The 10V supply is an AD587 with a 22 uF foil capacitor connecting its "noise reduction" pin (8) to ground which I had sitting around from the doubling phase noise experiment.
The signal goes like:
10V x (1 - 9.1k / {9.1k + R(T) } ) =10 V x (1 - 9.1k/{9.1k + 1.2k + dR} ) =10 x (1 - 9.1k / 10.3k x 1 / {1 + dR/10.3k } )
which by taylor expansion is:
10 x (1 - 9.1 / 10.3 x (1 - dR / 10300 ) = 10 - 10 x 9.1 / 10.3 - 10 x 9.1 / 10.3 x dR / 10300 = 1.165 V - 8.6e-4 x dR/Ohms V
for a 1 kOhm RTD, dR/dT is about 4 Ohms / K so we get a total signal temperature sensitivity that goes like (after the 560 x2 gain):
V(deltaT) = 2 x (1.165V - 8.6e-4 x dR/dT x deltaT /Ohms V)
V(deltaT) = ( 2.33 - 6.8 x 10^-3 x deltaT / K ) Volts.
We plugged this into an epics channel and confirmed that it was being recorded.
Tara regularly makes ~ 0.1K adjustments to the temperature of the can in the course of the refcav beat experiement. This will show up as a 700 uV change on the 2.3 V DC signal. I'm not positive that it's this is large enough for us to see, but we get it for free. If we can't see this, we can try a 1K step for a 7 mV change, and if that doesn't work I'll put something less crappy in to readout the PT1000 sensor.
The DC signal read as 2.6 V on a scope and 2.43 V with the existing calibration of the frontend. The 1.2K number I plugged in is wrong by a bit. I trust the systematic calibration (as recorded by the frontend) to ~10%
**WHY AM I DOING THIS?**
Me and Frank disagree about what the radiative time constant is between the cavity and the can, in the case of both the cryo cavity, and the as built cavity in the PSL. (**see this elog**).
- Envelope physics says it should be on the order of 41 hrs.
- The reference cavity + can + heater at the 40m
**has a time constant of 4 hours**
- Frank has said "the radiative time constant from the can to the cavity is like 30 minutes," (though he may have since rescinded that statement)
- I looked through the PSL elog for some info about what these time constants actually are. I found:
- This elog where they do a step in the heater temp, but I can't glean anything clearly useful from it
- Another elog claiming a 7.5 hour time constant just for the heater where Rana made a few criticisms on the measurement
- The other salient point is that Frank has talked to some people (who exactly?) in both LISA and at UF, and seems to have been told "the only way to get long time constants is to have really good radiation shields with a bunch of layers." This doesn't seem to jive with the second point above, where envelope calculations give you a > 1 day time constant. It would be awesome if Frank could elaborate on exactly what is known here to help remove one person from
**the telephone chain**.
My money is that the thermal time constant is totally dominated by the conductive pole through the stack, and that's what we see (in the PSL lab and at the 40m), so depending on what the time constant between the can heater and the stack thermometer is, we might be able to rule some things out.
Is there any model for the thermal conductivity of the stack + suspension? |