Heat treat oven calibration

[PhilipTreeWalker]

Hi all,

I have built a heat treat oven. It has a PID controller connected to a k type thermocouple. The thing is I don't know if it is accurate. Can anyone recommend a method of testing. Obviously a thermometer would be the easiest way so I have been looking for a high temperature one. The ones I have found are k type but I need to put the probe inside the oven with the door closed and I am concerned that, whilst the probe can work at 850 C (up to 1200 C), perhaps the connecting cable cannot?

Anyone have any experience with this?

Cheers.

 

[Oliver G]

You can get ones that plug into a multi meter, I've got a cheap one from maplins that you can plug a K-type in, RS do a suitable probe if you've got a small hole to stick it into:

https://uk.rs-online.com/web/p/temperature-probes/3428928/

Failing that there are loads of calibration companies that can pop over and carry out a calibration for you, bit costly though.

Have you got pictures of said oven?

 

[PhilipTreeWalker]

You can get ones that plug into a multi meter, I've got a cheap one from maplins that you can plug a K-type in, RS do a suitable probe if you've got a small hole to stick it into:

https://uk.rs-online.com/web/p/temperature-probes/3428928/

Failing that there are loads of calibration companies that can pop over and carry out a calibration for you, bit costly though.

Have you got pictures of said oven?

I could potentially drill a small hole for a probe. This is the only pic I have on me, not much to see really. Just imagine that all the welding is beautiful

 

[Oliver G]

Ahh, looks simpler than I thought, are you able to unplug the thermocouple? if not you may be able to splice in a plug and socket to allow you to connect a second device to verify. As you can't change the output of the thermocouple to calibrate it you will have to either change the input range / off set of the oven or just write down the deviation at certain points and use a ready reckoner.

https://uk.rs-online.com/web/p/temperature-sensor-accessories/4559764/ x 1 to the probe
https://uk.rs-online.com/web/p/temperature-sensor-accessories/4559758/ x 1 to the oven and x 1 to the secondary device

 

[pieinthesky]

Plugging a sensor into a meter is no different from plugging your thermo-couple into the PID. You wouldn't know which is more accurate, also you will get different readings in different places in the kiln.

Best thing is to take HT recipies as a guide, do a load of experimentation in your kiln and learn what temps on your PID get the desired result.

 

[PhilipTreeWalker]

Plugging a sensor into a meter is no different from plugging your thermo-couple into the PID. You wouldn't know which is more accurate, also you will get different readings in different places in the kiln.

Best thing is to take HT recipies as a guide, do a load of experimentation in your kiln and learn what temps on your PID get the desired result.

That is essentially what I have been doing so far. The whole point of building the oven was for accuracy but it is harder to obtain than I thought it would be.

 

[Oliver G]

If you want accuracy you'll want to do a heat map of the oven with a multi channel thermocouple meter.

What accuracy are you looking for? It may be worth doing some googleing to see if anyone else has done this before and what results they got.
If their oven is of a similar construction use their accuracy with some fudge factor and you'll have something to be working with.
With the mk1 eyeball and a temp/colour chart you'll get an accuracy of about 10%, based on the engineers black book I have knocking about.

 

[PhilipTreeWalker]

If you want accuracy you'll want to do a heat map of the oven with a multi channel thermocouple meter.

What accuracy are you looking for? It may be worth doing some googleing to see if anyone else has done this before and what results they got.
If their oven is of a similar construction use their accuracy with some fudge factor and you'll have something to be working with.
With the mk1 eyeball and a temp/colour chart you'll get an accuracy of about 10%, based on the engineers black book I have knocking about.

I am very knew to metal, so I was hoping to take as much guess work out of it as possible. I will have to stick to 'no longer magnetic, and a bit hotter than that'. The chamber of the oven is relatively small and well insulated so I wouldn't expect too much variation in the space, especially in the horizontal plane. Colour has appeared even throughout blades I have heated so far.

 

[Broch]

It all depends on how accurate you want the calibration to be.

A k-type isn't linear over its entire range; there's a change in gradient at around 450/500 deg C (see below) so any signal conditioning needs to take that into account.

The thermocouple itself is very accurate but you need to calibrate the whole signal conditioning circuit. If that's a digital system it will include an amplifier, analogue to digital converter and software. The only real way to do that properly is to use a calibrated milli-voltage source at the input and adjust the software offsets and gains until the 'displayed' temperature corresponds to the correct temperature for that voltage input. The interface chip or the software must take into account the 'kinks' in the calibration. If you can't make/find a calibrated voltage source you will need to use a calibrated DVM.

