On Mon, 18 Oct 2004, Mark Scoville wrote:

>>> I don't know if thermistors will survive at 250C (No experience). I know
>>> RTD's will go that high - but they are usually a pain to
>> interface. I like
>>> thermocouples myself.
>>
>> Why are rtd's hard to interface ? It's just a handy valued resistor (100
>> ohms at 25C is common). All you need to read it is an ohmmeter
>> and a table
>> (supplied). Your trusty dvm will work fine if you don't need wire
>> compensation, and unlike thermocouples they do not need special chips and
>> outstanding high temperature insulated wiring skills (you can use the
>> chassis/oven wall etc as return).
>>
>
> Hello Peter,
>
> I only meant that RTDs are a pain to interface in comparison to a 1-chip
> solution. Recall the original poster was looking for a solution that was
> easy to interface to a PIC. I have interfaced both RTD's and thermocouples

The op was looking for an easy interface to a one off homemade smd oven. 
For one off, unless you buy a ready made dvm with thermocouple option and 
a K thermocouple supplied in the box you are better off with a rtd imho.

> in industrial products. I have used the MAX6675 - it does the job and is
> only 1 part between the thermocouple and the processor (PIC in my case).
> When using RTD's we always end up using many more parts (more board space,
> more costly to assemble, more opportunities for assembly mistakes). For me
> the MAX6675 is an elegant solution.

I agree 100% for industrial product use. Also for home one-off iff one 
already has the thermocouple paraphernalia lying around. The funny 
connector, high temperature cable, crimping set for prolonging same, and a 
sample chip.

> If you are aware of a simple way (1 chip) to interface RTD's to the SPI bus
> (including open/shorted RTD indication) please tell me what it is. I am in
> the design phase right now of another product which has both RTD and K
> thermocouple interfaces (customer picks which one they want to use). Right
> now the RTD interface we use has op-amps and generates a voltage that is
> read on a analog input of the PIC. Not very elegant if you ask me... but it
> works. If you know of a better way (less parts, direct digital to SPI etc.)
> lay it on me - I'm all ears.

The opamp solution is the one I know. short/open sensing should be done by 
same imho (too high or too low input reading). The advantage I see is that 
you can make a rtd interface with a section of a lm324 and 4-5 resistors 
and it will work and you can check it out against a LM35 at room 
temperature with your dvm. Low cost rtds for room temperature can be f.ex. 
low cost inductors (several ohms of dc resistance for f.ex. 10mH).

It's the lack of need to calibrate between units when changing or 
installing the sensor at the customer that makes rtds a good idea for 
production imho.

>> Were the wing temperature readings on the doomed shuttle coming from
>> thermocouples ? They went off scale *low* as the wing exceeded
>> temperature
>> probably cooking the wiring and shorting the signals.
>>
> Don't know. Interesting to think about though. Maybe they use thermocouples.
> However, I don't think that shorting together thermocouple wires will cause
> a low reading. Maybe if the wires shorted to something else funny things
> will happen. But if you short the two wires of a thermocouple together you
> just get another thermocouple which measures the temperature at the point of
> the short - don't you?

Yes but it depends on the wires. I assume that nasa knows what they are 
doing and would not allow something failing to cause a major misreading 
(in the wrong direction too) on an instrument (or on several) without 
providing some check.

Peter
_______________________________________________
http://www.piclist.com
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist