Cook in progress - trouble with probe accuracy - Help me Obi Wan, You're my only hope

Hey all,

Got a problem with a cook in progress. Tried to provide as much detail as possible, hope its not too much to read so early in the morning!

HM Version: 20141003B
All hardware purchased from HM store with the pre-soldered TC module

Setup:
Probe 0 is a TC ordered from Auber - "K Type Thermocouple w/ alligator clip tip. Smoker (TC-K6A)" - this is clipped to the V-rack my pork butt is sitting in
Probe 3 is a PR-005 high-heat probe ordered from maverick housewares stuck into the pig
Probe 2 was added around 6am and is the same PR-005. It is currently aluminium foiled to the v-rack
Maverick wireless - my old wireless unit that I am using for comparison and a backup alarm since this was my first HM cook and I have a $55 pork butt on the smoker

Problem Description
I have a 10.5 lb pork butt on the BGE since about 4:45pm yesterday. My TC probe was reading about 7 degrees higher than my old maverick wireless unit which wasn't a big deal to me. Everything worked GREAT until about 4:45am (almost exactly 12 hours). At about 4:45am, the Thermocouple for some reason thought that the temp was climbing and sent me a text message alert that my grill temp was 235 degrees. I got out of bed, checked the smoker and both the maverick remote and the dome gauge said it was 205 degrees. I went back to bed. The HM continued to believe that the temp was rising and kept sending alerts, but the dome and maverick sensors said she was holding steady at 200-205. At 6am, I *gasp* opened the lid and installed a 2nd probe which I aluminum foiled to the grate and plugged into the heatermeter as probe 2 just to have a comparison to see if the TC probe was the issue.


Questions:
So, its now 8am and the TC probe thinks the grill is at 227, so its shutting off all airflow to bring the temp down - unfortunately, the actual grill temp (measured by the maverick AND the dome) is 190. Immediate question is; should I put in a 37 degree offset on the TC probe? Move my set temp up 37 degrees? something else? I still have some hours to go on this cook and need to get the temp stabilized. Also, any suggestions on what could cause the probe/HM to go haywire after 12 hours? Probe 2 is only off 15 degrees, reading at 205, but that's still pretty bad.

Any help would be most appreciated.

-Brian


Here is a graph of the cook so far.

Update:

So in the 10 min it took me to post this, the TC temp has dropped to 210. Its still over by about 20 degrees vs the maverick and dome, but now its "more accurate". Im wondering if I have a setting wrong or something. Here is a screenshot of my

Go to the AVR Firmware page and load the "hm-noise" firmware from the online repository and let's see what the input is looking like. Post the noise graph from the home page once it pops up and maybe there will be some information there we can use to help. The settings look fine.

Dome temperature is not accurate. It could be as much as 50°f off. If you are not sure about the accuracy of the thermocouple, then put it in boiling water and see if it holds between 210-212.

There should never be an offset on the thermocouple, unless you are using an adapter board.

Thanks for the reply Bryan. I have attached the screenshot below. One other note, the probe 2 just up and stopped working. Also, the temp difference from my maverick gets worse the higher the temp goes. At 220 deg on the maverick, the HM shows the temp at 267

@John, ya I should have made sure they were calibrated before I started - noob mistake. That being said, I know that my maverick is dead on as I calibrated it a few days ago when I was troubleshooting, and my dome temp is pretty consistently within a few degrees of the maverick on long cooks once the dome is fully up to temp and as long as I don't open the lid.


Here is a video of the input with the hm-noise firmware. It jumps around a lot.

http://screencast.com/t/Ur0znJpPFVk

Here are some screenshots in case you don't feel like watching a video.

That is really bizarre. I thought for sure we'd see the graph jumping all over the place like there's bad noise on the power, but that's exactly what you should be seeing-- the readings jumping between a number and that number plus one. If you're getting interference on the probe, or if the thermocouple amp isn't working properly, you'll see the values just constantly skewing from one value to another and those values being 100 numbers different on the noise graph. I am at a loss why you'd see the value going all over like that because the data indicates no problem with the HeaterMeter hardware. Although the green probe is also tracking along with it so if something's wrong it is with the whole analog system. Iiiiii am at a complete loss.

On the graph I see most of the cook running without the blower, and the blower ramping up and blowing like hell right before the pit temps shoots up... Those two things generally do follow each other, and the other probe seems to track with the TC (albeit slightly offset) so my guess would be the HM is showing you exactly what happened in the grill. Why IDK. Hard to really tell not being there to actually check things out though....

