Signal periodically fades

[len]

I have a HP 5985 which I have salvaged of eBay thats been stored for about 4 years. I worked on the vacuum and got it to about 1.8 e-6 at which point the galaxy manual says spectra can be done. I repaired a blown RF drive transistor and resistor in the ion focus supply at which point all optics and RF voltages read as they should. The galaxy scan box and main computer I got rid of as it needed a dongle and wrote a brand new GUI for it in visual C++.

At this point I was able to tune a spectrum. I have large signals at 28 and 32 (1/2 of full log-scale, correspondig to my vacuum I suppose. Trouble is after I let in a sample (I use acetophenone which has about the same vapor pressure as PFTBA) the sample peaks appear, and then the whole thing fades, almost all peaks disappaer - they do not shift - just colapse into the noise where they were. The source current stays as programmed - as otherwise the MS reports a fault. After I close off sample, the air peaks reappear taking about 1 min, and oscillating in and out before establisng themselves - but always at the same values 28 - 32. I guess this excludes the RFPA, and source cleaning is in order? Has anyone observed such behaviour.

Another question I have is to do with the EM. I read after two years it should be replaced. But what are the indications for this? My baseline (at no source current) is very low and flat, while the background N2 and O2 is enough to give 1/2 full log-scale - so is there any reason why one would want to replace the EM? I did need to tune a low x-ray voltage to get a spectrum - about 50V. At about -200V xray voltage the spectrum disappears completely and is replaced by off-axis noise (its not thermal noise as the noise disappears completely when ionisation current is turned off). I do not know if this is normal.

A third question is to do wth the pumps. I turn them off after work to extend their life and the life of the motor. Is that generally done? The motors are old - the E2M2 looks like its from the 1970's.

[len]

Nobody answered my question, but I now solved the problem, so I thought Ill answer it myself should anyone experience similar symptoms.

The problem was a 'dirty' source and slightly bent rhenium filament.

I opened the source, completely disassembled it. There was a thin carbonized layer on the input and output slits, as well as on the injector, and the rhenium filament had a very small kink in it. I scraped of the thin carbon film (which was not blocking) the slit and injector, and etched them in HCOOH/H2O2 solution till the latter started boiling.

It was difficult reasembling the source - in particular picking up the thin ceramic spacers (20 in all) which separated the draw out and ion focus electrodes was difficult - and one was broken by just slight tweezer pressure. It could nevertheless still be used as spacer and inserted at the end. The glass tube shileded wires (about 15 in all) connecting back to the terminal 15cm away was also a challenge.

In the end all was assembled. The pressure now went down to a factor of 10 lower to 2 x 10^-7 torr. The noise started dropping markedly, until after 5-6hrs worth of operating at that pressure with source at 130C, it was all below the floor. Letting in acetophenone now to aboput 2 x 10^-6 Torr residual (the standard inlet did not have a flow control so it all had to be controlled via the valve - which is difficult) gave a perfect acetophenone spectrum at 120, 105, 77 (+round), 68, 51. The 68 peak was dominant, which is the only difference from standard - but the acetopehnone was old. There was no longer any fading, and the electron multiplier voltage was reduced to 1300V - which shows that they are OK to store for years even at atmospheric pressure.