Make your own precision SS temperature probe

[bull8042]

I was planning on using the LM34 because it was already in F and linear, but what would be a good tracing software to use with it? And, what would be suitable for providing clean 5V? A computer PSU or Wall DC Adapter? How would the sensor connect to the PC?

The LM34CAZ is calibrated to 10mV/°F due to wafer-level trimming making it perfect for this application. I am using DAQFactory for my programming interface and a LabJack U3 (LabJack - LabJack U3 USB DAQ) for my DAQ. Hook up the three wires from the sensor to the LabJack and no other hardware is needed..... except for the PC of course.
As for what other types of displays are available that will work with this sensor, I can't say as I haven't researched anything else. I had decided to use PC control from the start....

 

[WhimWhamWhom]

Can we get this"calvelcade" aka Make your own precision SS temp probe threads put into the stickies category?
WWW

 

[Homercidal]

The LM34CAZ is calibrated to 10mV/°F due to wafer-level trimming making it perfect for this application. I am using DAQFactory for my programming interface and a LabJack U3 (LabJack - LabJack U3 USB DAQ) for my DAQ. Hook up the three wires from the sensor to the LabJack and no other hardware is needed..... except for the PC of course.
As for what other types of displays are available that will work with this sensor, I can't say as I haven't researched anything else. I had decided to use PC control from the start....

Yeah, no I won't be purchasing this for $100. There must be a very cheap and simple method of hooking this thing up. I need to sit down and get all my info figured out. I got a few projects I'm juggling and getting confused.

 

[Brewers Hardware]

Yeah, no I won't be purchasing this for $100. There must be a very cheap and simple method of hooking this thing up. I need to sit down and get all my info figured out. I got a few projects I'm juggling and getting confused.

You can use one of the Arduinos. For about $30 it'll read 6 or 8 LM34's depending on the model and it'll send the data back to the PC as a serial stream. You'll have to figure out what to do with the data from there of course but it's a pretty slick solution. If you really needed to you could cascade 4051 multiplexers and read as many as you wanted.

 

[freddyb]

You can use one of the Arduinos. For about $30 it'll read 6 or 8 LM34's depending on the model and it'll send the data back to the PC as a serial stream. You'll have to figure out what to do with the data from there of course but it's a pretty slick solution. If you really needed to you could cascade 4051 multiplexers and read as many as you wanted.

18B20's are a much better solution if you're going with an Arduino. It's digital and each sensor has a unique 64-bit identifier. Plus, they operate in a 1-wire mode (plus ground) using a parasite power method. I just hooked up 5 of them on a single I/O pin to confirm the operation. No problems. There really is no practical limit to how many you can connect to a single pin. The source is already written and readily available. It took me all of 30 minutes to modify it to function as a temperature controller.

I'll post the code as an example if anyone is interested.

 

[jcdillin]

Yeah if your going arduino I would only go with the DS18's, they work great and no analog to worry about. I've ran 10 so far on one input pin with no troubles at all.

 

[Brewers Hardware]

The DS18 is a great chip no doubt, but the poster was asking about the LM34 specifically. I'm using DS18's in order to free up pins as well. One other nice thing about the LM34 is that it's already calibrated in *F. As Bull pointed out on page one when I first suggested the DS; with .5*C accuracy, making the conversion to *F results in 1.8* accuracy where as the LM34 will be 1*. On the downside the LM34 will for sure need an RC circuit to clean the output and the DS18 will not. Bull found a great cap and resistor that can be fit in the probe body.

I don't think it's fair to say one chip is better than the other. I think it depends on the application. If you need pins free to control relays and such and it's a noisy environment then I think the DS18 is a better choice. If you just want to look at and log temperatures then it doesn't get much easier than using the LM34. Hell, you could just use a volt meter if you wanted too.

One other thing to consider is speed. In parasite mode the DS requires ~750ms to make the conversion and write to the Scratchpad if you are running it at 12 bit resolution. If you wanted to poll several sensors in order you could have several seconds delay in your code. When I was using the LM34 I was able to read and store readings from 6 sensors in 2 lines of code with no delays. Using the DS18's I needed to address them individually and drop the resolution to 9 bit to minimize the 750ms delay (which I'm running at 100ms @ 9bit without errors.)

None of this should deter someone from using the DS18 other than if they are just learning to code the DS is more challenging to implement. As a noob to coding when I first started my project and all I read was "use the DS18B20" and then I looked at the tech sheets on using it, it was very intimidating to me. I learned a ton playing with the LM34.

They're both fantastic sensors.

 

[freddyb]

Well, it might not be fair to say one is better than the other, but I'll say it anyway...DS18 is better than LM34!

googlefight gots my back

 

[Brewers Hardware]

Well, it might not be fair to say one is better than the other, but I'll say it anyway...DS18 is better than LM34!

googlefight gots my back

LOL! Oh yeah!?! Well I'm rubber and you're glue!....

 

[freddyb]

Ahem, Derrin, "The customer is always right."

