DIY temp probes with DS18B20

I've read some very helpful threads already -- summing up my plan w/ some questions. I could just buy these from Derrin -- but that's hardly any fun

https://www.homebrewtalk.com/f51/make-your-own-precision-ss-temperature-probe-101192/

http://theelectricbrewery.com/temperature-probes


Parts

- 4" stainless probe ends (from Brewers Hardware)
- M8 Male cable w/ leads (for the sensor end)
- M8 Female/Female cable (sensors to project box)
- M8 Male connector (for project box)
- DS18B20 (temp sensor)
- thermal paste
- heat shrink (adhesive & normal)


The Plan:

- solder the DS18B20 to the M8 Male w/ leads (may shorten the leads to ~4" outside the probe)
- cover all exposed connectors/solder joints with heat shrink
- add thermal paste to bottom of thermal probe
- insert probe/cable assembly and seal the top with adhesive heat shrink

It all seems pretty simple really (famous last words...)


Questions:

- I plan to run these in non-parasitic mode as it has been more reliable with my futzing around so far. Does it matter in practice other than needing to do a strong pull-up when doing the temp conversion?
- Do I need a cap in the probe end for anything? I've seen mentions of small caps across Vdd and GRD.
- Is there value in 4-wire cables? Are they actually twisted pair inside? If I read write, I should run DQ/GRD on one pair and Vdd on the other pair.
- Is the thermal paste going to potentially create problems shorting out sensors?
- Any cheap sources for M8 cables/connectors?

4" seems a bit long. The response time of the sensor will be greatly reduced. (I've tried, took almost a minute to register a change)

Heatshrink the three leads individually, but do not heatshrink the plastic.

Thermal paste usually has a high thermal conductivity, but a low to non-existent electrical conductivity. No need to worry about a short from that.

The caps would help on sensor like the LM34/35, but this is digital and not necessary for the length you are using (assuming less than 20-50ft).

Good choice on the DS18B20, I had 5 running on 1 wire and all within +-0.5*C. I would recommend not using parasitic power, just because of longevity and the fact that it is not necessary in this case.

In my probes I use a 4-conductor phone cable and use some old phone jacks so I have a modular design. I also use 1.5-2" stainless tubes, thermal paste inside, stress relief with silicone inside then the outer opening with 'WaterWeld" by JB Weld. (food safe, 300*F). Response is probably around a few seconds, never checked it.

Of course this is all linked up to an Arduino...

4" seems a bit long. The response time of the sensor will be greatly reduced.

Click to expand...

i dont agree. i use several probes made almost exactly as the OP plans, and they begin to register a change nearly immediately. they are 2, 6 and 8" long. obviously, the more material to heat or cool, the longer it will take to do so. normally it takes less than 10 seconds to register a 10-20 degree change in my experience. the amount of material directly around the sensor is what matters most, and tha tis the same regardless. it doesnt matter much if there is 2 or 20 inches of tubing behind it.

Thermal paste usually has a high thermal conductivity, but a low to non-existent electrical conductivity.

Click to expand...

that greatly depends on what exact thermal paste used. i didnt use thermal paste, as i only had arctic silver 5 on hand, and that is conductive. i ended up mixing 90-95% powdered aluminum, 5-10% silicone caulk, by weight. works just fine. as long as the paste is specifically non-conductive, its fine to use.

no capacitors are necessary. only a 4.7k pullup resistor on the data line. capacitors would prevent the controller from pulling the line low, and would prevent the reset() commands from working.

i agree that if you can easily avoid parasite power, do that. using only two wires /should/ be fine, but if using a 3rd wire isnt a problem at all, just do it.

audger said:

i dont agree. i use several probes made almost exactly as the OP plans, and they begin to register a change nearly immediately. they are 2, 6 and 8" long.

Click to expand...

I may have over exaggerated the probe length, but I would still recommend the smallest size to comfortably fit everything inside.

audger said:

that greatly depends on what exact thermal paste used. i didnt use thermal paste, as i only had arctic silver 5 on hand, and that is conductive. i ended up mixing 90-95% powdered aluminum, 5-10% silicone caulk, by weight. works just fine. as long as the paste is specifically non-conductive, its fine to use.

Click to expand...

Arctic Silver 5 is a thermal paste (compound) and is most certainly not electrically conductive. I use it on CPUs + GPUs all the time.
Arctic Silver Incorporated - Céramique 2

___

I would recommend keeping the 4K7 resistor outside of the probe to avoid space constraints and possible heat addition.

you dont need thermal paste if you go with the brewershardware probe ends. The sensors fit snuggly into the probes without much room between the sensors and the probe wall, i just built four of these senors in a netduino project last week. I used 3 10" probes and 1 4" probe.

I would go with longer probes because you want to have your sensor in the middle of your pots. With the 1/4" compression fittings also sold at brewershardware you can adjust the probe depth as long as you used the o rings and not the permanent metal compression pieces. Just my $.02.