Thanks Jan, (I couldn't get that link to work) I'm expecting to need more resolution than a thermocouple can give ~40uV/K... I'd like to be able to see 1-10 mK temperature changes.
Back in the dark ages (gradual shcool) we used Allen Bradely carbon composite resistors as temp sensors... It's been a while, but as I recall those only worked well at low temperatures. (maybe 50 K and lower) They also drifted a bit from run to run.
George H.
Hmm, thanks I'll order some.
Sorry, I just don't know. I've only got LN2 here every few months, maybe I can send one to Jan and he can put it in his cryo-cooler. (Or maybe he's got some in his parts box?)
The ones with lower forward voltage (upper curve in the plot) start falling off the linear line at ~12O K. Perhaps that's a hint something is 'freezing out'? But even if it looses it's 'diode connected transistor' action you've still got the B-E junction. I'd expect that to work down to 4K similar to lakeshore diodes.
George H. George H.
They do something 'different' below around 25K, with Vf increasing a lot faster with temperature.
Even 1N4148s apparently do strange things at 4K (like turning into a not-bad ICO), and diode sensors are known to occasionally have one or more negative resistance spots on the I-V curve.
The sub-77K area is like another world in some ways.
Yeah I've seen the curve from lakeside, but never any discussion of what the physics is behind the rapid increase. (What's the binding energy of an electron to a dopant atom?)
What's an ICO? (current controlled oscillator?)
I haven't done any liquid helium stuff in a long while. And only started playing with LN2 again a few years ago. Given the cost of helium and cryostats, I guess a few hundred dollar temperature sensor is lost in the noise.
George H.
Yes.
Yes. The calibrated germanium ones are a little harder to lose in the noise, but it's possible to see single-digit microkelvins.
On a sunny day (Fri, 16 Mar 2012 07:48:48 -0700 (PDT)) it happened George Herold wrote in :
OK, my fault, that was localhost, thsi shoudl work:
rge
but
if
Phil H. has a nice chapter in his book on thermal stuff. For a while a draft copy of the Thermal chapter was on his web site. I'm pretty sure he wouldn't mind if I sent you a copy. I guess you start with more than ten times lower dV/dT. Then for the diode since you provide the stimulus you could always do some coherent detection if the noise was a problem. (I've never done that with a diode thermometer.) A good voltmeter will measure a uV, that's better than 1mK without even breaking a sweat.
George H.
Tin-lead solder seems to hold up down to -77C..
On a sunny day (Fri, 16 Mar 2012 11:28:10 -0700 (PDT)) it happened George Herold wrote in :
Thank you very much for the pdf in the mail. I am reading it ATM, and found some interesting things I did not know. Now I understand why you want to use the transistor. Yes the smaller BDXXX packages should have a bit lower time constant. I have had problems with those packages in the past as the leads did not really seal hermetically, maybe a heating problem (expansion). They did seem more fragile than the TO220.
environment
point.
You are being dissed. It's up to you to figure out how to manage the organization to succeed at your goals.
?-)
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.