Public Beta 4.0 Beta 2 Issues

Things to come.
crane
Posts: 24
Joined: Sun Apr 07, 2013 9:42 am
Bot?: No

Re: Public Beta 4.0 Beta 2 Issues

Post by crane »

When I first built my panel I noticed that when I would move some switches on the front panel that were located near the displays that it would sometimes cause them the reboot(the I2C address would pop up like it does when I first power on the panel). I separated the low and high voltage wires and haven't seen the problem since. I don't think this has anything to do with EMI from AC switching as I have never seen this problem on the 3.7 FW, nor did it occur on 4.0 beta 1. What I am seeing is different from the EMI issue I saw when I first built the panel in that the display still outputs a temperature and setpoint, but it gets stuck with the same value and doesn't keep updating as the temp changes until it switches states.

I am currently in the process of brewing a large amount of beer for my friends wedding coming up and I reverted back to 3.7 FW in the mean time and won't have time to test this out further until after the wedding, so I'm not sure if it is 100% repeatable each time it enters the states where I observed it.
gthielen
Posts: 7
Joined: Sun Mar 08, 2015 2:19 pm
Bot?: No
Location: San Clemente, CA

Re: Public Beta 4.0 Beta 2 Issues

Post by gthielen »

Three problems with Beta 2:

1) Process step exit conditions defined on temperature probe setting exited at apparently incorrect temperature readings. The temperature does not appear to have met the criteria for exit, but the process was on the target step. I haven't been able to catch it when it happens, I just see the results occasionally.
Process step exit condition definitions.
Process step exit condition definitions.
Screen Shot 2015-03-30 at 6.30.09 AM.png (29 KiB) Viewed 7772 times
2) Count-down timer does not stop at 0:00:00 and goes negative. Not sure if this is intended or is part of the cause of the next problem.

3) Step Exit conditioned on the above countdown timer <= 0:00:00 won't trigger if step is entered and continued timer is already < 0:00:00. On 3.7, the countdown timer stopped at 0:00:00 and the timer-based exit was triggered if the step was entered and the timer was already 0:00:00
Step 1 Timer 0 definition
Step 1 Timer 0 definition
Screen Shot 2015-03-30 at 6.34.40 AM.png (16.12 KiB) Viewed 7772 times
Heat Delay (step 2) Timer 0 definition
Heat Delay (step 2) Timer 0 definition
Screen Shot 2015-03-30 at 6.39.07 AM.png (15.88 KiB) Viewed 7772 times
Let me know what other details of these issues you need.

Thanks.
JonW
Site Admin
Posts: 1726
Joined: Sun Jul 18, 2010 7:51 am
Bot?: No
Location: Huntington Beach, CA
Contact:

Re: Public Beta 4.0 Beta 2 Issues

Post by JonW »

1 - With such a narrow expected range for your exit conditions, I would bet a small fluctuation is triggering the exit. You should datalog it and see where your temps are when the exit occurs. This is likely operating as designed. Any reason you're not using a heat and cool setpoint instead of multi-state exit conditions?

2 - Timers going negative are by design so that you can see the additional overrun time.

3 - It does look like a bug that the exit condition doesn't fire for <=0 when it is negative.
brahn
Posts: 543
Joined: Thu Dec 13, 2012 11:01 am
Bot?: No

Re: Public Beta 4.0 Beta 2 Issues

Post by brahn »

I agree with Jon on all points.

I'll need to take a look at the <= 0 exit condition when entering a state with a zero/negative timer value. That should trigger.
gthielen
Posts: 7
Joined: Sun Mar 08, 2015 2:19 pm
Bot?: No
Location: San Clemente, CA

Re: Public Beta 4.0 Beta 2 Issues

Post by gthielen »

Jon - Thanks for your responses.

