Problem with latest Firmware Snapshot


 
R

RalphTrimble

TVWBB Diamond Member
I fired up the pit tonight for my first cook with the latest firmware snapshot installed and have run into a problem. I am getting "Probe Noise" on the pit probe, which I thought I liked at first until I realized the probes were already warmed up enough to register... If I went to manual fan mode the pit probe seemed to register a steady 75 degrees even though I know it was hotter than that and getting hotter, if I go back to auto mode I get "Probe Noise" again.

I rolled back to the release firmware and all the probes read fine again, then flashed the snapshot and same deal. I pulled the probes from the damper end and plugged them directly into the HM and they did register then, back out on the CAT5 cable and I've got "Probe Noise" again. I also noticed that I was getting "Probe Noise" even when I had no probes plugged in anywhere, just the CAT5 running out to the damper with the fan and servo connected.

I'm not sure if this has to do with the changes in the servo motion or the "Probe Noise" feature, but whatever is going on in the snapshot firmware it is making the probes pick up more interference over the CAT5 wire it seems, or making it more sensitive to it somehow....
 
Thanks for reporting this. I think I might just make the noise threshold a configurable value. I'd rather have the UI just show a noise indicator but there's a ton of code that would have to be changed for that.

The new code just reports when the noise gets bad, the old code would just report the bad value. You've always had the noise, but the old firmware wouldn't check it so the only time you'd run into a problem is when it came close to the top of the ADC range where the probes would drop off way below where they should (such as the ET-732 probes not working at room temperature even though it was inside their range). The new design is technically more correct, but less convenient if the probes or power is noisy.
 
OK, well, I got plenty of noise here over this long CAT5 cable, so if you need code tested I'm your guy! LOL For now I rolled back to the release version so I can cook in the meantime...
So you are pretty sure that it's the change in the "noise" code that caused this issue, not the change in the way the servo moves? What about the blower drive, you had talked about the two modes it can run in now, did the blower drive method change in this firmware version? I didn't notice any option to select which mode to run the blower on the Config page....
 
Oh I am absolutely sure it is just knowing about the noise it didn't know about before. Before it would just take instantaneous samples 4 times a second and average them together so it couldn't really tell much about the shape of the data in aggregate. Now it samples 256 samples ~0.1ms apart, across 6 inputs, 6 times a second so it can take at least a cursory look over the data variations in those windows. When there is more than a ~3% variation in the input data, that's declared "noisy" because that's 9.29F at 225F. Looking at the raw data it is rarely completely symmetrical so HeaterMeter can end up chasing a phantom temperature.

I'm probably adding back the "ignorance is bliss" mode because it certainly does feel dickish to stubbornly refuse to report the temperature because I (me personally) feel it's not a good value. Plenty of decisions are made with bad data!

The blower Voltage drive mode is available in firmware but it defaults to the Pulse mode. You need a v4.2 board to do Voltage mode, and the UI is in the linkmeter package so you wouldn't see the option. If you install that then put the latest snapshot firmware on you can see your noise level (but there's no way to turn it off).
 
OK, I'm glad to hear it is the changes in noise detection rather than the change in servo movement that is the source of this issue, cause I really like(d) the way the servo was moving with the new 50 step method.
 
You must log in or register to reply here.

 

Back
Top