Temp sensor fluctuation

I need to calibrate mt temp sensors on my new rig.
I am using Brewershardware sensors with a BCS.
The first problem I ran across, was the temps fuctuating approx 7 degrees.
With the sensor in an insulated mug,with ice water, My old digital thermometer stays steady 37.5 degrees, BUT, the temps read by th BCS change a few tenths at a time, and over the course of maybe 30 to 45 seconds (didn't time it)it went from 33 degrees to 40 degrees and back.
Any advise?

It sounds like you're picking up noise, but I hesitate to say that because the brewershardware stuff tends to be really well made.

Are all of your connections good and clean? Is your probe and wire away from anything that would generate noise?

I don't know where it could pick up anything.
Last nights test, nothing was running except the BCS. No elements,no pumps, nuttin.
The rig is not near any other applainces, ect.

Did you update the temperature probe settings for the Brewers Hardware probes? The BCS ships w/ coefficients that match our probes, the BH probes have their own set. If you try to use settings that don't match the probes you'll get erroneous readings.
http://www.brewershardware.com/temperature_sensor_instructions.html

Hmmm.... the sensor wires run through a DB25 cable along with the trigger lines for the SSR's. That DB25 is shielded though so it shouldn't be picking up any noise, and all the lines are low voltage DC.

maybe try using the coefficients from Derrin again and see what happens.

Thanks for the compliment Malfet!

Layne, I sent you the actual spec sheet from Vishay on the thermistors used to build the probes and the coefficients I came up with and have been using are here: BCS Temperature Probe Instructions

If there does turn out to be a problem with my sensors I'll get them replaced asap.

Is it typical for the temp to 'move'?
My handheld digital thermometer stays steady for minutes ant a time, but all 4 of the temps on my BCS creep up and down. This can't be right.

Layne said:

Is it typical for the temp to 'move'?
My handheld digital thermometer stays steady for minutes ant a time, but all 4 of the temps on my BCS creep up and down. This can't be right.

Click to expand...

It depends what you mean by move. In fully analog systems, there will always be a little bit of noise, and even in digital terms you can expect to see a sensor flip-flop a high and low bit when between readings. That should be small, though. How much depends on your system, but generally it should be less than a degree (possibly much less). Your digital thermometer has this too, though it might have some kind of display hysteresis to prevent you from seeing it. If you are getting a 7 degree swing, something is wrong.

Just looking at every possibility.
I have wireless surround sound in the brewroom. Could that possibly cause interference?

Layne said:

Just looking at every possibility.
I have wireless surround sound in the brewroom. Could that possibly cause interference?

Click to expand...

I guess in theory, but (like Boerderij mentioned) it seems unlikely if everything is shielded. My money is on some kind of configuration error or some kind of power stability problem. I'm not very familiar with the BCS, though, so I'll leave that for others to speculate.

Layne, are all the sensors acting the same? Or is one different from another?

They all fluctuate constantly.
I have not compared them to see if it is a similar amount, or the same time/pace, but I do know that on two different nights, I tried to calibrate. One night on the phone woth you BK, and last night. And those were NOT the same sensor, but had similar results.
I don't remember how much range the first time, but last night, it was 7 degrees.

This is the cable that runs from the sensors in the main box to the remote box:

http://www.l-com.com/item.aspx?id=5882

I am of the understanding that the fully shielded nature should protect the sensors from noise???

I did run the sensors past some 220VAC wires in the main box (unshielded) but i didn't think that would make a big enough difference...

In this pic, you can see where the AC runs (orange wires, the other hot leg isn't installed in that pic, but takes the same route) and where the sensor wires enter (the four white cables).


Here you can see how they are next to each other for about 4 inches in the conduit.


I have had a lot more AC/sensor interaction than this with no noise issues before, but maybe the BCS is very sensitive to noise. I need to figure this out.

I am really sorry about this headache Layne. When I fired it up, the sensors read evenly but were just about 1-1.5 degrees off from one another. Have you tried re-entering the factory coefficients???

Layne, are you operating any of the other outlets during the calibrations or readings?

Additional info, the shielding is grounded.

Nothing else is on.
I just put in Derrins coefficients and it's WAY better.

They are still off, but at least now I believe I can adjust them, now that they are staying constant.
I am brewing right now. I'm using my hand held temp gauge.

A couple questions, is the BCS powered by a wall wart power supply?, and are you a ways from town or a power substation. The condition shown on the screen and what you describe are symptoms of poor voltage regulation, as the thermistor circuits are quite sensitive to voltage fluctuations. If the service voltage is fluctuating and the power supply is not regulating very well, the thermistor readings will be bouncing around like shown on the screen shot.

This is good. A tenth here and there is MUCH better than 7 degrees!
All the temps are off from my hand held thermometer, but at least they are consistant.
Tomorrow I will try to fine tune it.
Thanks a million guys.

How sure are you that the hand-held digital is correct? I would check that too...

The coefficients provided with the Brewers Hardware probes are ok, but far from perfect. Could be the nature of the beast, but I found that the original BCS probes were more accurate with the provided coefficients.

Yes, the last screen shot looks much better. Varying a few tenths is pretty normal and well within spec of these probes, particularly when they're not submersed. You might have to calibrate them to remove the offset error.

Just to answer some of the questions above, I don't believe this is an issue with voltage regulation. Using the default temperature coefficients that are meant for a different sensor can have the tendency to magnify the error/sensitivity, which is what you were seeing. The BCS is sensitive to noise, but most people get by without using shielded cable. Its always advised to use shielding if issues arise, particularly if you're running the wires near noisy EMF generators, like compressors, motors, etc.. If you're going to run it in conduit with AC lines, its a good idea to use shielded cable. It sounds like you are already a step ahead, by using shielding so that you won't have to worry about it.

