Better Probe Calibration technique

Things to come.
dtfeld
Posts: 61
Joined: Thu Oct 27, 2011 7:02 pm
Bot?: No

Better Probe Calibration technique

Post by dtfeld »

After owning and brewing with the BCS-462 for several months, the one thing that drives me crazy is the probe calibration/recalibration procedure for a new install or replacement probe. Its an exercise in futility.

What I would like is a way to set the temperatures that i read on screen to a known temperature value, not a resistance value. This would be exactly like an adjustable dial thermometer where you simply rotate the dial to match your "known temp" from a lab grade calibrated thermometer. Since its probably not linear, we could have three points of measurement measure to creat on offset to make the probes read the correct temperature.

Of course I'm open to other ways of improving this, but calibrating 8 probes spread throughout the basement is impossible since its physically impossible to get them all in 1 container without a LOT of work!!! Its a lot easier to move my one good lab grade thermometer!
User avatar
ECC
Posts: 676
Joined: Fri Sep 12, 2008 12:29 pm
Bot?: No
Contact:

Re: Better Probe Calibration technique

Post by ECC »

Do you really calibrate every new probe? I've found that the manufacturers default coefficients work very well, usually within a degree. I can count on one hand the number of times that I needed to calibrate a probe.

It is possible to 'tweak' probes without having to record resistances. For example, with the probes that we sell on our store (which the BCS has the coefficients set as default) you have these three pairs of temp/resistance (more info here)
32 29481
104 5595
158 1991
And from the calculator, the coefficients are calculated as:
A=0.0010281797
B=2.3924024E-4
C=1.5657565E-7

Now, if you always know that a probe is 2deg too high (over the full range), just subtract 2
30 29481
102 5595
156 1991
Run those through the calculator and get your new coefficients.

In this way you can take empirical data to adjust the probes instead of painstakingly recording temperatures and resistances. All probe manufacturers should have readily available a chart of temp vs. resistance, so its usually only a matter of picking 3 pairs, and modifying them if necessary.

If there is enough interest, maybe we could find a volunteer to write a utility to do what you're asking. All of the BCS interface is documented, so this utility could read in the temperature configs, provide the user with a nice gui to increment/decrement the setpoint, and then re-calculate coefficients based on the Steinhart-Hart equation, and finally write the config to the BCS.
dtfeld
Posts: 61
Joined: Thu Oct 27, 2011 7:02 pm
Bot?: No

Re: Better Probe Calibration technique

Post by dtfeld »

Actually, I do calibrate every one as they are close but not dead on. I built an apparatus with 5 Tee's with a place to put the lab thermometer and up to 4 probes then pump water through at 104 and 158. ( I do use an ECC probe as my standard as it is very close to my lab thermometer). I use a ice bath to get the resistances at 32. Once the temp has stabilized, I take the 20 readings and average them out then run through the coefficient calculator.

Using this procedure I get very accurate and repeatable results across multiple probes. Its certainly very close to the glass/mercury thermometer over the range I use. Maybe a little anal, but that's why I chose the BCS.

I have a quasi feed forward system running and it helps to know that the displayed temp difference between the HLT an the MT is actually a temperature difference, not an instrument error and the appropriate offset can be programmed. On the fermentor side, I lager and I want to make sure that I know the "real" temp otherwise I'd have a broken glass caraboy shard slushy if it were reading high!

With all that said, an app would be great as my proccedure take a couple hours.
User avatar
bbrally
Posts: 210
Joined: Sat Mar 27, 2010 3:59 am
Bot?: No
Location: Vancouver, BC
Contact:

Re: Better Probe Calibration technique

Post by bbrally »

I have the same issue. It drives me nuts to see two identical temperatures reading .2 degrees off.

I thought I'd try my hand at writing an app to do what I think dtfeld was asking for.

It's an adobe air application. (adobe air can be loaded from from http://get.adobe.com/air/ )

I wrote this sitting in a hotel room far away from any BCS, so I haven't been able to test it. Perhaps someone can try it out and let me know if it works. (My only concern is the save to BCS function)

The only parameters that are adjustable are the url address and the setpoint adjustment. It should be pretty self explanatary.

