Mounting of DS18B20 in Keggle

Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum

Help Support Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

waldzinator

Well-Known Member
Joined
Feb 17, 2013
Messages
147
Reaction score
11
Location
Waxhaw
Hi,

I just started all-grain brewing, but I am a hobbyist who enjoys all things computers and programming. My current setup is two converted kegs for my HLT and Kettle, and a 10 gallon cooler for my MLT. I have some experience with using the Arduino Uno. I have read a lot of good things about the One Wire protocol on the Dallas temp sensors. I can handle the electronics side of things. I have a more practical question. How are you mounting them in your keggles? I see brewinghardware.com sells the sensor with pre-attached thermowells. What's the best way of mounting these in the keggle? It appears they use a compression fitting. What size hole do these fittings take?

I realize this is kinda a broad question. Just trying to get some ideas before I go drilling into a perfectly good keg and realizing I screwed up.

Thanks.
 
1/2" NPT threaded thermowell w/ a 1/2" NPT nut and o-ring. I siliconed the temp sensor directly into the thermowell after wrapping the electrical connections in shrink tubing.
 
Makes sense. I was torn between doing it that way and this . I guess I was kinda leaning towards the compression fitting since they also sell a probe epoxied into a SS tube. However, I guess the end result is the same, and I wouldn't have to piss around with a compression fitting. Thanks!
 
I made my own Tri-Clamp ones (although this is probably not much help to you).
Had the end of some 1/4" .020 tubing TIG'd shut (you can buy them premade), then silver soldered in a triclamp end-cap. Hasen't been cleaned up in the pic yet.

I soldered a DS18B20 onto the end and also added a 1k res and 0.1uF cap for filtering (suggested by the brewtroller guys), it was a pain. Excuse the sloppy soldering, I was getting pretty tired of making them at this point.

I then just inserted the probe with a bit of silicone heatsink compound (response is very good), crimped the tubing on the cable and added some adhesive heatshrink for good measure.

I highly recommend brewershardware.com, I have bought alot of gear from them and the service is outstanding (although they can be slow to respond at time, they seem to be a bit overwhelmed)

IMG_0897.jpg


IMG_0881.jpg
 
After I had to replace the second homemade probe I built, I started buying them from Brewers Hardware as well. I love them. I have them in my kettle, hlt, rims, and mlt.

Order a stock length; they were real busy last winter when I placed my last order and couldn't do custom lengths.
 
I made my own Tri-Clamp ones (although this is probably not much help to you).
Had the end of some 1/4" .020 tubing TIG'd shut (you can buy them premade), then silver soldered in a triclamp end-cap. Hasen't been cleaned up in the pic yet.

I soldered a DS18B20 onto the end and also added a 1k res and 0.1uF cap for filtering (suggested by the brewtroller guys), it was a pain. Excuse the sloppy soldering, I was getting pretty tired of making them at this point.

I then just inserted the probe with a bit of silicone heatsink compound (response is very good), crimped the tubing on the cable and added some adhesive heatshrink for good measure.

I highly recommend brewershardware.com, I have bought alot of gear from them and the service is outstanding (although they can be slow to respond at time, they seem to be a bit overwhelmed)

Great input everyone. Question, Allan. Were your running in normal or parasitic mode? I've seen varying documentation that in parasitic you can use a pullup res on the board to hold the bus high. In normal mode, I've seen the use of a 4.7k res on the bus. Thanks! Btw, your tired soldering still beats my normal soldering haha!
 
I use a brewtroller, which can do parasitic or normal. I use normal as it is a quicker measurement time. If you've got 3 wires I would use normal, when I've done parasitic with my own hardware you can install a transistor as a switch to provide a stronger pullup (MAXIM has some app notes on suggested ways). You need the 4.7k regardless, but it should be on the controller end.

The 1kohm and 0.1uF I mentioned are to filter the power to the device to provide noise suppression.
 
Ah, gotcha. I thought you meant you were using the 1k in lieu of the 4.7. Thanks. I'm still just piecing together my gear. I initially want to use the ds for passive temp monitoring, but I would eventually like to go all in for full (possibly PID) automation
 
I use cat-5 cable and a phone jack for each sensor. I have four phone jacks on my control box. They are wired to a single pin on my Arduino. I use a 4.7k pull up resistor on that pin. I run the sensors in powered mode because I can take the sensor readings faster and don't have to worry about how heavy the data pin is loaded with parasitic capacitance.
 
PTFE insulated hookup wire is good if you have it, especially for the brew kettle, because it can stand up to 100°C. Cat5 is rated to 60°C so the insulation might suffer from the heat, although doubtless loads of people have used it without any problems.
 
PTFE aka Teflon wire is good to 200C usually (used in military and aerospace primarily), and is massive overkill for this. It is also pricey and rare as it has to be silver plated.

I'm not a fan of using Cat5 though, pick up some 3 conductor cable instead, which is usually good to higher temps (not as high as teflon though). Also as per the MAXIM specs, you should avoid "STAR" style connections (all temp probes going back to the control panel and being jumper-ed together there). While it can work, these probes are meant to be connected in a "daisy chain" fashion.
 
OK, I was under the impression that standard PVC insulation was not good to 100°C but I may be wrong about that. PVC insulation for appliance cable is certainly not rated for 100°C but it may be different for hookup wire.
 
All the PVC insulation I've seen commonly is good to 105deg C (not to be confused with the cables rating). In the case of cat5, the cat5 standard only spec's 60deg and hence the rating.
 
To clarify my use of cat-5: I bought Brewer's Hardware sensors which are the One-Wire sensor sealed in a stainless steal 1/4 tube with three feet of four strand wire. I used the cat-5 to extend the cable. It never sees close to 200F. I assure that the cable Brewer's Hardware is temp rated appropriately, but never thought to check.

About the cat-5 twisted pair. I use three different pair for signal, power, and ground. I float the twisted partner for signal and power, but ground both wires in the pair for ground of the sensor.

I read the Maxim sheet about not using a star point configuration. My first setup I ran from sensor to sensor in one daisy chained network. When I had issues polling each sensors serially w a 750ms delay between the "calc temp" and "get temp" commands, I stated experimenting both with my code and there hardware configuration.

I stopped tinkering when I could poll all four sensors in less than 2 sec while they were in a star point configuration. I send the "calc temp" command to all four addresses, wait 1.6 sec, and then send the "report temp" command.

I agree that this might not be the optimum arrangement, but when I reached this point I stopped having measurement issues.

Sorry if I took OP thread too far off track.
 
Back
Top