Thermocouples.

[bigegg]

Ive just ordered this:

http://m.ebay.co.uk/itm/100-240V-Di...ouple-Probe-/131720388818?txnId=1241202757003

With a view to building an electric furnace/kiln/burnout oven/heat treatment oven.

The thermocouple is a k-type nominally rated to 400c (which is ok for some of the uses i have for the oven)

From reading, the k type is a particular combination of metals so *theoretically* is good for the max 1200c i need to read.

Is there an extra component (resistor?) which restricts the temperature to 400c? Or is it a result of the wires/body of the thermocouple - and would therefore need mounting inside a protective sleeve (i believe a graphite tube?)

I know a thermocouple isnt expensive so this is a mental exercise as much as anything.

 

[tinman123]

if all that is cumin for £11 odd quid i wouldnt say its up to much tbh.

i used mypin pid's on a project last winter, even they are hard to get genuine even tho they are cheap at the best of times.
a propper pid would set you back hundreds on its own.

the pid should be able to hold a 400c temp if you have it set up right, it ill shoot past it initially but will sit right on it not long after.

 

[daleyd]

Normally it's just what the thermocouples themselves are tested to, if they come in an enclosed casing that may well be the limiting factor.

As you say k type in general (chromel alumel according to a handy guide I use at work) are good for up to 1372 deg c theoretically, though anything above 1200 deg we would normally perhaps look at higher rated alternatives.

There shouldn't be any resistor as thermocouples generate a very small voltage at the hot junction which is then measured by the instrument to give the temp reading. For the price they go for (couple of quid on the bay of e) you might as well get one rated to above the max temperature you will be using.

 

[ronan]

Thermocouples ? I remember the "Wheatstone bridge" from college, we seemed to do an awful lot with thermocouples, including measuring temperatures of a disused factory, over five floors. All the information was processed and stored on a laptop i seem to remember.

 

[bigegg]

if all that is cumin for £11 odd quid i wouldnt say its up to much tbh.

i used mypin pid's on a project last winter, even they are hard to get genuine even tho they are cheap at the best of times.
a propper pid would set you back hundreds on its own.

the pid should be able to hold a 400c temp if you have it set up right, it ill shoot past it initially but will sit right on it not long after.

I don't expect it to last long - it's just a cheap way of getting started.
I'm a bit out of date with computer control, but I have a BBC micro I can use if the pid isn't up to much.
I've also got raspberry pi in the shed somewhere as well I think.

What.sort of voltage, btw? I think the BBC Analogue port gives 256 steps over 5v - Im guessing I'll need to amplify the thermocouple signal for that level? (That gives an accuracy of +- 5deg c up to 1280c)

 

[daleyd]

It's the order of millivolts - 500 deg c is around 20.6 mv. You bbc micro steps might be a tad coarse for this application!!

Those cheap pid controllers should be absolutely fine for your application, I think @brewdexta has used one with good success for his brewery projects.

 

[daleyd]

If you find its no good (I'd be surprised if it wasn't though)I have a couple of spare units from work, but I'm not sure of the output signal so I'd need to check (ours are modular and you can have different output types, eg relay/analog/time proportional etc) - not all work with an ssr.

 

[bigegg]

It's the order of millivolts - 500 deg c is around 20.6 mv. You bbc micro steps might be a tad coarse for this application!!

Those cheap pid controllers should be absolutely fine for your application, I think @brewdexta has used one with good success for his brewery projects.

That's good to know - because I've just been reading up on it, and there are other issues with using the Beeb
Non linear response.
Reference temperatures.
Noise suppression on the amplifier.

 

[brewdexta]

It's the order of millivolts - 500 deg c is around 20.6 mv. You bbc micro steps might be a tad coarse for this application!!

Those cheap pid controllers should be absolutely fine for your application, I think @brewdexta has used one with good success for his brewery projects.

As long as you get the right PID to match your thermocouple you are fine, some can be changed between K type and PT100 etc. some can't.

These guys are the ones to talk to for thermocouples you can't get off fleabay, I got some nice 600mm long PT100 probes for the fermentation vessel from them.

 

[bigegg]

If you find its no good (I'd be surprised if it wasn't though)I have a couple of spare units from work, but I'm not sure of the output signal so I'd need to check (ours are modular and you can have different output types, eg relay/analog/time proportional etc) - not all work with an ssr.

Thanks!

