Sujet : Re: Instead scopes
De : alien (at) *nospam* comet.invalid (Jan Panteltje)
Groupes : sci.electronics.designDate : 01. Sep 2024, 08:36:34
Autres entêtes
Message-ID : <vb15i5$20n23$3@solani.org>
References : 1 2 3 4 5 6 7 8 9 10 11 12 13
User-Agent : NewsFleX-1.5.7.5 (Linux-5.15.32-v7l+)
On a sunny day (Sat, 31 Aug 2024 11:57:03 -0700) it happened john larkin
<
jl@650pot.com> wrote in <
t5p6djh6vdfl5kmosb54bdpnn7e9np0smg@4ax.com>:
Here's my millikelven temperature controller.
https://www.dropbox.com/scl/fi/emxfaurnyj35t0y84tvwl/Oven_Cables_pub.jpg?rlkey=jpcnmnt1pcooz9nj7d0j4rah0&raw=1
That's a dual-stage Mach-Zender e/o modulator whose extinction is
much better if the temperature is stable to milllikelvins. The big
hogged-aluminum box is heated by six mosfets on the bottom, and there
are four thermistors for feedback. It runs at 30C.
>
Why four thermistors? The only temperature that matters is that of the
modulator itself. People who need to minimise temperature gradients do
need more than one temperature sensor, but it isn't clear why you would
have. I used a second thermistor to monitor the temperature of the
exhaust side of my Peltier cooler (which does matter) but you seem to
have used mosfets as resistive heaters which is rather easier.
>
One thermistor is on the heater board, on the bottom of the big block.
Three are on the platform that mounts the e/o modulator. We really
don't need three up there, but we wanted to error check and snoop for
gradients and optionally do some averaging if we had noise.
>
The EOM platform is spaced off the bottom of the big block, which
makes us a 2nd order thermal system. The main block has a 75 minute
time constant, and the platform inside is 17 minutes. Our control
algorithm uses the difference as, essentially, a derivative term.
>
Coolers have lots of problems, including condensation. Heating to 30C
worked fine. This is in maybe the world's biggest single clean room
and the local air is always 20C.
Decent temperature stability over several years:
https://panteltje.nl/panteltje/tri_pic/ https://panteltje.nl/panteltje/tri_pic/tritium_decay_experiment_black_box_open_on_main_PCB_with_beeper_IMG_3874.GIF https://panteltje.nl/panteltje/tri_pic/tritium_decay_experiment_closed_black_box_IMG_3868.GIFMuch better time constant, much cheaper, took an hour to code, maybe 2 to build
Uses less power, can run on batteries if main fails,
:-)