1 - I am using this with a chest freezer and plastic fermenters, so the temperature swings are extremely small and gradual. I am not using heat and cool setpoint because by the time 5-10 gals of wort cools to the desired setpoint, the freezer has way overshot the low temp and will continue to cool the wort even more, triggering an unnecessary heating cycle. Heating is roughly the same, but the heating element I have is much more gradual so I could probably use a setpoint with hysteresis for that. To compensate for the huge cool temp overshoot, I use an ambient sensor in the freezer and exit the cooling step when it reaches a low point, then have it return to a compressor delay so it doesn't cycle too much. Maybe there is an easier way to do this, but this was what I came up with.

Incidentally, It would be great if the controller allowed a temperature value to be set somewhere, then referred to in the exit conditions with or without an offset value. This way I wouldn't have to change each step for a different fermentation temperature.

2,3 - Ok no problem on the timer going negative as long as it triggers correctly.
JonW
Site Admin
Posts: 1726
Joined: Sun Jul 18, 2010 7:51 am
Bot?: No
Location: Huntington Beach, CA
Contact:

Re: Public Beta 4.0 Beta 2 Issues

Post by JonW »

gthielen wrote:Incidentally, It would be great if the controller allowed a temperature value to be set somewhere, then referred to in the exit conditions with or without an offset value. This way I wouldn't have to change each step for a different fermentation temperature.
You can do this with the recipe utility at http://embeddedcc.github.io/bcs-recipes/

When you define your variables in the "Variable Setup" portion, use the same name for multiple locations where you want a single temperature set across those locations. When you go to the "Recipe Values" input tab, you will only need to set the value once and it will write to each of the defined locations (that use the same variable name).
gthielen
Posts: 7
Joined: Sun Mar 08, 2015 2:19 pm
Bot?: No
Location: San Clemente, CA

Re: Public Beta 4.0 Beta 2 Issues

Post by gthielen »

Thanks for the info regarding the recipe utility.

I had started writing my own php script to update my process using just settemp and differential values similar to what the recipe utility would do - it's now working well. I have to say I like the 4.0 json api much better than the 3.7 api! (You probably already know this, but the docs for the /process api are missing the "states" array).

Regarding my issue #1 (exit incorrectly firing base on temp value), I turned on the datalog and based on that, the temperature never hit the exit value but the exit was triggered. The exit value is 65, and the datalog shows that the lowest temp value over a much longer time period than when the exit was triggered was 65.2 (I set the datalog axis to 65 - 66 to get a very granular view). Could there be some kind of floating point inaccuracy that might account for this?

Thanks!
gthielen
Posts: 7
Joined: Sun Mar 08, 2015 2:19 pm
Bot?: No
Location: San Clemente, CA

Re: Public Beta 4.0 Beta 2 Issues

Post by gthielen »

I just realized that the datalog was set to sample every one minute, so I might have missed a spike that dropped the temperature to the exit trigger value.

That being said, I see a lot of people talking about temperature fluctuations on this board. I have tried to follow some of the examples to reduce this fluctuation (namely tying a BCS ground lead to AC ground) and it seems to have helped, but I still get +/- 0.1-0.3 degree fluctuations from sample to sample. I know this is probably minimal especially in the context of a fermenter, but would it be feasible to provide a temp probe configuration that says "use the average of the last n samples"? This might help smooth out some of the temperature fluctuations.

Also - any idea when a fix for the timer not triggering the exit condition will be made available? I have to keep watching the controller and manually activating the cool state when the timer expires.

Thanks!
brahn
Posts: 543
Joined: Thu Dec 13, 2012 11:01 am
Bot?: No

Re: Public Beta 4.0 Beta 2 Issues

Post by brahn »

It is already using an average of the last 32 values. I can try tweaking this and see if it makes much difference.
gthielen
Posts: 7
Joined: Sun Mar 08, 2015 2:19 pm
Bot?: No
Location: San Clemente, CA

Re: Public Beta 4.0 Beta 2 Issues

Post by gthielen »

What is the sample frequency of the temperature probes on which you're basing the average? Are you sampling once/second and using a 32-second average for triggering exits, or sampling multiple times/second?

Are you triggering exits based on an individual sample, or the averaged value? If on an individual sample, this might explain why I never see the trigger temperature get hit but the process acts like it was.
Post Reply