Mystery Noise

[Jason Hendricks]

Hello,

I have a problem that has me stumped. On at least two occasions we have seen a noticeable amount of noise in the baseline of our UV detectors. We do not believe that it is related to the LC system for various reasons. We feel that it is electrical in nature Here is our setup…

Situation One: 10-27-06
1. Waters 2487 UV Detector
2. Channel A output going to PerkinElmer 970A Nelson Interface box via a shielded cable.
3. Data collected on TotalChrom.
4. Signal wire shield was connected to ground at one end.

Attempt to troubleshoot:
1. Changed signal wires and connectors.
2. Disconnected detector from AC power supply and attached it to a UPS.
3. Disconnected Nelson box from AC power supply and attached it to a UPS.
4. Disconnected UPS from the power supply so that it was running off of its battery.
5. The detector was isolated from the LC system (disconnected inlet and outlet) and at times it was powered off.
6. This noise is very small (0.1 mV) in amplitude and about 0.02 min period, but it is noticeable and not usually present.

We spent a few days looking into this and then all the sudden it cleared up and stopped.

Situation Two: 12-11-06:
1. Shimadzu 2010 HPLC with built in UV detector.
2. Channel A output going into PerkinElmer 970A Nelson Interface box via the RCA wire at the back of the instrument.
3. Data collected on TotalChrom.

We noticed the same type of noise during an entire run (16 25-minute injections) and then the next morning it was gone. This system is in the same lab but on the other side of the room. I am not sure if it is on the same circuit or not but I don't know if the noise is coming in through the AC line or not because on the first incident we had everything plugged into a UPS with it diconnected from the AC line so everything should have been issolated. Other people were running at this time and they did not see this type of noise on their runs. It may be picking up some type of interference but we can't think of what it may be.


A scanned pdf of the file can be found at the following QuickSharing link for download :
http://s28.quicksharing.com/d/2835863/1 ... _Noise.pdf


I appreciate any ideas.

Thanks!
Jason

[Mark Tracy]

Your pdf file is unreadable. Try uploading it again please.

[Jason Hendricks]

I just checked and I can read it...on this site you have to click on download file for it to open. Let me know if you have problems.

Jason

[Mark Tracy]

I got it on the second download. You are right, this is strange. The period is much too short for the ususal things like the pump. When you see the interference, does changing the sample rate on the Nelson box make a difference? Another possibility is ground loops. You want the shield on your signal cable grounded at one end only. Those are about the only ideas I have.

[gbalock]

Jason,
Mark seems to think the problem is electronic, but the only time I can remeber seeing a baseline like that is when I had a bubble in the flow cell. Does the noise go away when you turn off the flow? If so, then it probably is a bubble. This can be corrected by putting a flow restrictor post detector. Also, pumping a high organic MP seems to wet the cell better.

If the problem is electronic, is there a stir plate close by? This may expalin the intemittent nature of the proiblem.

[Jason Hendricks]

Thank you both for your ideas...

When you see the interference, does changing the sample rate on the Nelson box make a difference?

That is an idea I didn't think of. If we see this again I will try it. I believe the last time it occured we had the sampling rate on 3 pts/sec and the first time it was on 1 pt/sec.

Another possibility is ground loops. You want the shield on your signal cable grounded at one end only. Those are about the only ideas I have.

We have had grounding problems in the past so we learned our lesson a few months ago. The first time this happened with the 2487 we had a new cable on the signal output and we had the shield grounded only at the Nelson box. The last time on the Shimadzu the RCA cable doesn't have a seperate sheild wire so we only had the + and - hooked up to the Nelson box + and -.

Does the noise go away when you turn off the flow?

The second time this happened the other day we didn't have a chance to troubleshoot because the noise only occured during a sequence of 16 25-minute injections mixing water and methanol (freshly filtered and degassed with the degasser on) online under gradient conditions. In the morning we could repeat the noise under the same conditions so we had to give up troubleshooting the problem.

The first time this happened using the Waters instrument we had the problem for about three days so we had some time to extensively troubleshoot the system and during this process the LC and detector were both off and we still saw the noise so that is why we concluded that it was electronic.

If the problem is electronic, is there a stir plate close by? This may expalin the intemittent nature of the proiblem.

That is another good idea that I did not think about. There is a stir/hot combo plate in the lab on the other side of the room (I am not sure if it is on the same circuit or not because I don't know which benches are on which circuits.) I will have to check and see if it is in use when/if this happens again. I do know that the first time this happened on the Waters instrument it was during our PM so I had a flowmeter hooked up and it seemed that the amplitude of the noise was greater when the flowmeter was plugged in. When I unplugged the flowmeter the noise was still there but at a lower amplitude. This seemed repeatable, but it never went away even with the flowmeter unplugged and removed from the vicinity. On the Shimadzu instrument the flowmeter as nowhere nearby because we were only running a simple run to train a new user.

Also we are wondering if it is somehow related to constrution that is occuring in our building. Over the past several months our institute has been remodeling an area in the same half of the building. Maybe this could somehow be related?

I appreciate your help,
Jason

[Mark Tracy]

Since the period is about 1.6 seconds, look for something in the environment with the same period. Also, does shorting the inputs on the Nelson box have any effect?

[Jason Hendricks]

I will look for something else in the lab with that period...good idea. We did short the nelson box and it appeared to make the noise disappear (if I remember correctly...we did this on the Waters LC a couple months ago).

We believe that the signal wire may be acting as an antenna picking up something but it seems too low to be RF interference.

Jason

[Sallybeetle]

I have seen this type of problem in the following situations:
1. Bubble in the detector.
2. Gradient run not plumbed through post-pump external gradient mixing chamber.
3. Post-pump external gradient mixing chamber not plugged in to electrical outlet.
4. Air conditioner feedback through building wiring.
5. Sometimes never figured it out.

Hope this helps.

Michele Sanders
Director, Analytical Development
Sparhawk Laboratories, Inc.

[Jason Hendricks]

Thank you for your response Michele. I will look into these things. Both instruments mix online via a gradient proportioning valve on the low pressure side of the pump and I feel that this is working properly because the runs were very very tight. (0.07% RSD for RT and Area on the Shimadzu 2010).

The AC noise might be a thing for me to look into. If this happens again for any significant amount of time I might check with our building supervisor to see if anythign could be going on.

Thanks again,
Jason

[unmgvar]

If you suspect the PerkinElmer 970A Nelson Interface box why have you not taken it out of the loop?

connect one of those instrument to their original software without the use of the box next time this happens to you. this can be easely set up

[Jason Hendricks]

Thank you for the suggestion unmgvar but in our lab none of the instruments have software. We control the instruments via the front panel controls and we output our data to TotalChrom and the Nelson box is the only way to convert the analog signal to a digital signal that TotalChrom can read. When the problem originally occured we suspected the PerkinElmer box and we changed it out with another box which gave us the same amount of noise.

Thanks,
Jason

[unmgvar]

still you are not ruling out that it is either something influencing the boxes or something else in the system.
can't you loan an original software from the tech people of your instruments? install it has a backup on the pc conected to the instrument and switch between the 2 when it happens to you again.

[Jason Hendricks]

Unfortunately in our lab we don't have any PCs connected to the instruments. We have a client-server setup. I agree the boxes may be acting as an antenna picking up somethign but I don't believe we have any way around it. The next time it happens we may try one of the new PerkinElmer NCI boxes that we have on a few instruments. Maybe these are designed to filter out some of that noise?

[unmgvar]

but you could have a stand alone pc ready to use then with the original software?