Want thermocouple support? Wait!


 
Bryan Mayland

Bryan Mayland

TVWBB Hall of Fame
Despite the thermocouples showing good performance sitting on my workbench and hooked up to the oscilloscope, in practice they're really working terribly right now. In addition to the addon board being difficult to use, the values are just all over the place when on the grill.


They have these strange spikes where the temperature will shoot up or down 30 degrees, hang there for a couple minutes, then return to normal. I've even added a giant filter to the output and the temperature can swing up to a degree for a tick. I've tried bunch of probes, from bare thermocouple to isolated and grounded. The bare was the worst because I think it picks up small changes in heat that are normally dampened by a slow-reacting probe.

So if you're gung ho on a HeaterMeter with thermocouple support and are about to buy parts, I advise you to wait. The HeaterMeter PCB itself will probably require changes, and it might not work very well at all in the end.
 
Bryan Mayland said:
So if you're gung ho on a HeaterMeter with thermocouple support and are about to buy parts, I advise you to wait. The HeaterMeter PCB itself will probably require changes, and it might not work very well at all in the end.
Click to expand...

That's a bummer....
Do you mean the HeaterMeter V4.1 PCB itself is likely to change a bit before it is finalized? Or just the thermocouple circuit?
 
Yes, I mean the HeaterMeter PCB itself may very likely require an update. This is why


The temperature was seldom right, fluctuated all over, and finally just dropped off. This may be related to the connection to the thermocouple itself. It may be possible to salvage an existing 4.1 board by using a completely redesigned thermocouple board and then soldering the wires to it but at this point I can't say anything for sure.
 
We appreciate the update. Good luck getting the gremlins whacked!

For those of use sitting on v4.1 parts (but no PCB) and waiting to build a HeaterMeter to use with regular Maverick probes, would you recommend just sitting tight too? I also want to get a 3D printed box from one of the guys selling them here on the forums, so if the PCB changes, it may affect that too.

Decisions, decisions!
 
I had a similar problem with my initial AD8495 TC board test. Just for yuks, I grounded the braid and everything fell into place. I already had a 1M ohm resistor between TC- and GND on the board, but ok. Now that I have it mounted, it doesn't seem to matter if the braid is grounded or not. I've smoked a turkey, ABT's and burgers, and just kept two other pit probes on, and there was about a 20F variance at 350, about 5F at 250, but that could be placement and the fact that I calculated my own coefficients for my probes. The TC was running the show, and everything came out as expected when expected, and was pretty good. I guess what I really need is to use something properly calibrated for a third opinion;) Ice water and boil test were good, but those other variances make me hesitant of posting it.

I was looking at Thermocouple-Board-Success-(sort-of), did you ever solder a TC to a mono plug, and the jack onto the board? That seamed to help the stability for me. I just glanced at GITHub and didn't see an eagle file for your AD8495 board in the eagle folder, was i looking in the wrong place?
 
Bryan,

I saw on the development thread that you have been looking at TC terminals.

I'd also like to remind everyone that if they use a material that is not the same as TC's they are using, they can drastically throw off their measurements. For example, if the ambiant temp outside is 60 degrees and the temp inside the enclosure is 90 degrees, the TC measurement will be off 30 degrees.

You might also want to consider RMJ-K-R.

They use a 5/8" hole and mount to the enclosure. If one were to mount this jack and use a short piece of TC wire between it and the PCB it would allow them to plug the TC into the enclosure.
 
I'm not a fan of panel mount jacks because I take apart and put together devices so frequently that having wires run to the case is cumbersome to work with. Would that work any better inside a case though, because your cold junction is still going to be inside? I've picked up some of the PCB mount from Omega for the next attempt.
 
What you linked to will work fine as long as there are no other materials between the TC tip and terminal (ie copper).
 
I spoke with the guru of TC's here at work and he had a couple of suggestions. He thinks a good portion of the variation appears to be poor junction compensation and poor isothermal connections. The TC amplifier add on board is also not a great design from an isothermal standpoint in his opinion.

Here's an HP/Agilent Application Note that will describe every detail you ever wanted to know about TC, thermisters, RTD's, etc.

http://cp.literature.agilent.com/litweb/pdf/5965-7822E.pdf
 
I'm not sure there's much one can do if there's heat generated inside the case and the amplifier is located inside, is there? I was unaware how thermocouples worked with regards to their connectors too so the first TC board, while it sort of worked on the bench, was just not well designed for the final application. I've got some of these out at the fab right now and I'm hoping they'll work better. You'll see a proper thermocouple connectors, integrated filtering for output common mode and differential rejection, separate analog ground lines (although power still goes straight through the middle of the board).

MP1jyfy.png

l3x6MGJ.png


It's probably not ideal but I'm hoping it will be ideal enough for control purposes.
 
The temperature compensation is needed at the junction where the metal changes to copper on the board, this is called cold junction compensation. The chip AD8495 is located right next to the + / - terminals on the pcb with .250 inches or so, The temperature of the terminals and the chip will be the same even if the HeaterMeter generates some heat inside the cas
 
You must log in or register to reply here.

 

Back
Top