If you haven't already, check out our newest firmware release, v3.3. It has dynamic charts and some other nice features. It also as an improved oversampling/filtering algorithm that will smooth out your temp readings even more.

Kudos on the panel build. Very clean and well laid out. Looking forward to reading more!

Sounds like a bit of engineering is needed for this platform with wandering sensor readings and noise sensitivity problems that need addressed.

kladue said:

Sounds like a bit of engineering is needed for this platform with wandering sensor readings and noise sensitivity problems that need addressed.

Click to expand...

Not sure what you're referring to. I think we've ruled out noise sensitivity and wandering readings.

Yeah, so far, the main problem has been operator error :0
We're close now, and as soon as confirm wether my hand held digital thermometer or the sensors are correct, I'll be able to finish the fine tuning.
Thanks everyone for your help.

I was referring to post #15 display of the sensor readings where it appears that the samples were taken on 3 second intervals and there was noticeable fluctuation in the readings where there should not be any fluctuation. The other issue that has been noted in other posts is the sensitivity to noise generated by the direct spark ignition systems which seems to cause major upset of the platform when it is near the ignition modules.
This is just my observations of published accounts, I just compare this platform to the commercial and industrial platforms and sensors I have worked with for the last 20+ years, which is probably an unfair comparison. I am not trying to be critical but it just looks like there is room for improvement here in both hardware and software.

kladue said:

I was referring to post #15 display of the sensor readings where it appears that the samples were taken on 3 second intervals and there was noticeable fluctuation in the readings where there should not be any fluctuation.

Click to expand...

Yes, I believe this is a free air measurement so its hard to tell if the fluctuations are real or not. But regardless, I read a maximum of 0.4deg fluctuation reported in the 21 samples, or over a full minute. The ECC sensors sold on our site are spec'ed with an accuracy of 0.2degC, and I believe the BH are similar. Considering these variables, I don't see anything unexpected. With the latest firmware, along with the new oversampling/filtering that i mentioned above, we implemented data logging to the 10th of a degree, and the ability to export it to a csv (and view it dynamically online). I would love to see more data here.

kladue said:

The other issue that has been noted in other posts is the sensitivity to noise generated by the direct spark ignition systems which seems to cause major upset of the platform when it is near the ignition modules.

Click to expand...

Agreed, the the BCS is not an isolated system, which imho is a requirement in that application.

kladue said:

I am not trying to be critical but it just looks like there is room for improvement here in both hardware and software.

Click to expand...

There's always room for improvement. Thanks for your constructive criticism.

The thermal mass of the probes would prevent free air fluctuations like shown in post #15, there has to be another reason for the displayed variations, voltage or analog input conversion fluctuations.
The Fluke 51-2 thermometer with exposed junction thermocouple that we use for calibration does not have this type fluctuation when exposed junction is in free air,and it's thermal mass is considerably less than the probes.
All of the calibration test equipment we use receives a NIST traceable certification in a cal lab on a semiannual basis, and the fluke is in mid cycle. Just for grins and giggles I should buy one of the BH probes and have the techs drop it in the dry block calibrator and plot the results, and then free air the same probe to see if I can reproduce the effects shown.

I was in the same situation as you, Layne. BCS-460 with Darrin's probes and some odd fluctuations. I then did the calibration as described here:

http://www.embeddedcontrolconcepts....f=6&t=10&sid=a8cafb2f9edd02ef6c06b171be02c67e

My fluctuations are pretty tiny now, within a tenth of a degree or so. It would seem that this may help you out even more since this calibration would be specific to your system rather than taking generic coefficients.

My only question would be, on the Thermistor Calculator page, it has an option to select your model. The context is that the model selection is completely optional, but does anyone know if any of those models correspond to the one's Brewer's Hardware sells?

TheAleMaster said:

does anyone know if any of those models correspond to the one's Brewer's Hardware sells?

Click to expand...

No, that's just a calculator that we hijacked from another company. We should really get our own. It just solves some equations to spit out the coefficients.

kladue, you're not comparing apples to apples with an exposed thermocouple and a thermistor probe, although maybe the reduced thermal mass of the bare thermocouple makes up for the increased sensitivity of a thermistor. An interesting experiment nonetheless.

Thanks guys.
Right now, I'm trying to locate a 'correct' thermometer to use for the calculations.
I've asked all my brew buddys, and most use something that is 'close enough'.
So, I'm on hold till I can find a known accurate thermometer for calculating.

There are 2 possible ways to test the probes, crushed ice in water, and boiling water with altitude boiling point correction. For the crushed ice method you need to give the ice melt process about 5 minutes after mixing ice and water to let temperature stabilize, then the temperature should be as close to 32 degrees as practical with this method.

OLD THREAD REVIVE!

(x-post from ECC forums)
I'm having a similar problem hope its okay that I piggyback on this thread. Checked ground wires, I have one ground going from GND0 on the BCS to the A/C ground and one going from the SSR Ground jumper to AC ground. AC power is about 8 inches away from my probe wires in the box. Readings are looking like this: these are from 2 separate "runs" one at 32 degrees and one in 62.6 degree water verified by a calibrated thermo. Probes are from Brewhardware and I used the numbers from their site in my BCS settings. Also tried calibrating it with the calculator found on ECC forums. The behavior seems the same. Any other ideas, gentlemen?

32 degree ice bath:


62.6 degree water: