How to get my DS18B20 into my ferm bucket?

So I spent the afternoon wiring up a DS18B20, along with the proper resistor, etc, to some female pin headers, then writing a program on the RPi to log that temp.

Unfortunately, the temperature from a stuck-on DS18B20 is nowhere near what either the fermometer or digital are saying, usually about 15 degrees cooler.

is there a way, other than drilling a second hole and dropping a legitimate thermowell down into the bucket, that I'm going to get an accurate reading?

this is the sensor I have

Did you try putting insulation over the sensor and giving the system some time to come to equilibrium?

What sort of insulation would you suggest?

What about drilling 2 holes in 1 stopper?

I am doing the same thing at the moment and may end up with something like this.

http://morebeer.com/view_product/16672


Once the sensor is in the thermowell I was thinking of filling it with a non conductive liquid or filling it with epoxy to get a better reading.

I tried adding foam tape to the outside of my stuck on DS18B20 and was still a few degrees off.

-Jason

I use the plastic buckets, so if anything, i would need to drill another hole and put another grommet that would (hopefully) hold the thermowell in place...

I just put a bunch of folded up paper towels and 6-7 layers of folded bubble wrap, duct taped over. seems to be a bit better, only about 8 degrees off.

if that consistent, I can at least take that into account to get a fairly accurate idea of what my wort temp would be.

what are some better ideas for insulation?

I have a thermowell but don't use it much anymore because when paired with my controller I get some pretty significant undershoot on cooling (like 8F degrees or so). It would probably would work better once the system got closer to the set point, but I've just found it not to be worth the extra effort of sanitizing it etc. So currently I place my J-type thermocouple in contact with my plastic or glass fermentation bucket and cover it with a drywall sanding spounge. Any similar size spounge would probably be OK.

When you see that temp diff are you fermenting? If you are reading that sensor with 9 bits, it seems like the total inaccurracy could be as high as 1.4F. Throw in some inaccuracy for the other meter and the exothermic action of they yeast, and 8F might not be too unreasonable.

make your own thermowell, you can for like 10 bucks. buy a chrome toilet line, they are solid brass covored in chrome so it's safe. they come long like 22 inches, so you need to cut it with a dremel or something ,just measure it so it's half way in your wort, then seal the end with aquarium silicone, you can get it at any pet store, it's 100 percent silicone once it cures,so it's safe in your wort. it fits a ds1820b perfectly. I drilled a hole in my bucket lid just big enough to fit the lower bit in and then push the flared bit in for a tight fit. I seal the top with some extra silicone and slide my sensor in. works great, I don't bother sealing the top.

you can see my arduino reporting to cosm here. if interested u can supply info to recreate my setup, the uno and Ethernet shield we're like 30 from dealextreme,and you already have the sensors

https://cosm.com/feeds/83208

the supply lines look like this
http://ecx.images-amazon.com/images/I/11mNg+mPBeL._SL500_AA300_.jpg

they are 4.49 here at Lowe's. the silicone was like 10 for a tube that can beused to seal many things for brewing

Of course when i post to go look at mine it goes down lol...i think im still having power issues and possibly water getting into my tube screwing up the sensors.

More work to be done!

When it does work its great though lol...at 12 bits you get 0.11 degree resolution...a bit over kill but whatever.

Thanks. Mind if I ask what gauge wire you used to solder to the leads on the sensor? This was my first soldering project, so it's obviously awful, and will probably have to wire up another sensor to get it down in such a small tube.

For reference, the difference I was referring to being 8F was between what my mercury thermometer sitting in the water in the bucket was compared to the sensor on the outside of the bucket.

I actually went to home depot and bought some "Thermostat" wiring, its like 33c per foot and comes with 4 wires.

I picked it because i wanted a thin cable that had a minimum of 3 wires, and was flexible...everything i could find was solid core and too rigid for what i wanted.

I honestly have no idea what gauge it is, its pretty small though, if i had to guess i'd say 20 gauge.
Something like this


It came with 4 wires, i just snip back one of the wires. As a reference, its flexibility is similar to a Cat5 network type cable, but obviously thinner...another thing i wanted to try and find but couldnt was just regular telephone wire like you would use inside to connect a phone to the wall as its flexible and small...but all the stores only had phone wire for inside the walls which was rigid.

The tube is surprisingly large, you should have space to spread the legs out a small bit to solder and it will still fit.

I had issues with taping the DS1820B to the side of the bucket, no matter how much insulation i put on(and i put on so much it was ridiculous) you could watch the sensor begin climbing the exact moment my home heater kicked in...sure my ambient was going from 66-70 and the insulated sensor was only going from 66-67, but it was still annoying and ruined what i wanted when every 90 minutes my temperature spiked up and down when the heater came on.

Then on the far end of the cable that connects to my arduino in pins, an inch or two above the end i stripped the shielding off and soldered the 4.7k resistor between data and 5v.

