AD590s on both thermal shields are not working. I was wrong when I checked them at the first time.
The temp sensors in the vacuum tank for monitoring temperature on heat shields are wired as shown in the picture. The resistor,R, is 30k ohms. According the the datasheet, the current from AD590 should be ~ 300uA, (30kx300uA = 9V). But what I read from the voltage across the readout R was 20V which was over the input range of EPICS (+/-10V). This happened on both of them. I compared the readout with a left over AD590, and got ~ 9.3 V readout which was expected at room temp.
At first I thought it might still be working linearly and useable if I just switched to lower R. However, with R=12 k, the readout voltage was 18V (I expected 20x(12/30) =8V). So certainly, this is not working.
I think the reasons they are broken is that they were overheated when I soldered them. I tried to be careful, but, apparently, that was not enough.
I'll check if there are spare AD590s in the lab or not, otherwise I'll order some more.