Non-linear DC amplifier?

** Yeah - why not use a type K instead.

Plugs direct into some DMMs or use a cheap adaptor.

What sort of temps are you trying to measure?

........ Phil

Fair comment Phil, but we need to use R-type for the higher temperature (working in the range 700 to 1400 deg C) and longer life.

One solution might be to take the R-type mV signal, add a bit and amplify a bit so it looks like a K-type signal!

A

maybe you would get some kind of idea looking to my CpuDiode measuring converter under Electronics on my site ...

--
                  Regards ,  SPAJKY ®
   mail addr. @ my site @ http://www.spajky.vze.com 
    "Tualatin OC-ed / BX-Slot1 / inaudible setup!"

Thermocouples are linear devices, all you need to do is adjust the gain and the offset.

I don't have the specifications for an R-type thermocouple, how many microvolts per degree?

what range do you want to measure

do you want output in 1mV per degree Kelvin, Celcius, or fahrenheit? or something else?

Bye. Jasen

"Jasen Betts"

** But only in New Zealand - Jasen.

On the rest of the planet they are non linear.

** Baaaaaaaahhhhhhhhhh

** No Google in NZ ???

........... Phil

Wrong.

that doc.

.Thermocouples are made by welding two pieces of dissimilar wire together and .using the welded junction as a temperature sensor. A voltage is created that .is proportional to the difference in temperature between the welded sensing .end and the other end where the measurement is being made.

sounds pretty linear...

.Actual thermocouples have Seebeck coefficients that vary widely over their .temperature range.

Bugger!

Only "mostly" linear over small ranges then...

Bye. Jasen

On Tue, 13 Dec 2005 09:01:19 -0000, Jasen Betts put finger to keyboard and composed:

In fact it *is* linear over the range (700 - 1400 deg C) and accuracy (+/-20 deg C) specified by the OP.

Here is a "Thermocouple Reference Table Type R (Platinum 13% Rhodium/Platinum)":

Using two roughly equispaced points (950C and 1200C), I came up with this relationship:

T = 74V + 221 where V = mV, T = deg C

The following table compares the temperature, T', predicted by the formula for various values of V (as in the above chart).

V T T'

--------------------- 6.743 700 720 7.950 800 809 9.205 900 902

10.506 1000 998 11.850 1100 1094 13.228 1200 1200 14.629 1300 1304 16.040 1400 1408

If the OP cares to do a proper linear regression over the desired range, the resulting line of best fit should be somewhat closer than my guesstimate.

- Franc Zabkar

--
Please remove one 'i' from my address when replying by email.

"Franc Dakar"

** It is NOT linear from 20C upwards - making calibration in the range 700C to 1400C a major problem.

** IF ** the OP had a reliable, independent means of calibration in the desired range, then a simple correction chart is all that is needed for his type R probe.

......... Phil

On Wed, 14 Dec 2005 19:31:18 +1100, Franc Zabkar put finger to keyboard and composed:

BTW, I believe you should be able to realise this formula using an instrumentation op amp with a gain of 74 and an offset (in mV) given by ...

221 + 74 x Vcj

... where Vcj is the cold junction compensation voltage.

For simplicity you might prefer a fixed voltage at a fixed ambient temperature of 25C, say, or you might like to use a device such as the LM335 to sense the actual ambient temperature.

If a fixed offset is good enough, then I believe you would need to add

0.141mV to the measured values. The formula then becomes ...

T = 74 x (Vmeasured + 0.141) + 221 = 74V + 10.4 + 221

The multimeter will display the temperature in millivolts.

- Franc Zabkar

--
Please remove one 'i' from my address when replying by email.

Thank you Franc. That is the sort of thing I am looking for. Can you send me a little more information. I only know enough electronics to have an idea of what's possible, but not enough to know how to do it. I'd be very grateful if you could send something to ceramics AT cof DOT com DOT au

Cheers

AW

Neat. I'll put that data in a spreadsheet and see what I can make of it.

(pdftotext gave me one row for each column in the table...)

Bye. Jasen