 

[daleyd]

Yes, the mv curves for thermocouples are not generally linear. And as you rightly say as the voltages are so small any amplification would need to be spot on otherwise you'd end up with all sorts of irregularities!!

Don't forget you'll need the correct compensating cable if you need to extend the connection between the probe and the controller.

 

[ronan]

Aren't thermocouples calibrated with freezing water and boiling water ? 0 degrees and 100 degrees ?

 

[daleyd]

Aren't thermocouples calibrated with freezing water and boiling water ? 0 degrees and 100 degrees ?

You can do and it's a coarse check to see if they are reading ok. For anything needing a degree of accuracy though you would use signal injection and/or a proper calibration temperature source, ideally up to or above whatever temperature you will be using it at. Calibration is a job in itself, though I get involved in it occasionally we have a dedicated team that solely does that so I wouldn't claim to be any expert on that side of things!!

 

[norlander]

You can do and it's a coarse check to see if they are reading ok. For anything needing a degree of accuracy though you would use signal injection and/or a proper calibration temperature source, ideally up to or above whatever temperature you will be using it at. Calibration is a job in itself, though I get involved in it occasionally we have a dedicated team that solely does that so I wouldn't claim to be any expert on that side of things!!

Oi Jack, he is building an oven, not hunting or driving roundy things with funny propulsion sneakin about below thermoclines and thingies like that.

 

[daleyd]

Oi Jack, he is building an oven, not hunting or driving roundy things with funny propulsion sneakin about below thermoclines and thingies like that.

Oi wings, get back in yer trench!!

I wasn't suggesting that mr egg should do any of that, just a bit of background in calibration (not that I know much about it like...)

Of course there's ovens and then there's OVENS!!

 

[Kayos]

I use one of the cheap pid's from eBay, as recommended by @brewdexta infact he even sent me a pt100 as I needed the better resolution

I can't speak for the K type but on mine it holds a bath of water and chemicals at 37.8 degrees perfectly, tested against a proper old school reference Mercury thermometer and £1k of calibrated flue gas analyser I use at work. Not bad for £10

 

[tinman123]

I don't expect it to last long - it's just a cheap way of getting started.
I'm a bit out of date with computer control, but I have a BBC micro I can use if the pid isn't up to much.
I've also got raspberry pi in the shed somewhere as well I think.

What.sort of voltage, btw? I think the BBC Analogue port gives 256 steps over 5v - Im guessing I'll need to amplify the thermocouple signal for that level? (That gives an accuracy of +- 5deg c up to 1280c)

your probably getin too technical for me when your talking about steps and volts but i used two here to heat waste oil for a job,
one is set at 80 dgs and the other iirc is set to about 400 dgs so im a long way off your 1200 dgs.
i used e bay thermocouples, one short like yours and one longer one.

i needed some shielded cable and iirc it was cheaper to buy a long tc and cut the wire off it than buy it on its own.
i also used K type plug and socket to aid removal of the unit its on if needs be.
the tc is 220v and i have a transformer supplying me with 12v's iirc for the SSR, of all of the items i used the SSR's are the most fickle, that is unless your willing to part with 50-100 for a good one, i wasnt.

 

[rosinante]

The thermocouple will eventually degrade but for home use should be fine, a pt 100 4 wire will give best resolution but will be more expensive, we used type s thermocouples in the furnaces at work , good up to 1600c, changed 3 monthly for accuracy, switching of the heating element will cause the most wear so best use a Solid state relay , over rate it for reliable use,
hth
chris

 

[wookie]

but I have a BBC micro I can use if the pid isn't up to much.
I've also got raspberry pi in the shed somewhere as well I think.

What.sort of voltage, btw? I think the BBC Analogue port gives 256 steps over 5v - Im guessing I'll need to amplify the thermocouple signal for that level? (That gives an accuracy of +- 5deg c up to 1280c)

Erm... 1.8v rings a bell with me for the BBC analogue port rather than 5v. The analogue port has 4 channels, if you only have 1 input then use channel 0 and switch the others off and it speeds up the conversion, I still play with my old 8 bit Acorn kit from time to time

 

[Parm]

We use K type thermo couples at work to measure gas temps in big industrial chimneys. We have to have them calibrated to comply with our UKAS accreditation (that's another waste of time storey). It's cost more to calibrate than buy. So now we treat them as a consumable item and just buy them in calibrated