many thanks for your answer. The GC does have a split/splitless inlet (which seems to work fine, at least as far as heating and pressurizing is concerned) so I can definitely try that out.

Regarding the software part, do you know if the makeshift solution could work? The idea would be to take a computer which is connected to a [currently operational] 6890, disconnect that GC and then hook the computer up with the GC I'm working on to take some chromatograms. Or is the software 'tied' to a given GC, so that it cannot be done?

[LCbob]

Hi

Not too experienced with Openlab but in general if your 6890N and the other one is also a 6890N then it is just a case of making sure you change the 6890's IP address to mimic the working one. Things get interesting if it is a 6890A or plus but still doable I believe.

Hope this helps

[rb6banjo]

I run all of mine with a computer but I'd bet there is still an analog output from the 6890 so that you can take your FID and/or TCD signals to a strip-chart recorder or stand-alone integrator. You can control the entire instrument from the keypad on the front. That would help you figure out if the instrument is working well before you attack how you're going to control the instrument.

So, the PC (and the software) that once controlled the GC is no longer available? Your Agilent rep. might be able to help you with that. If you can verify with them that your institution purchased that chromatograph, they might be able to go back through the records and find that you purchased the software too. Agilent might be willing to give you a copy so you can set it up again. It's likely that they're not selling too many copies of that old ChemStation. If I'm them, I'd rather you keep using one of my old instruments than have you go buy something from a competitor!

[Gaston]

I emailed Agilent asking if they could help us out for retrieving the license information, but I was told that unfortunately that was not possible. They did offer to sell a new license, though.

So what I did today is that I installed OpenLAB CDS Chemstation Edition (whatever that means, but it looks very new) in another computer and connected the 6890 to it. Apparently, the software can be used for 60 days in "Trial Mode", which should be more than enough time to run some tests on the 6890 to check if everything is OK before buying a new license.

However, in the meantime, I found a CD in a lab drawer with "MSD Productivity ChemStation Revision D.01.00 SP1" (again, I'm clueless about the versions and stuff) with a software registration number. I'm thinking this might be the software that was originally provided with the 6890. If it is the case, that would be great, because then we wouldn't need to buy anything extra

[KM-USA]

Agilent used to put a sticker on the side of the instrument with the 10 character license number. So check for that, maybe....

[Gaston]

Well, the software issue is solved for the moment... thanks for the tips.

Now, I noticed that the column appears to be connected to the two detectors simultaneously: the column ends into some kind of "round" Y, and then this is connected to each detector. Is this a usual configuration? Can I run tests with that or should I connect the column to each detector individually?

Again, many thanks for your help.

[LCbob]

However, in the meantime, I found a CD in a lab drawer with "MSD Productivity ChemStation Revision D.01.00 SP1" (again, I'm clueless about the versions and stuff) with a software registration number. I'm thinking this might be the software that was originally provided with the 6890. If it is the case, that would be great, because then we wouldn't need to buy anything extra

Hi
this is the Mass Spec version of chemstation, it is quite different from the usual chemstation. Also I am not too sure whether you can configure an 'online' 6890 without a mass spec. I am sure someone else will answer this question though.

[Steve Reimer]

The MS version can run a 6890 with no MS connected. I have done so while debugging connectivity issues. I haven't collected data with it though. When you configure the instrument don't check the MS box.

There are a couple of ways to split the flow from/into a column. What you describe sounds like one of them. If you don't care what the flow onto each detector is for testing it may work for now.

[MSCHemist]

Chemstation MSD can run an FID just fine but the analysis software can bit a bit quirky to setup calibration tables and such. I use Chemstation MSD B.02 to run my 5890SII-FID. If you just want to collect data and get a printout of the area counts it is more than sufficient. It just treats the FID signal as a TIC.

[Gaston]

Many thanks on the information about ChemStation MSD. I think I will be mostly using the TCD, so some inconvenience with the FID could be an acceptable compromise.

I just did a test run now with both detectors turned on using the sampling valve. The TCD signal shows a peak at a RT of 1 minute which I guess is air. The FID has a pretty high baseline (around 26 pA) and I read that it should be less than 20... is this correct? If I turn off the flame, it goes down to 0 pA immediately.

Also, the FID signal shows two "jumps" when the switching valve is turned on and off. Is this normal?

[MSCHemist]

That sunds normal. The FID may just need a day or so to stabilize or the jet and collector may need a cleaning.

[muGC]

Are you sure its not the FID but the TCD that "shows two 'jumps' when the switching valve is turned on and off"?

The TCD is extremely sensitive to changes in flow and your "jumps" are a typical result of valve changes. You just ignore that portion of the chromatogram in the integration events.

[Gaston]

Are you sure its not the FID but the TCD that "shows two 'jumps' when the switching valve is turned on and off"?

The TCD is extremely sensitive to changes in flow and your "jumps" are a typical result of valve changes. You just ignore that portion of the chromatogram in the integration events.

It's definitely the FID, I think (unless I misconfigured the signals or something like that). Here's the chromatograms:



There's also a peak in the FID signal concurrent with the TCD one at 1 min, and those two peaks at the start and 0.25 min are the "jumps" with the switching valve. However, I suppose all of these are quite small, right? Likewise, there's some drift of the baseline, but it seems to stay within 0.5 pA or so.

Considering that the GC has been collecting dust for years, I wouldn't be surprised if there were some cleanliness issues going on with the detectors (and even other parts), I suppose.