Flow Proportional Counter Backflow Gas Regulation

[Probeman]

Yes, exactly.  All these regulators (if you read the company details) are vacuum pump backed for positive flow.

I had a similar (homemade) differentially pumped detector flow system at UC Berkeley in the 1990's based on Henke's design from LBL. It ran pure propane gas at sub-atmospheric pressures for better detection of nitrogen. Obviously the (separate) vacuum mechanical pump was vented to a fume hood!  30" of vacuum is all one needs for such a system.

But I am hoping that these regulators will run at 15 or 30 psi (absolute) and as you say, because they are referenced to vacuum, might provide more pressure stability.

Of course what we really need are solid state detectors for our WDS spectrometers!

[sem-geologist]

Of course what we really need are solid state detectors for our WDS spectrometers!

I am completely not sure about that. Why? (1) it probably could be ok replacement for low pressure counters; However providing very stable cooling of such detector is very challenging. We probably could use cooled nitrogen and same capillary to transport heat from the SDD. To cool SDD means there should be enough pressure of nitrogen, or lower pressure nitrogen but cooled to some low temperatures (then lots of liquid nitrogen is required). SDD and its counting electronics drift depends from temperature and humidity (We need to calibrate EDS SDD on SEM much more often than on EPMA, as the air conditioning is less sophisticated there, and SEM experiences more temperature swings (the SDD itself is cooled at stable set temperature, the decalibration is clearly a response to temperature and humidity variations by pulse counter unit (EDS SDD electronics))). This is why I am also not 100% sure that PHA drifts are directly produced alone only due to atmospheric changes, and rather suspect that electronics react to atmospheric changes too, in particularly that EPMA shaping and counting electronics is much less sophisticated from EDS pulse processing units, and low pressure counters most often works at lower BIAS and much higher GAIN (which means that gain-OPAMPs can swing the signal much more, than on high-pressure low gain spectrometers).

Partly problem with low pressure counters are the PHA drift. But if You use integral mode, there is no problem, as You still are counting same amount of counts (the counter bias can be set higher or lower in some margins which does influence pulse height, but not so pulse density).

SDD can look as that nice stuff, but there are many artefacts which would influence measurement linearity in particularly at low energies and is very hard to resolve:
incomplete-charge - no such thing on proportional counter
Si esc peaks - on proportional counter we have Ar esc, but for low energies it is not applicable.

Generally new design of proportional counter electronics could solve most of proportional counter problems (PHA resolution, pile-ups, increase linear throughput, intelligent bias offset depending from room temperature), while there would be no need to hardware changes inside spectrometer.

Actually there is no obstacle to use the last one NOW. Stick some digital thermobarometer (hygrometer) in the room. Log the values, apply some low-pass filter. Make correlation curves between bias and environment. Add some code to PfS to adjust bias depending from room environment -> profit.
   

[Brian Joy]

Below is a final update on my plot of the effects of variation in barometric pressure and atmospheric water content on Si Kα pulse amplitude distribution using JEOL P-10 gas-flow counters.  Previous plots weren’t contoured quite correctly because I’d made the approximation that the anode bias required to keep the distribution centered at 4 V varies linearly with respect to dew point temperature.  In truth it varies linearly with H₂O partial pressure, and the plot now reflects this.

As before, I’ve contoured the plot by eye, with contours represented by the following linear equations:

Channel 1:  bias [V] = 7 V/kPa * P_atm [kPa] + 7.5 V/kPa * P_H2O [kPa] + 905 V
Channel 4:  bias [V] = 7 V/kPa * P_atm [kPa] + 11.8 V/kPa * P_H2O [kPa] + 922.5 V

Over the relevant range of temperatures, P_H2O is given well enough by the following relation:

P_H2O [kPa] = 0.611 kPa * 10^((7.5 * T_dew [°C]) / (237.3°C + T_dew [°C]))

Based on scatter in the plot, it’s obvious that other variables have important effects on the pulse amplitude distribution at given count rate, but I don’t necessarily know what they are.  Time is undoubtedly one of those variables, as equilibration certainly doesn’t occur instantaneously.

[sem-geologist]

Brian,
That is really nice work! Do I understand clear - T in a room was all the time stable and changing humidity and atmosphere pressure were moving this dew point (so severely)?

