High temp thermometer.

the

u are

Hi Jamie, I have to agree with Spehro. Something is 'not right' in the instrument. It sounds like the device that is (supposedly) measuring the junction temperature is not at the temperature of the junction. Maybe a bad design? But 10 degrees seems like an lot of error! And perhaps it is mis-wiring of the box.

I assume you've got a front panel with the proper thermal couple (TC) connector and plug. (Both of the same TC material.) (I use a lot of type T TC's, copper and constantan, because it's the least magnetic of the common TC's) If your box is like the ones I've built then you've got a section of wire running from the input connector to the device that is measureing the TC voltage and junction temperature. These wires also must be of the same TC material! (Are they?) Are they wired correctly? Hey, maybe you can be a hero and fix this problem!

George H.

Sounds to me like the controllers have poor cold-junction compensation, which is a too-common problem. I've seen thermocouple gear that had not-very-isothermal tc terminations with cheap semiconductor or even diode ref junction sensors. They are usually calibrated to be right at 20 C or something like that, but the slope is poorly controlled and the ref junction correction goes to hell at other ambients. One degree error in 10, or even 5, happens.

In our stuff, we use thinfilm platinum RTDs for ref junction sensing, ratiometric ohms sensing against a good resistor, and we use the proper math to convert the resistance to temperature, and that temperature to the TC voltage correction. It isn't really hard, but most TC circuits and appnotes still use dumb analog diode things.

(the platinum RTD is on the bottom of the pcb)

Connecting the extension wires backwards, or using the wrong kind, would be interesting. He could get some freeze spray and spritz all the junctions in the cabling, and see what changes.

The compensation circuit needs to measure the temperature at a single specific place: the connector used to terminate the thermocouple (or extension) wires. So it is not a matter of "seeing the ambient environment" or keeping the cabinet the same temperature as the environment. Although this may help of course. There is no reason why a PCB mounted sensor located at the thermocouple connector cannot be within a degree or so of the true value.

Backwards will give an error of roughly double the temperature difference. J-K errors are a bit more subtle- about 20% of the temperature difference shows up as an error.

It's complicated by the balkanized color codes used worldwide for T/Cs. If you have an Italian machine bought in the US with Japanese controls just about anything might be in there.

Helps to have a magnet handy, except for T, S, R, B .

Yeah, he could have the extension wires wrong. The color codes *are* awful.

You know what, I've work with this shit long enough to know the difference on wire connections. You guys act like you never leave the controlled environment. Red = (-), white (+) for J and yellow (+) For K etc...

It would be nice if you had a work place that could maintain temperatures inside and out at a even level, but since you obviously don't work in harsh environments, you wouldn't know.

I'll give you a classic example, we have these unit located in various places to test polyrad after it has been irradiated. This is to insure it got irradiated and the process didn't fail. The unit has a fire rode heater mounted in a 4x4 basic electrical box that has a heat shield cover the operator can lift, this box also is mounted to a NEMA 4 enclosure that houses the heat controller. after some operating time, the side of the enclosure gets hot and it heats up the inside and the process controller gets hotter than the ambient temperature where the J couple cable is leading, which is a J-Jack on the outside.. a 10+ Bias must be applied to all these units and the units have to be on for a while before they are accurate.

We have units that use various venders of controllers to perform this job from all parts of the world. They all seem to behave the same.

When these units are cold and the firerod reaches temp, they read correctly, if you have the offset set to 0. when the enclosure gets hot after an hour of so of just sitting there, which is what they do, all day, all week etc.. the offset needs to be applied. If we open the door to the enclosure and let it vent to get balanced, it then does not need the offset.

The product developing lab has the same issues with controllers that deionize water. Some of us call them lobster pots, but the controller sits just below the tank of water and it gets up to 95C, an offset on those must be applied, too. If you attached a heat controller and let it sit on the table where It does not get any hotter than the surrounding environment, then it is with in the spec with 0 offset.

This is all basic stuff to me, I don't understand why it is such a hard concept for others.

jamie

Thanks kindly. In other posts, I gather, the subject is your specialty.

That would be a kind of bench mark (hiccup). Sold as a "potato purifier" that eliminates Ethyl at a quart per day might be marketable.

Cheers Ken

That's the American convention. There are different colors for europe and, I think, Japan.

Obviously something is wrong in your system if you have to recal just because the cabinet temp changes.

As far as getting out goes, I have laid on asbestos matting on the top of a live steam valve on a 32,000 horsepower turbine in the middle of the Gulf of Mexico, in mid-summer, trying to turn a trimpot. I once enjoyed a 2-day cruise on a container ship down the West coast, finally locating a loose screw on a terminal block. I don't think Spehro or I lack experience getting dirty in the field.

