Thermocouplers anyone?

[G. Hoffman]

OK, I'm curious if anyone here knows anything about thermocouplers?  More to the point, designing with them.

I'm working on a guitar design that uses a top which is laminated with hide glue, and clamping it will take much too long (I think, I haven't actually tried it; but I'm pretty good at estimating clamping times, and hide glue has an open time which can be measured in seconds), so I need to be able to heat it up under while clamped up in the mold.  This is related to a fairly traditional way of doing laminated surfaces, but this one has too many curves for traditional hammer laminating. 

So, what I need is to heat the entire workpiece to 150 degrees, plus or minus 5 degrees, but no part of it can ever go above 155 degrees.  Because I'm familiar with them, I'm planing to use a silicone heating blanket, which would easily get WAAAAAAAAAAAAAAAAY to hot, and what I'm thinking is that I can use a thermocouple to watch the temperature of the blanket, and then another to measure the heat at the edge of the jig.  Coordinate with a Microcontroller, and turn the blanket off if it goes above 155 (actually, probably 160 so that it heats up a bit faster, but it is critical the hide glue never goes above 155), and on again if it goes below 145.  When the thermocouple at the edge reads 155, let it cook for 15-20 minutes while tightening the clamps even more (because of glue squeeze out), and then turn the whole thing off.  A lot more complicated than using a man made wood glue, but hide glue dries harder, which seems to sound better.  Oh hell, it DOES sound better, though I can't prove it scientifically. 

So, what I REALLY need to know is, how do you use a thermocouple?  I've been looking around on-line, and I've watched the recent eevblog about their use, so I've got some ideas.  The two things I'm most seriously considering right now is either using a MAX31855 thermocouple-to-digital-converter, or just buying a couple of Microchip Thermocouple development boards.  Either one gives me the data over an SPI bus (which I've used before, for other experimental stuff).  Since this would be a one off device, using the development boards has some appeal, to be sure, but then I do like doing this stuff myself as well.

I don't need super high precision - if it goes up to 157-8, that's fine; and as long as it gets above 145, it's fine - and the resolution can be pretty rough, but I do need it to be fairly consistent.  Any tips?


Gabriel

[vattern]

Why not just get something like :http://dx.com/p/digital-temperature-controller-66334 $18 ...

[G. Hoffman]

Why not just get something like :http://dx.com/p/digital-temperature-controller-66334 $18 ...

Seems a bit easy. 

I'll have to look into more industrial controls, I guess.

[Nasse]

I think in the old days at the uni guys did diy thermocouples by charging a suitable cap to certain energy to weld those two wires together... needed some calculating or experimental...

SOme other probes have tooo much inertia or are just too slowly, but nicely done thermocouple is fast and can go high as house fire or brick or ceramic manuf... National semi used to make some ic to connect thermocouples in the 80´s wonder if such interface is still availabl... I know some fire safety regulations measures have too rounded edges just because they used probes with inertia...

sorry my non eng language

how bout a room with adjustable highest temperature like finnish sauna, you could set it for many hours if needed

[Gus]

As posted by vattern a PID type temp controller might be the way to go.

A solid state relay(SSR) control of a heating system with a type K thermocouple(TC) or RTD etc attached to the heating blanket.
The trick is going to be how to attached locate the the TC for good feedback.

Some controllers have a learning mode so you don't have to set the PID etc settings.

It sounds like a temperature controlled enclosure(heating the air) might be another way to do what your want

Look at this link http://www.omega.com/.  Good information

There are other sites selling controllers SSR's, TCs, RTDs, extension wire etc.

[Brossman]

Last year, I built a rig that connects 45 K-type thermocouples to 15 digital readouts for an organic geochemist in my department. Seems pleased. He has them in some high pressure tubes (~3kBar) at 60-100°C. The couples have to be welded at the tip. The readouts are pricey (IIRC, ~$250).

However.... I have a DMM that came with a K -type thermocouple and has a temp setting 0.o ...I recall paying about $50 for that DMM from a website that seemed to specialize in DMMs. It also has a capacitance meter too  ... anyway...

[Seljer]

A fellow in the lab I'm writing my diploma in is doing some stuff with thermal modelling of losses in electric motors. He frequently has to rig two dozen thermocouples up to different points on a motor, He uses an Agilent DAQ unit with a relay based multiplexer addon card. When its running you can hear the clicking of all the relays every second.

[Jdansti]

Something that I haven't seen mentioned yet is the issue of wiring thermocouples. If you buy a thermocouple that has a connector on the end that can be inserted into a socket on a meter designed for thermocouples, you're ok. This is because thermocouples produce a voltage based on the temperature differential between thermocouple junctions. Normally, one junction is at the end of the two dissimilar  metals and the other is inside the meter where there is another junction (actually two). The meter has a thermistor in it to measure the temp of the internal junctions. It uses the voltage of the thermocouple system and the internal junction temperature to calculate the temperature at the external junction.

But if you add wiring to the thermocouple, and the new wire is different than the thermocouple wire, then each new joint junction becomes a thermocouple and the measured voltage will be affected by the temperature of the new thermocouple joints junctions (unless you know the temperature of the new joints junctions and compensate). This applies to external wiring and the wiring inside of an enclosure that you might build to house multiple thermocouple sockets and a switch for example. There are ways to deal with this, though.

[Jdansti]

Something else to think about is how to place the thermocouple so that you'll get the best measurement.  There are various ways of doing this depending on the type of surface you want to measure. Here are a few examples:

[Brossman]

This applies to external wiring and the wiring inside of an enclosure that you might build to house multiple thermocouple sockets and a switch for example. There are ways to deal with this, though.

Sorry if this is OT, but...

o.O are you kidding me? Does that mean that the rig I built is bollocks?

[Jdansti]

Did you install copper wire and sockets/switches made of various metals between the thermocouple wiring and the meter?  If so, every junction of dissimular metals acts as a thermocouple, similar to the junction you are using for measurement. The magnitude of the EMF of the new junctions will be a function of their temperature and the make up of dissimilar metals. That's illustrated by looking at a table of thermocouple types ( K, J, R, etc. ).

This write up does a good job of explaining it, in particular the section called, "The Problem of measuring temperature with thermocouples":
http://pcbheaven.com/wikipages/How_Thermocouples_Work/

Other info:
http://www.omega.com/temperature/Z/pdf/z021-032.pdf

http://www.w-dhave.inet.co.th/Thermocouple.html

Believe me, I ran into this problem a few years back at work and it took a while for me to understand the thermocouple "rules" after much discussion with an EE.

[Brossman]

Well I spoke with the lab coordinator. He told me he knew of this. He said that its fine as long as those junctions are the same temp as the rest of the line except the measuring point. As a means to ensure this, he installed fans on either side of the chassis to pull fresh air through, keeping those junctions effectively nullified.

Phew, I almost thought that this year's research was screwed.... pheew again!!

[Jdansti]

Well I spoke with the lab coordinator. He told me he knew of this. He said that its fine as long as those junctions are the same temp as the rest of the line except the measuring point. As a means to ensure this, he installed fans on either side of the chassis to pull fresh air through, keeping those junctions effectively nullified.

Phew, I almost thought that this year's research was screwed.... pheew again!!

Glad all is well!