[Brian Joy]

Brian,
That is really nice work! Do I understand clear - T in a room was all the time stable and changing humidity and atmosphere pressure were moving this dew point (so severely)?

No, temperature in the lab is not necessarily stable.  I have no choice but to work with the data I’m able to collect while I work to minimize temperature fluctuations.

Temperature is taken into account in the plot in the sense that, as temperature rises, the maximum dew point temperature also rises.  For instance, the points that plot as open circles were collected during the past two summers during periods when the air conditioning was overloaded.  Since the temperature in the lab was anomalously high (up to ~24°C), this allowed the dew point temperature to reach ~17°C (65% relative humidity at 24°C) on a few occasions.  Typically, with lab temperature at ~21°C, I wouldn’t expect to see values in excess of 15°C (69% RH), which is already way too high for a variety of reasons.  At any rate, most data on the plot were collected when lab temperature was in the range 20.5°C +/- 1°C.

For each X-ray counter, warm, humid summer days plot in the upper middle (black dots), while cool, stormy days plot to the lower left, and cold, dry winter days plot on the far lower right (blue dots).

[sem-geologist]

Brian,
That is really nice work! Do I understand clear - T in a room was all the time stable and changing humidity and atmosphere pressure were moving this dew point (so severely)?

No, temperature in the lab is not necessarily stable.  I have no choice but to work with the data I’m able to collect while I work to minimize temperature fluctuations.

Temperature is taken into account in the plot in the sense that, as temperature rises, the maximum dew point temperature also rises.  For instance, the points that plot as open circles were collected during the past two summers during periods when the air conditioning was overloaded.  Since the temperature in the lab was anomalously high (up to ~24°C), this allowed the dew point temperature to reach ~17°C (65% relative humidity at 24°C) on a few occasions.  Typically, with lab temperature at ~21°C, I wouldn’t expect to see values in excess of 15°C (69% RH), which is already way too high for a variety of reasons.  At any rate, most data on the plot were collected when lab temperature was in the range 20.5°C +/- 1°C.

For each X-ray counter, warm, humid summer days plot in the upper middle (black dots), while cool, stormy days plot to the lower left, and cold, dry winter days plot on the far lower right (blue dots).

Where we are in the UK, I don't think there's a dehumidifier in existence that could get our humidity below 60%.

I see improvement at summer with second (cheap) Air conditioning working as dehumidifier. It keeps RH below 60%. Well, We are in the middle of Europe, maybe that does not compare at all to wet climate. Also rooms for our probes are pretty small, it is enough just for 2 probes. I saw some pictures of other labs, and those looks spacious - it would not make me surprised that simple dehumidifier would not work. Maybe solution (financially possible) would be to put the probe together with AC and 2nd AC as dehumidifier in small space, leaving control PC and control panel outside (Something like plexiglass separation of the room).

[Brian Joy]

I see improvement at summer with second (cheap) Air conditioning working as dehumidifier. It keeps RH below 60%. Well, We are in the middle of Europe, maybe that does not compare at all to wet climate. Also rooms for our probes are pretty small, it is enough just for 2 probes. I saw some pictures of other labs, and those looks spacious - it would not make me surprised that simple dehumidifier would not work. Maybe solution (financially possible) would be to put the probe together with AC and 2nd AC as dehumidifier in small space, leaving control PC and control panel outside (Something like plexiglass separation of the room).

That’s a good idea, and it’s something that I discussed last summer with the department manager.  The lab is in the basement and is partially underground, but it has windows, and so a window unit could be a possibility.  Also, last summer it came to light that physical plant services had been adding the wrong refrigerant to all the AC units in the building; the resulting refrigerant mixture was highly ineffective.  If lab temperature and humidity are still difficult to control this summer, then I’ll probably push for a window unit.  Unfortunately, nothing is cheap at Queen’s -- we’re forced to work with contractors chosen by the university, and then the university marks up the price by something close to a factor of ten (no joke).

Other improvements (like venting heat from the building properly and pre-drying air) would help but are prohibitively expensive.  It seems a little silly to me to have to worry so much about heat and humidity in a place where a 30-degree summer day is considered a “scorcher.”