Don't get obnoxious without knowing what you're talking about. It can make you look silly.

One problem is that temperature is an ill defined quantity/quality, much more complex than is usually understood, likely an obsolete term in General Relativity and Quantum Mechanics, having no general consensus. The term 'temperature' applies to a bulk of material. Regards Ken

You know, maybe JT is correct about you. I really thought you were better than that. He may have been taking an off sided approach to it but I can see he made some good points about you.

You do not understand what you think you do about the subject. And maybe other things you could have been sucking off the group over the years.

My how things come out in the strangest times, you should have kept quiet.

One day, you'll find yourself in this same situation, and look back however, by the time you realize that I could of been correct, and I am in this case, it won't make any difference.

To put it bluntly, you are mistaken this time. You have been a few other times also. I just did the courteous thing of sitting back and watching.

Have a glorious time sucking information from others. It seems from what I've seen here lately, you are good at that. Put out danglers and have others refine it for you. what a way to get your work done for you for free. In my eyes, that isn't much different than the free loaders we have in our society.

If that is what you expect from others? "Feed the pig" so be it. I feel sorry for some here that have been nice enough to offer you suggestions that actually helped you. Unfortunately, they too, will be one day, approach by the likes of you in the same manner. I can only hope they learn or have already learned their lesson.

There is nothing like actually experiencing life and duties over just sitting there and theorizing them in your fantasy world.. Does that ring a bell for you? You really should get out more in the field, on site, and actually see what goes on out there. It's blurring your vision, but It does not matter because you've grown accustom to picking fruit off of what ever tree is in blossom when you need it.

Jamie

On a sunny day (Thu, 31 May 2012 15:36:26 -0400) it happened Jamie wrote in :

Hey, I think you could be wrong. Not that I am a thermocouple genius, but I did study the subject when I build by little thermocouple sensor thing, and wrote the soft. Actually John was a great help suggesting the circuit, and pointing out stupid errors in my first software release.

If you keep the cold-end sensing where it should be, *at the connector where the thermocouple wires come in*, with a sufficient low thermal constant, there should be zero errors. Maybe your wires touch some metal some where, or some other problem. Nobody knows everything, and I forget the details of many things I did many years ago. Wrote some nice software today and got my temperature, pressure, altitude and magnetic bearing module from ebay working. Anybody wants the C source? Hey it ended up multi-threaded...

Have a cookie, a good night sleep, and maybe tomorrow fix the problem with your temperature sensors.

ious

this

at

Well that's mostly true for the TC stuff I've done. Most of the time the controller is near room temp. But next time I turn on one I'll blow a little heat on the controller and see if there's any change.

George H.

Or you could just throw it in a bucket of ice water to stabilize it. ;)

Cheers

Phil Hobbs

I try to be civil and helpful to anyone who is interested in electronics, at any level. But if you elect to go personal and insult people who are trying to help, which is exactly what you did, don't expect them to keep being polite.

How helpful is JT being? I bet he knows exactly what your problem is, but he's decided to keep it secret. Lately he does that a lot, tells us how smart he is, but keeps the proof to himself.

I designed this

and this

and this

and three thermocouple-based NMR temperature controllers, so far. I know the basics about thermocouples. Spehro and I were trying to help, and you reply by telling us that we don't know what we're talking about.

The next time you have a problem, I sure will.

If it does, the cold junction compensation is junk. Which it often is.

Well maybe not quite zero but certainly very small, depending on the accuracy of your CJC compensation. A $200 industrial thermocouple input module will do this quite well, but a $50 Watlow temperature control board with thermocouple input will not do a very good job at all, having a poor CJC circuit located far from the thermocouple connection on the same terminal block as the power and heater terminals, and management might not see the value in switching. I am sure there are lots of other cheap controllers on the market, suitable for +/- 5 C or so, and if you stick to the $50 category, that is probably all you get.

People have been doing accurate thermocouple measurements in harsh industrial environments like steel mills for about a century now. I find it curious that it is still botched so often, even by Omega, one of the largest manufacturers of thermocouples (some of their cheap TC meters and controllers are quite poor).

or K.

Pretentious nonsense.

"Temperature" is an exquisitely well-defined concept in thermodynamics and quantum mechanics.

This is right to the limited extent that temperature is a property of a population of objects, but "bulk of a material" doesn't really capture the concept.

-- Bill Sloman, Nijmegen

Yeah, but I don't think the controller will like the iced water :)

Jamie