Here's the link to the file:

https://docs.google.com/folder/d/0B3hDP ... U4aEU/edit
Last edited by bbrally on Mon Sep 17, 2012 9:33 am, edited 1 time in total.
User avatar
bbrally
Posts: 210
Joined: Sat Mar 27, 2010 3:59 am
Bot?: No
Location: Vancouver, BC
Contact:

Re: Better Probe Calibration technique

Post by bbrally »

I've added a flash version that can be run in a flash enabled browser.

https://docs.google.com/folder/d/0B3hDP ... U4aEU/edit

Just put the three files in the same folder and double click on "thermistor coeffifient changer.html"
Last edited by bbrally on Mon Sep 17, 2012 9:33 am, edited 1 time in total.
bruin_ale
Posts: 6
Joined: Mon Feb 14, 2011 10:27 am
Bot?: No

Re: Better Probe Calibration technique

Post by bruin_ale »

Couldn't get the flash version to work, but the adobe air one works really well! I was a little confused about what to enter for the setpoint adjust. If my readings are 2 degrees F high, do I enter +2 or -2?
User avatar
bbrally
Posts: 210
Joined: Sat Mar 27, 2010 3:59 am
Bot?: No
Location: Vancouver, BC
Contact:

Re: Better Probe Calibration technique

Post by bbrally »

I actually forgot all about this. Since I got back home I haven't even tried this out.

The way I envisioned it was to take the temp reading that was felt most accurate (or the average) and then plus minus the others to bring them all in line.

for example, t1 = 150.6, t2 = 150.9, t3 = 151.1

In order to make all temps equal to t2:

t1 setpoint = 0.3
t2 setpoint = 0
t3 setpoint = -0.2

When I get home tonight I'll plug in the BCS and try this out. I'm brewing on Monday, so I'll definitely confirm this before then.

I stared playing with the idea of programming in HTML5 instead of Flash so that there would be less "flash" problems, but that curve seems pretty steep at the moment and the only time I usually have to do this sort of thing is when I'm at work. (pretty tough when the only free time you have is when you're working:).
User avatar
bbrally
Posts: 210
Joined: Sat Mar 27, 2010 3:59 am
Bot?: No
Location: Vancouver, BC
Contact:

Re: Better Probe Calibration technique

Post by bbrally »

I made some time and found that there were problems. Mainly my calculations were using Celsius, but my setpoint was in Fahrenheit.

I've changed the program and tested it and it seems to work, all of my temps which were different by +-0.5 degrees are now within 0.1

I played around using smaller and smaller setpoints to get the numbers even more inline, but noise in the sensor probably starts being greater than the actual difference in temps.

If you make multiple adjustments to the setpoint, remember that the "calculate" button will adjust the values from the coefficients in the "current BCS coefficients" column, which doesn't change when a save is made. So you may want to load new values between each adjustment so your calculations are using the current BCS coefficients.

Just to be safe, you should save a copy of your BCS system settings so you can reload your original coefficients if there's the need.

Here's the new files:

https://docs.google.com/folder/d/0B3hDP ... U4aEU/edit

If you want to use a flash enabled browser, just download "thermistor coefficient changer.html" into a folder and then allow folder access to that folder through the flash player settings on your computer. Double clicking that file should run it in your default browser. The settings are located here: http://www.macromedia.com/support/docum ... ger04.html
Last edited by bbrally on Mon Sep 17, 2012 9:34 am, edited 1 time in total.
dtfeld
Posts: 61
Joined: Thu Oct 27, 2011 7:02 pm
Bot?: No

Re: Better Probe Calibration technique

Post by dtfeld »

bbrally.

Thank You Thank You Thank You. :D :D :D :D

That is beautiful and has made my life much easier. I can now calibrate to a "known" value I measure with a lab grade thermometer and everything looks really nice.

You Da Man!!!
Drew
Posts: 12
Joined: Wed Oct 26, 2011 9:16 pm
Bot?: No

Re: Better Probe Calibration technique

Post by Drew »

Is there a new link to calibration program? It appears that google docs is now google drive and the above link is no longer valid. Thanks!
Post Reply