OT - Temperature - CLOSED

[Gabriel]

Hi All,

Sorry to digress from CCS.

I need to measure temperatures at -80C on a medical freezer.
The sensor will be continuously operating at that temperature.
-55C seems to be the limit on most sensors and -40C for thermistors which is worse.

I have an easy SPI option with a K-Type thermocouple and a Maxim ASIC.
I have no Experience with RTD's and development time is NIL.

Any suggestions are appreciated.

G.
_________________
CCS PCM 5.078 & CCS PCH 5.093

[temtronic]

If there's an 'off-the-shelf' sensor/amplifier unit I'd go with that as -80C is really, really cold even for a crazy Canuck like me. As it's for a 'medical freezer' reliabilty is important NOT the cost of the sensor. Cheap is NOT a good idea.
As R&D time is nil, that says to buy 'off-the-shelf'. Something that you can easily source, have a couple of 'plug in' spares that do NOT need calibration.
The older I get the more I want to make life easy for ME ! Charge the client appropriately for a good, working solution.
I'm not to sure how you'd calibrate a sensor for -80C and I'm assuming it's +-1*C resolution.

Food for thought ?

Jay

[ezflyr]

Gabriel,

The Maxum MAX31855 thermocouple interface IC is really easy to use, and there are lots of breakout boards available (Sparkfun and Adafruit). I had mine up and running from scratch in a few hours! Order a good quality probe from Omega, and you will be good to go!
_________________
John

If it's worth doing, it's worth doing in real hardware!

[Ttelmah]

I'd suggest you talk to a temperature sensor specialist company.

Platinum resistance elements for instance, go down to only a few degrees Kelvin. The curve flattens badly at about 30K, but this doesn't matter for you.
Look at something like:
<http://www.gekon-trading.de/en/products.htm>

Nice thing about these is that they will operate up to several hundred C, and down to something like -250C. Their 'standard' types are only certified to -50C, but this is just the casing, and calibration. I've used type to -200 from the same company.

Silicon diode sensors like these (UK company, but they have US suppliers, and there are US companies doing similar products):
<http://www.omega.co.uk/subsection/superconductive-wire-cryogenic.html>

have ranges down to 4K, and up to a couple of hundred C.

However cheapest/simplest option is something like a Pt100 element:

<http://www.tcdirect.co.uk/Default.aspx?level=2&department_id=245/1>

These are specified to -200C, and reasonably priced.

[Gabriel]

Hi All
Thanks for the suggestions!

im going for this:
http://www.mikroe.com/click/thermo/

its has the MAX31855 that Ezflyr mentioned and a thermocouple by Omega as well!

This post has given me the confidence needed to go forward with this sensor.

Thank you all!

G.
_________________
CCS PCM 5.078 & CCS PCH 5.093

[temtronic]

Wow that's a neat little sensor board and cheap too !!
Perhaps once you've made the driver you can add to the 'code library' ?
Oh yeah...remember it's a 3V device !! Sigh, seems darn near everything is 3 volts.......

Glad a solution was found and in under 24 hours !!

Jay

[Gabriel]

Hi Jay,

I've been using their products for a while now and I've been very impressed with their quality.
I'll post my driver when the boards arrive.


G.
_________________
CCS PCM 5.078 & CCS PCH 5.093

[joergn]

I have just made a 4 channel 31855K sensor board with a 18F45k22 and an FTDI USB 232R chip and control this over virtual com port driver out of Visual C++ for temp data recordings. FTDI chip gives you 3V3 for all the ICs so minimum parts required. The thermo click from mikroe seems for me to have a missing emi filter for the K-sensor wires itself - compare it with Adafruit. If I want to test for instance inside a switch mode power supply the heating of power transformer I don't want to get the noise coupled into the 31855K inputs. A 2m probe can collect lots of noise, if you have bought that expensive item also. It was claimed that the 31855K has some poor performance at very low temps. I can not cool down that low just down to -25°C and realised that at my lowest temps bad sporadic positive readings occur. Maybe inside a buffer it results in bad averaged values? So better to sort this out of the statistics rgds Joerg

[asmboy]

Two years ago i did some work with a thermo-acoustic refrigeration system being developed by a physicist i've know for years

-80c is 193 K and in the sweet spot really for
a part i've used:

http://www.lakeshore.com/products/cryogenic-temperature-sensors/silicon-diodes/dt-670/Pages/Overview.aspx

paired with an Analog Devices low offset opamp and read by precision 16bit LT1609 ( i already had a nice driver for) plus a bit of PIC code and the client was very happy .

the part above spans your range (darned near CENTERED on it !) with excellent linearity.

[Gabriel]

Hi Asmboy, them diodes look good... and thing that those are exactly what the freezer I'm measuring uses!

Since this thread has been resurrected i will mention that the thermocouples worked just fine.
Noise was not an issue but i do alot of SW filtering.

Temtronic, you can find my driver on the library... its under another users post which i hijacked.
http://www.ccsinfo.com/forum/viewtopic.php?t=49599

It took a while to get it working after bad chips and nist tables.

G.
_________________
CCS PCM 5.078 & CCS PCH 5.093

[asmboy]

clients $$ not mine for the parts. low self heat number etc - ideal for desktop thermo acoustic sterling engine refrigeration control.

[joergn]

I am using also Diodes for temp sensing and will soon have the US granted Patent doc in my hands - after paying the fees....... ;-)

But those are Power LEDs and I made an even simpler method than NXP patented before.

You can find it on my web site http://www.led-temperature-protection.com
I have soon to update the site, when I have done my data collection performance docs ready.

I started this with a small LED buck regulator out of a PIC18F45K22, it can not go that fast with the DCDC frequency but it was fast enough to get some results indicating that the function is working.

The CCS compiler was a great help for me in the past to get all those direct PWM regulators working ;-)