However, having said all that, for the accuracy you need, any good quality DVM that's not been abused will probably give you a good enough calibration.

If it's an analogue system you may have calibrating 'pots'.

 

[pieinthesky]

Make sure your PID is set up for a k type thermocouple. If you have not done this your readings could be a long way out.

I did a lot of re-search a few years back when trying to calibrate my home made kiln and never found a workable solution. If you do find one I will be all ears. I did read about one guy who used wax cones which are calibrated to melt at certain temperatures. They are made for use in firing ceramics - when they melt, they trigger a switch to turn the kiln off, maybe worth a try?

I am pretty sure that the temp read out on your PID will be reasonably accurate. If you had bought a kiln it would also have a thermocouple plugged into a PID, not sure if the manufacturer does any further calibration or not - anyone know?

They are not cheap but a Rockwell tester (which is easy to calibrate yourself) will give you feed back on how your HT is doing and will (in a round about way) help calibrate your kiln. Better to spend money on one of these than to throw too much at calibrating temperature.

Welcome to the world of knife making - where nothing is straight forward!

 

[Imagedude]

I use Tempilstik temperature indicating crayons to gauge the accuracy of my kilns...

 

[Broch]

Actually, I should have said, the way I checked my kiln at near 800 was to put a piece of sterling silver in (925 silver). I set the temperature to hold at 750C, when it got to temp I checked to see if the silver had melted, ramped the temp by 10C and repeated. In my kiln the silver hadn't melted at 890C but had at 900C (it's supposed to melt at 893C). On that basis I decided my kiln was accurate enough around the 800C mark and was happy to heat treat between that and 1000C.

Just don't raid your partners jewellery box to do it

 

[PhilipTreeWalker]

It all depends on how accurate you want the calibration to be.

A k-type isn't linear over its entire range; there's a change in gradient at around 450/500 deg C (see below) so any signal conditioning needs to take that into account.

The thermocouple itself is very accurate but you need to calibrate the whole signal conditioning circuit. If that's a digital system it will include an amplifier, analogue to digital converter and software. The only real way to do that properly is to use a calibrated milli-voltage source at the input and adjust the software offsets and gains until the 'displayed' temperature corresponds to the correct temperature for that voltage input. The interface chip or the software must take into account the 'kinks' in the calibration. If you can't make/find a calibrated voltage source you will need to use a calibrated DVM.

However, having said all that, for the accuracy you need, any good quality DVM that's not been abused will probably give you a good enough calibration.

If it's an analogue system you may have calibrating 'pots'.


View attachment 56900

Wow. You know I usually just play with wood, right? I'll send this to my father, he will understand lol.

Make sure your PID is set up for a k type thermocouple. If you have not done this your readings could be a long way out.

I did a lot of re-search a few years back when trying to calibrate my home made kiln and never found a workable solution. If you do find one I will be all ears. I did read about one guy who used wax cones which are calibrated to melt at certain temperatures. They are made for use in firing ceramics - when they melt, they trigger a switch to turn the kiln off, maybe worth a try?

I am pretty sure that the temp read out on your PID will be reasonably accurate. If you had bought a kiln it would also have a thermocouple plugged into a PID, not sure if the manufacturer does any further calibration or not - anyone know?

They are not cheap but a Rockwell tester (which is easy to calibrate yourself) will give you feed back on how your HT is doing and will (in a round about way) help calibrate your kiln. Better to spend money on one of these than to throw too much at calibrating temperature.

Welcome to the world of knife making - where nothing is straight forward!

I am pretty sure it is set for K type, but will double check. Rockwell tester is on the list, but it's a long list and I am already in trouble for worldwide exotic wood purchases (she doesn't know the half of it )

I use Tempilstik temperature indicating crayons to gauge the accuracy of my kilns...

Super simple, just ordered one at 816 C. I had never heard of them so thanks for that.

Actually, I should have said, the way I checked my kiln at near 800 was to put a piece of sterling silver in (925 silver). I set the temperature to hold at 750C, when it got to temp I checked to see if the silver had melted, ramped the temp by 10C and repeated. In my kiln the silver hadn't melted at 890C but had at 900C (it's supposed to melt at 893C). On that basis I decided my kiln was accurate enough around the 800C mark and was happy to heat treat between that and 1000C.

Just don't raid your partners jewellery box to do it

That's a good idea, similar to the Tempilstick crayons mentioned above. I think the mrs still has some giant silver hoop earings from back in the day...