Im still trying to get a better way set up to have a set of 3-4+ cables that are easily interchangable and i can plug/unplug at will...right now everythings hard wired to my arduino so if something goes wrong its a pain to swap cables. Not to get onto my project tangent, but just for your information i was hoping to use a telephone/cat5 jack, and my cables have a male connector with 2 sensors each. Then adding/removing sensors is as easy as plugging unplugging a telephone wire.

I have used these in the past with a stopper I drilled an extra hole in. Sensor fits just right. Not to tight and not to loose.

http://www.brewershardware.com/Straight-Wall-Thermowells/

rockytoptim said:

I have used these in the past with a stopper I drilled an extra hole in. Sensor fits just right. Not to tight and not to loose.

http://www.brewershardware.com/Straight-Wall-Thermowells/

Click to expand...

Thanks i may actually end up going with these rather than my DIY version.

No matter what i seem to do, i cant seem to get it to stop leaking in with silicone.

My only other option is to silver solder the bottom, but i have no silver solder and that would be 10-15 bucks im sure...

Anyone ever had issues with these DS18B20 sensors reading a few degrees low in a thermowell? I have one in a 14" brewers hardware thermowell down in my wort. The sensor in the thermowell reads about 3.5 degrees lower than the others whether it's in the air or in the wort. With 3.5 degrees added, the thermowell sensor reads 67.7 while one taped to the bucket reads 64.6 with no compensation. I'm not sure which to trust. I've always trusted the one stuck to the bucket before getting the thermowell. Now I'm concerned I've been fermenting 3.5 degrees warmer than I thought.

What happens when you switch the sensors Spintab? Do you still get a 3.5°F difference? Plastic is a pretty poor conductor of heat so I would tend to trust the reading obtained from the thermowell DS18B20 over that obtained from the outside wall of the bucket.

That is the one thing I haven't tried. I'm using a 4 wire phone cable which requires it's insulation be shaved down a bit to fit in the thermowell. I didn't want to do that quite yet but I will have to soon. I have tried taking the sensor out of the thermowell and putting it right next to another one. They read the same. Also my dial thermometer, which I usually really trust, reads the same as the thermowel sensor +3.5 right next to it in the wort. Take the entire thermowell out and stick it next to the air temp sensor and they read the same after the +3.5 adjustment.

I was in mid panic when I first posted so I didn't do as many tests as I should have. I finally will have some time this weekend to do a better job and I will let you know what I find. Thanks for replying.

Spintab said:

That is the one thing I haven't tried. I'm using a 4 wire phone cable which requires it's insulation be shaved down a bit to fit in the thermowell. I didn't want to do that quite yet but I will have to soon. I have tried taking the sensor out of the thermowell and putting it right next to another one. They read the same. Also my dial thermometer, which I usually really trust, reads the same as the thermowel sensor +3.5 right next to it in the wort. Take the entire thermowell out and stick it next to the air temp sensor and they read the same after the +3.5 adjustment.

I was in mid panic when I first posted so I didn't do as many tests as I should have. I finally will have some time this weekend to do a better job and I will let you know what I find. Thanks for replying.

Click to expand...

I test all of my temp probes with both an ice bath and in boiling water. That will tell you whether the probe in your thermowell is actually off or not. When doing the ice bath you have to constantly stir the probe around otherwise you won't get all the way down to 32, especially if your thermowell is much longer than the glass of water you are using. In that case the part that is exposed to air is transfering thermal energy into your ice bath creating a localized pocket of warmer water. I use 1% thermistors for all my temp sensing needs and consistently read within 0.1F at both ends without any calibration. After they pass those two tests I can always trust what they are reading.

crane said:

I test all of my temp probes with both an ice bath and in boiling water. That will tell you whether the probe in your thermowell is actually off or not. When doing the ice bath you have to constantly stir the probe around otherwise you won't get all the way down to 32, especially if your thermowell is much longer than the glass of water you are using. In that case the part that is exposed to air is transfering thermal energy into your ice bath creating a localized pocket of warmer water. I use 1% thermistors for all my temp sensing needs and consistently read within 0.1F at both ends without any calibration. After they pass those two tests I can always trust what they are reading.

Click to expand...

+1 to this. I'm a a little behind your guys since I just now ordered by arduino uno and DS18B20's but my plan of attack is very similar to what crane suggests. I will bring RO/DI to a boil, insert the thermowell with temp probe and note the temperature. Then call a local weather station and get the absolute pressure. This will let me correct the boiling point away from 212°F. Then immerse the thermowell+probe in an ice/water slurry and stir well continuously. Note the temperature reported by the probe. These two data points will allow me to construct a two-point calibration curve for each of my DS18B20's. If I am feeling particularly anal (happens often), I'll do a third point in boiling acetone provided I can control the vapor. A three point calibration curve with a midpoint smack in the middle of the temperature range we are interested in is pretty damn hard to beat. Couple that with the fact that we can read these temperature sensors to the nearest fractional degree F and we have some serious precision and accuracy.