I presume the FID is detector B ? What is detector A ? If it thinks detector B is not installed then you should not even get a signal or a solvent peak ? Have you checked the gas pressures and flows to the detector ? Have you confirmed that the detector has ignited ?

Yes, it is detector B that we are connected to, and detector A is also an FID. I thought it was weird too, but we have had the same set up for some time, and it was working just a couple of weeks ago. At the start of the new week, I lit it up, and got no signal. I did all of the things I mentioned above, then got a signal, but very poor sensitivity. The gas flows are fine, and I confirmed that it was lit by holding a watch glass above the detector, plus you can hear it lighting...

When you remove the brass castle assembly and just have the base electrometer do you have a signal above 0? With the castle assembly installed but no flame what is your signal?

When you light the flame does it instantly jump to an elevated signal? Is the signal the same whether the column is installed or if the column is removed and the flame is lit with a blank nut installed? What about if you run with no make up gas (only H2 and air)?

It COULD be a gas issue, but having worked on lots of 5890 it is certainly possible that something in the electronics has died. Do you have any other instruments you can swap parts from?

Are you using a computer/data system or integrator?

No flame, signal =0. And when you do light it, yes it jumps VERY high. I have not turned it on with a blank nut or run with no make up gas. Once I did the maintenance mentioned above, I put the column in our other instrument, (we were running with the both of them dedicated to one column), and that column worked fine in the other instrument. They run off the same gas, and yes, we operate from an old version of ChemStation...

Very high like what, 8 million? If you just stick a metal screw driver inside and short the detector does it jump to the same value?

At first, it leveled off at around 2500. I ran a low standard and the solvent peak was about the intensity I'd have expected from the analyte, and the analyte was not detected. When I started it up the next day, it leveled off around 2500 again at first, but then shot to 8.3 million and stayed there, guess that's where the old model tops out. And I have not tried shorting the detector.

The only reason I ask about shorting the detector (just placing something metal against the collector to provide current) is you can check that it is capable of even "detecting" the maximum amount of signal. Now that I know your signal is around 2000 it is helpful--sometimes people talk about having "really high signal" and that signal is like 830000 or 8300000 whatever the max is (you can see for yourself by sticking a screwdriver or paper clip or something in the collector with the electrometer on--it won't hurt anything.)

Even if you have a very high background signal due to contamination or whatever you should still see a giant solvent peak.

First, do you have attenuation/threshold turned on? I know you said you're using chemstation, but make sure those settings are 0. Check the easiest stuff by making sure your method settings are correct.

Second, you said this instrument has two channels right? Is the other channel working and also SSL inlet/FID? And are both channels running the same gases?

If Channel A works but channel B is acting bad, install your column (which has been verified to work in another instrument) from channel A's inlet to channel b's detector. Shoot your sample. Then reinstall the column to channel B's inlet and shoot it again, compare. Repeat going from inlet B to detector A, then inlet A to detector A. I think you get what I mean here, but basically use your working components to half split the problem. Once you figure out if its the detector you can use the working parts (electrometer, signal board) to figure out which part is bad. hell, on a 5890 it could be the actual port the signal board seats into.

From the front panel you can press front detector and then A or B, which assigns the front or back detector to a given data channel. What kind of connection are you using between the GC and computer--is it analog to digital converter or GPIB?

Check for constant insertion of impurities (probably from the gas)

Not sure about the connection, but I may try that out when I get a chance. My lab is very busy right now, didn't get a chance to look at it today. Thanks!

follow me for check system step by step

1- syringe ( volume , manual test , me be dirty or closed)

2-septum ( check kind septum ,should be not dry)

3-Liner (check kind liner ,should be not dirty or close)

4- column (when is head cut column or tail ,should be ring full and distance input 2 mm in liner or detector )

5- gas ( carrier gas , flame gas , make up gas , pressure )

6- detector FID ( sensitivity test )

If you need to explain any step told me so.

I have another GC daisy chained to the same gasses as the problematic GC, and it is working fine. That's part of the reason I'm leaning toward electrical problems. I have dealt with a lot of GC issues myself before, but nothing electrical.

Yeah this is why I asked if you were using the same gas on two channels of the same instrument or with multiple GCs, if you have other systems sharing the gas and they're ok generally it's not a gas issue. That said I've absolutely seen contaminated gas give a constant response of around 2000 pA so it is certainly a reasonable possibility. Most GC problems are the inlet or syringe but if you remove the column from the FID, cap it off with a blank nut, and then relight and the signal is the same then the problem is either gas, the fid weldment, or electronics. Once you can get a normal fid signal (3-20 pA as someone said) then reinstall the column and evaluate the rest of the system.

If the signal is still pegged at 8 million my guess is its the interconnect (spring that makes contact with the collector) and or the electrometer (the metal box connected to the spring) - the signal board goes bad less frequently

Hmmm... I'm wondering if maybe it was a cleanliness issue before (I have had to clean or replace the jets plenty) and maybe I was too rough on the interconnect... Is that something that can be reinstalled or would I need a new one/take one off the Front detector?

Yeah if you stretch or bend that spring (easily done over time) you can cause problems with noise that won't go away till you replace it.

Thankfully, replacing it is pretty dang easy. Power off the GC, remove the castle assembly (as if you were replacing the jet) and then remove the right hand side cover of the GC.

Here is a picture that shows all the components of the FID: http://i.imgur.com/W69y2z8.jpg

Notice that the interconnect connects to the electrometer with some screws. The 5890 service manual is available with a little googling if you want some more detailed instructions.

http://csa-service.com/store/index.php/detector-parts/19231-60710.html

Here is the part.. $381, its easy to replace though :).

You shouldn't need to clean the jets often, everything non volatile generally stays put in the inlet or first inch or two of the column. If you're seeing white crystalline residue in the jets or collector/castle assembly you may be heating your column too high and causing it to bleed. What kind of column are you using? What is your oven temperature profile.. also what kind of samples are you running (what solvent are they

dissolved in and whats your injection volume)? If you think you're having problems with contamination there is definitely some things we can improve with your method/technique.

If you're getting graphite from the ferrule up in there either switch to graphite/vespel or don't tighten as much. If you're breaking off column in there its also possible you're over tightening or.. something else. I don't know, it happens.

My mantra with GC has become its better to undertighten than to overtighten. Its always possible to snug things up.. if you install your column and run a blank and see a big baseline rise with temperature increase just snug it up with a quarter turn of so. If you over tighten you run the risk of warping ferrules or causing damage to the connectors. This might not have anything to do with this problem but its good advice for anyone running GC or GCMS.

Oh also, since this one is tricky to identify and a total horrible pain to fix make sure and be careful not to crossthread your jet when tightening it down. If you do it won't go all the way in and the bottom of the collector will touch it, making your signal 8 million. When you try to remove the jet it will wreck the threads and you have to replace the entire base of the FID.. not fun :P.

Dear Alex, I am having same problem. My question is : the wire must touch the collector in order to assemble properly?

thanks