Could there be any ill effect from the HM getting really hot? last night I had the HM in a ziplock bag, which I took off this morning. Shortly after, the temp came down. Then, the sun came out. It is 102 in Texas here today and my HM was sitting out in the sun. I went out to check on it and it was so hot I couldn't pick it up. I took it out of the case and put a cardboard box on its side to create some shade. Some time after that I noticed that probe 2 had stopped working al

Is it possible that the extreme ambient temp is causing the electronics to freak out?

hey Ralph, I assume that behavior is related to the setting above where I have the fan only coming on when the servo is at 100%.

I don't disagree that its accurately reflecting what is happening, but the problem is that its happening because of false readings on the probes. The probe is reading 7-37+ degrees off from my maverick wireless redi-check, so while it thinks its adjusting to the set temp, its actually not. Its also not off a static amount, the disparity between the HM temp display and my known good maverick gets more disparate the higher the temp goes. In other words, at room temp, its off by about 7 degrees, at 200 degrees, its off by about 17 degrees. At 225 its off by almost 40 degrees. But even that correlation isn't consistent.

If I were looking at a graph where the HM% was consistent and suddenly the pit probe started reading high it would point to a probe anomaly, but the ramping up of the HM% right before the rise in temp seems to indicate the pit temp should be rising? Then again, the fire could have been at the point where it needed the boost of air flow to ramain steady, and the false rise in the pit probe reading made it drop off... Like I said, it's hard to read a graph and really tell what is going on with a fire...

At any rate, you may be on to something with the ambient temp analysis.... The TC amp is sensitive to ambient temp, although if the amp and the TC junction (TC jack) are held at the same temperature ambient temp should not effect its readings. I guess you can test that by letting the HM sit in the sun and then in the shade (while the probes are in a stable temp, like perhaps a cup of water) and see if the temp readings sway with the HM ambient temp.

I would also suggest that you reflow the solder on the components that are behind the probe jacks, the two resistors that are standing up and the caps that are with them, as well as the four pins on the ATMega where the probes connect (pins 25-28). I fixed a couple HM boards recently that had probes reading radically different, after soldering these components in properly all the probes were reading exactly the same. Heat inside the HM could effect the resistance of a bad solder joint and thus effect the readings on the probes. Although the SMD TC amp parts came pre-soldered to the board you had to install the resistor and cap for the TC mentioned above.

The ambient temp on my internal thermistor peaked at 130F today with no ill effects while the sun was beating down on it. Outside temp right now is 33C ( 91F ).

Yah, the ambient temp shouldn't be an issue.... I'm starting to think there may be some poor solder joints on the board, ambient temp will effect the resistance of poor solder joints....

Thanks Ralph, for all of the responses. Poor solder joints is a definite possibility...this was the first time I have done any real soldering. I am really Type-A so I thought I checked and double checked but its still very likely that I messed something up. As soon as this cook is over I will do as you suggested and re-solder those joints.

Where can I find the ambient temperature? I don't see that in the UI anywhere.

There is an optional ambient temp thermistor that you can solder to your board, it's a little blue Vishay unit that many people think is a capacitor... It would solder from the switched leg to ground on the last probe jack typically, then you set that probe type to Vishay in the HM config.

You can triple check a solder joint with your eyes, it may look good on top but not be connecting well below. The HM is a two sided board, there are traces on both sides, so solder needs to flow through the board to make connections on both sides. So just go real quick and reflow solder on the parts between the probe jacks and the ATMega (leave the SMD stuff alone since it came pre-installed), melt the joint and add a little solder, wiggle the leg of the component a little with the soldering iron, this helps solder flow down the component leg... try to be as quick as possible on each component though. Of coarse you should unplug the power, remove the rPi, and I always suggest you remove the ATMega from its socket before you reflow solder on it. (I make that statement because I have seen people make each of those mistakes and cause themselves more troubles)

I am embarrassed to even post this photo. But alas, maybe it will help someone else. Gonna resolder now and see if that helps.

Ralph was correct, I re-soldered basically the entire board per his instructions and just did a boil test. The TC only shows 208, but both probes showed 212 on the nose.

Most importantly, the pork butt was amazing

Thanks for all of the help and advice everyone!

208° on the TC seems to be off a bit. Where did you buy the thermocouple. My TC that I have been trying my hardest to destroy, still reads 211 in boiling water.

Isn't that what the offset is for, to adjust for slight variations in probes? Mine reads 208* as well at boil so I just have a 4* offset.

There should really be no need to adjust any offset, especially when using probes from the dropdown list or a thermocouple.

The nature of a thermocouple is 2 known dissimilar metals that output a voltage. If you're 4 degrees off at boiling, then something's amiss somewhere else. Adjusting the offset is not the fix in this case.