Thermocouple variation -defect?

[MartinB]

Put my probes in ice and boiling water to check offsets, since wasnt reading same.

Noticed......thermocouple reading at 212, bounces 215-217. Tried another, bounced 216-217.

This reflects what i see cooking too...the reading is unsteady by 1-2F. The reading is not unsteady in my multimeter.

The thermistor probes are not unsteady either

Thus....seems to be a defect in the heatermeter thermocouple amplifier? Bare wire thermocouples could bounce around rapidly....a probe cannot.

Anyone else see this? Is there a fix? I've observed others online cooks that bounced around by one degree or more also.

As someone who used a lot of lab equipment that stable at 0.1 degree.... This seems kind of poor. It did not fluctuate like that at cold temperature on ice.

Is there some kind of grounding issue?

 

[Bryan Mayland]

HeateMeter is definitely not a lab-grade device, but if the temperature is fluctuating that much then there's no way it would ever be able to control a pit as the fluctuations would confuse the PID controller.

You can see live graphs of the internal ADC measurements by installing the LinkMeter -> AVR Firmware -> Online Repository -> Flash hm-noise.hex. Then go back to the home webui and in a few seconds you'll see a new graph with a cycle's worth of ADC. It is easier to see from there if noise is being coupled into the probe somehow from there. You should see something like this

with the variation extending one or maybe two ticks on each side of a center. It's ok if there's a ton of blips, as long as they're only 1-2 units tall. If you're seeing a jaggedy line going all over then that's noise coming in from the AC power, or being coupled through the braid of the thermocouple. If it matches mains AC frequency, you can cancel that out in the config page by turning on the "A/C input line noise filter" in the config.

 

[MartinB]

HeateMeter is definitely not a lab-grade device, but if the temperature is fluctuating that much then there's no way it would ever be able to control a pit as the fluctuations would confuse the PID controller.

You can see live graphs of the internal ADC measurements by installing the LinkMeter -> AVR Firmware -> Online Repository -> Flash hm-noise.hex. Then go back to the home webui and in a few seconds you'll see a new graph with a cycle's worth of ADC. It is easier to see from there if noise is being coupled into the probe somehow from there. You should see something like this

with the variation extending one or maybe two ticks on each side of a center. It's ok if there's a ton of blips, as long as they're only 1-2 units tall. If you're seeing a jaggedy line going all over then that's noise coming in from the AC power, or being coupled through the braid of the thermocouple. If it matches mains AC frequency, you can cancel that out in the config page by turning on the "A/C input line noise filter" in the config.

Why not smooth this noise in thermocouple?

Ive used ac line filter, no effect

It doesnt really occur at room temp, only at higher temps

 

[RalphTrimble]

RE: grounding issue... I do prefer ISOLATED thermocouples as opposed to grounded ones.
What kind of thermocouple are you using?

 

[Bryan Mayland]

Why not smooth this noise in thermocouple?

It is smoothed / oversampled / filtered but if your noise is bad, even the hundreds of samples that are combined will have enough variation that your number is going to jump up and down by a few degrees. Remember a few degrees is around 5mV peak-to-peak, which is not a lot compared to what can be coupled into a HeaterMeter. Did you try the noise firmware to get a visualization of what you're dealing with?

 

[MartinB]

It is smoothed / oversampled / filtered but if your noise is bad, even the hundreds of samples that are combined will have enough variation that your number is going to jump up and down by a few degrees. Remember a few degrees is around 5mV peak-to-peak, which is not a lot compared to what can be coupled into a HeaterMeter. Did you try the noise firmware to get a visualization of what you're dealing with?

Not yet.

I will try a shorter thermocouple and twisted pair instead of parallel too .

 

[MartinB]

Ok, did as instructed
Below . First is without ac filter. Second is with. At room temp.

Dont look like yours....

Thx

And with it hot.....with the filter.

 

[MartinB]

And a pic on 12 v dc in my truck....not running...no alternator noise. Direct battery voltage

 

[MartinB]

I tried on a 5a regulated fitered 12 v dc supply i have.....same thing.


Then i changed thermocouples to different one with stainless cover over wire. (Auber)

At least in my house when calibrating on boiling water, appears lower noise than other and more stable. Still has the sinusoidal rhythm. Height of variations are 1/2 tic.

Now can i get rid of this add-in wo reflashing firmware?

 

[WBegg]

There's some interference somewhere. Have you tried on an isolated AC outlet? Grabbing at straws at this point.

 

[MartinB]

Well despite whatever seems to be going on with the sine wave, I put the new thermocouple in the oven and it behaved in a Rocksteady fashion. This should make controlling the smoker easier. The Auber thermocouple is fairly small diameter and responsive, way more responsive than the thermistor probe from thermoworks or Maverick. What I have found so far is it trying to control off of those thermistor probes is poorer then a thermocouple because of the added lag time of the higher Mass probe.

 

[MartinB]

And the pit test is successful. since I had some coals left for last night I just lit them again with a propane torch, and let it heat up to check things out. It's an order of magnitude less variation on the thermocouple. Yahoo!

Control is now basically a few tenths of a degree.

I still get some stray spikes of maybe even up to 5 to 20 degrees, but they are momentary and don't effects control. You can see them on the trend however.

So again I'll plug the Auber thermocouple as being a pretty good cheap option. Much better than the one I had at least.

 

[WBegg]

Awesome news! I actually make my own thermocouples. It's an easy deal. Just have to have a way of welding the tips of the two wires together on K-Type wire. I have an old Jewelry spot welder I bought on ebay cheap as chips.

And NO, I'm not selling them.

 

[MartinB]

Awesome news! I actually make my own thermocouples. It's an easy deal. Just have to have a way of welding the tips of the two wires together on K-Type wire. I have an old Jewelry spot welder I bought on ebay cheap as chips.

And NO, I'm not selling them.

You can see the occasional split-second spikes in this screenshot

Contrast this with the control of last night's ribs

 

[WBegg]

That's still some frequent spikes. I get those maybe 1 an hour? My solution is to hit the refresh button and they go away. It's the same thing i do with my truck. When it makes a noise I can't figure out, I just turn my stereo up louder. I have a bigger stereo now.