re. What got up my nose was Phil Hobbs taking the chance to exhibit his bro ad-spectrum expertise where it wasn't actually serving any useful purpose ( except in image building, an activity which Phil does seem to enjoy).
king
on..
t, not all that parenthetical.
Phil - who both told you that the first temperature sensor in the loop sho uld be put as close as possible to the heat source you are controlling, to keep the loop as fast as possible. Phil then went on to point out that a se cond temperature sensor close to the thing you want controlled can make sen se - you integrate the difference signal and use it to offset the primary c ontrol loop to give you exactly the temperature you want at point you want controlled.
good idea, of the kind we expect from Phil (and he will almost certainly h ave tried it somewhere).
Hi Bill, yup, I know the arguments. Most of the time we try to keep heater , thing and sensor all as close as possible so the question is moot. I did try it (several years ago) with a permanent magnet thermal control. TEC's and heatsinks outside of magnet yoke, long (1-2 inch) copper rod that connects to magnetic material...(there's a thin SS "magnet gap" in the ther mal chain I can't recall exactly where we ended up putting it.. but it's there for ma gnetic reasons.) And then a thermistor shoved dwon a hole in the copper rod to be as close to the magnetic material as possible. The time constants involved were mos tly terrible.. hundreds of seconds.. (I'd have to check my notes.) So doin g a lot of experiments was painful, (time wise) And there was a several thousand second time constant to drag the whole ste el yoke of the magnet around. (That was basically ignored in the design.) I tried sticking the thermistor right into the cold plate (opposite side fr om heat sink to air) of the TEC. (That was certainly much easier mechanically .)
The time constant for the system to track changes in the outside air was hard to quantify (and to control... dang air currents.) However there was also a set of gradient coils and field sweep coils stuck to the magnet pole pieces. (creating heat.) And the time constant for dealing with that changing heat load was ~ twice as long.. since heat had to go up and back d own the copper rod... (I've said all this before, but it's not always bad to r elive past battles.)
Anyway, where you put the temp sensor is not all about speed, but also where the source of thermal forcing is coming from. If a lot of the variable heat load is coming from the thing you are controlling, then stick the sensor near the thing.
For me this has been more common than a situation where the forcing is from the external air. But for things like thermal control of sensors..(I'm thinking of APD's) I can see it... and if I ever do one I'll try sticking the thermistor on the cold plate.
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.