Thermal Pressure Gauges

The single-wire pulsed thing is cute, because it can correct for ambient temperature, too. It's just slow.

? Well I hope you can put in the implied question marks.

Much as I dislike TC's a temperature difference makes me think of a TC on the heater and another stuck on the vacuum can as ambient. Then I get the difference.

George H.

How about a tungsten filament as Pirani... a ton of hits

Geo,

How about two thinfilm surface-mount RTDs on a PCB-type substrate, kapton flex maybe. Kapton is a rotten heat conductor. Put a heater next to one of them, a resistor or a zigzag trace. Maybe use those cool SOT-23 Zetex nickel RTDs.

Make it t-shaped, with the RTDs on the wings and the middle thing a long connection pigtail that mates with a vacuum feedthru connector. Mounting hole somewhere.

Or do the pulse thing with one RTD and a bunch of math.

I guess radiation cooling will be a problem.

Some employ a bridge in the manner of Sir Charles Wheatstone FRS, the sensitive element/heater being a fine wire of Wolfram encased in Element 79.

There was an old publication, where someone hacksawed a metal-case tube open and fitted thermocouple elements to the tube wires, then soldered or brazed the metal case to his vacuum fitting. So, the usual run of thermocouple gages use an 8-pin octal tube base, because the commercial units started out as copies of his homebrew device.

The nice thing about thermocouples for temperature differencing, is the absolutely perfect zero you get at zero delta-T.

So the heater is only on one side of the Tee? I'm not really seeing what the second RTD is for.

Yeah, I was thinking maybe a nickel plated brass tube with RTD stuffed in one side and heater in the other. Radiation goes as T^4 and the gas conductivity ~T^1 so there should be some sweet spot where the radiation is small.

(If I did different temperatures, maybe I can tease out the radiation part.)

Then I just have to think about conduction down the wire leads (Phosphur bronze... or maybe even stainless?)

George H.

Yeah, I've got some small tungsten bulbs resistance of ~120 ohms at room te mperature. (You can see the resistance increase as you measure it with the DMM... Time constant of a few seconds.) Resistance is ~ proportional to the absolute temperature. So not as much sensitivity as I might get from a 10k ohm ther mistor. Hmm I could use two of them though and leave one in it's bulb. Th at might help with sensitivity at low pressures.

The one element trick is cute, but I can't help think that separating the h eating and sensing give me more options.

George H.

Yeah I like that it gets rid of the effect of changes in ambient temperature. (I don't have a good feel for how much of an error that is... I guess it gets smaller as I raise the operating temperature of the thing.)

What I'd really like is a decent "zero". I don't like the big DC offset in the thermistor /Pirani circuit. And thermal couple gauges are a bit flaky near ~10 milli-torr. (Wiggle the contacts or bump on the scale and the zero moves around.)

I'll have to put in some numbers.

George H.

For a really good 'zero' at zero pressure I imagine you're looking at creating a reference vacuum. Otherwise a bridge and some compensation.

Speed of response is a big benefit. The time constant gets traded away like the capacitance of a PD with constant bias when you servo the temperature.

Maybe manganin.. 3 thou should be ~80-90 ohms/m.

Also handy for making precision WW resistors.

I guess the least radiation happens for a small diameter, shiny (like platinum!) wire.

A wire filament is really not bad. A thin wire will have a low thermal conductivity. A broken light bulb starts to look pretty good... just hack half the bulb off with a Dremel, or even just punch a hole in it.

I haven't seen that; the heater is on a regulated current source (often AC) and the sensor wiring hasn't ever been bothersome. The lower vacuum range, we always used ionization detectors.

An eBay search for "Hastings vacuum" pulls up lots of dissectable examples of the commercial thermocouple gauges.

The car people have a different problem - they want to measure the mass of air flowing into an engine, not pressure, and at much higher pressures than you - but they seem to have settled on a separate sensor and heater. Their sensor has to get fairly close to the right answer to start with - I don't *know* but I think it's on the order of 0.5-1% - but they also have another control loop later (O2 sensor/injectors) that can compensate for some errors in the mass air flow measurement.

Their product volume is such that they can have whatever they want, and whatever it costs gets multiplied by 12 million a year (in the US), so there must be a reason that they spend "extra" money on the two separate components. :)

Matt Roberds

Yeah this is some cheap TC gauge. But to be honest, that's just an excuse. It's much more fun to build it yourself.

I put the Thermistor circuit together with to 10k (T= 25C) ones. (putting in numbers the V^2 term adds up fast.)

I didn't measure room temp so delta T is a bit uncertain. (No vacuum so I stuck the thermistor in a piece of foam.) at 2.5 k ohm (~60C) it stabilized at 19.1 V ~145 mW. I was thinking I would like to throw away more power. Maybe some Aluminum foil stuck onto the thermistor?

George H.

Yeah lots more options when separated. I was thinking the loop might be better if it was linear in power. A FET heater a constant V with current control. Then let the thermistor measure T without the heater stress... I mean, you could worry that current/heat changes the R and the calibration.

How do they (auto industry) measure flow? A know impedance and measure the pressure drop?

I could tell a funny story about flow impedance and my first Helium-4 pot. I made a factor of 10 mistake in my calculation.. converting from Pascals to atm. or Torr.

To make it you crammed nickle wire (I think) into thin walled stainless. (really small ID) making a factor of ten mistake I crammed x10 more than needed. (A lot more work!)

Worked for s*it, which sent me back to my calculations.

Measure twice and cut once, George H.

The Bosch document that Lasse linked claims +/- 4% for one series of parts and 3% for another. I wasn't *that* far off.

Lasse linked you the Bosch document, which has a good explanation. Older ones had a wire sitting in the airstream as the heater, which would sometimes get crud on it and affect the calibration. One partial solution was to heat the wire hotter than normal for a minute or so, after you shut the engine down, in an attempt to burn off the crud. When the car had a *lot* of miles on it, sometimes you had to clean the crud off yourself with solvent spray, or a very gingerly applied Q-tip (the wire was pretty small gauge).

The old way (1970s) was a flap vane connected to a potentiometer. Yes, really! It fed an analog fuel injection computer.

Some cars don't have a mass air flow sensor; they measure the temperature of the incoming air, the intake manifold pressure, and the RPM. The computer knows the engine displacement, so it can calculate the air mass from all those numbers.

I've gotten kHz and MHz confused before, which has caused my code to do strange things. :)

...and have enough stuff on hand to do it over.

Matt Roberds