I Love the SS thermowells and probe ends that arrived last week. You do great work.

Speaking of rubber and glue, what's your preferred method of mounting a DS18 inside of one of these bad boys?

 

[Brewers Hardware]

Ahem, Derrin, "The customer is always right."

I Love the SS thermowells and probe ends that arrived last week. You do great work.

Speaking of rubber and glue, what's your preferred method of mounting a DS18 inside of one of these bad boys?

Oh well in that case who needs the LM34? lol.

I've tried silicone but found the silicone I was using didn't like to stick to SS. So I roughed up the interior of the tube with a .22 caliber SS bore brush and that seems to have worked better. I didn't know I had an issue until I tried to remove a probe and the compression fitting was still tight. The cable and sensor all slid right out with a perfect silicone impression of the interior of the probe end.

I have swage being made with the intention of selling the probes with either a DS18 or LM34/35 already installed with bare-leaded cable and then swaging the tube to the cable... but that doesn't do you much good at this point.

Try roughing the interior and epoxying or siliconing the cable in. Rough the cable jacket with sand paper as well.

 

[BrewBeemer]

Oh well in that case who needs the LM34? lol.

I've tried silicone but found the silicone I was using didn't like to stick to SS. So I roughed up the interior of the tube with a .22 caliber SS bore brush and that seems to have worked better. I didn't know I had an issue until I tried to remove a probe and the compression fitting was still tight. The cable and sensor all slid right out with a perfect silicone impression of the interior of the probe end.

I have swage being made with the intention of selling the probes with either a DS18 or LM34/35 already installed with bare-leaded cable and then swaging the tube to the cable... but that doesn't do you much good at this point.

Try roughing the interior and epoxying or siliconing the cable in. Rough the cable jacket with sand paper as well.

Derrin; I bet there is a oil on the surface inside the SS tubing from when they were rolled and welded. Silicone will stick to any clean surface even before it has been sanded for a better bite. The silicone will release acid which might not be too good on the temp probe. On your tubing / probe fitting problem I would look into the MSC catalog for a small reamer and run it thru from both ends to prevent any recalls with probes not fitting because of the weld seam flashing. heck you can sell 100 probes and only have two that not fit and fail, see which one your remembered by remembered the longest time?

On adding a "swage" would that prevent pulling the probe out and replacing it if needed? The probe is far cheaper than the stainless tubing with the end cap added plus Tig welded.

On the LM 34 that will not work with the BCS-460 control unit, it would need a different probe that is a lot smaller than the stainless steel probe body.

 

[Homercidal]

The DS18 is a great chip no doubt, but the poster was asking about the LM34 specifically. I'm using DS18's in order to free up pins as well. One other nice thing about the LM34 is that it's already calibrated in *F. As Bull pointed out on page one when I first suggested the DS; with .5*C accuracy, making the conversion to *F results in 1.8* accuracy where as the LM34 will be 1*. On the downside the LM34 will for sure need an RC circuit to clean the output and the DS18 will not. Bull found a great cap and resistor that can be fit in the probe body.

I don't think it's fair to say one chip is better than the other. I think it depends on the application. If you need pins free to control relays and such and it's a noisy environment then I think the DS18 is a better choice. If you just want to look at and log temperatures then it doesn't get much easier than using the LM34. Hell, you could just use a volt meter if you wanted too.

One other thing to consider is speed. In parasite mode the DS requires ~750ms to make the conversion and write to the Scratchpad if you are running it at 12 bit resolution. If you wanted to poll several sensors in order you could have several seconds delay in your code. When I was using the LM34 I was able to read and store readings from 6 sensors in 2 lines of code with no delays. Using the DS18's I needed to address them individually and drop the resolution to 9 bit to minimize the 750ms delay (which I'm running at 100ms @ 9bit without errors.)

None of this should deter someone from using the DS18 other than if they are just learning to code the DS is more challenging to implement. As a noob to coding when I first started my project and all I read was "use the DS18B20" and then I looked at the tech sheets on using it, it was very intimidating to me. I learned a ton playing with the LM34.

They're both fantastic sensors.

Yeah, I am not fantastically skilled with programming etc. I just want cheap, simple and effective. I was under the impression that the LM34 could utilize a free application for readings and logging temps from the parallel port. But maybe confused with the other one.

Anyway, I don't mind doing a conversion from C to F if need be in order to make it as simple to set up as possible. If it gets too costly and complicated, I may decide it's not worth the hassle. It's not like I need to do this to make great beer. I am just curious to see what temp fluctuations happen in the wort while it's fermenting. Or I could just set up my cooler/heater and look at the thermometer once in a while, but then I wouldn't have a cool graph showing the temp that I can show off to impress people and stash away for future reference if need be.

 

[beerthirty]

Derrin, you mentioned earlier about silver solder flux being corrosive. Yes it is a muriatic acid based flux. I had the same problem with with concrete.
To seal my temp probes I placed a very small zip tie on the wire near the probe, then shrink wrapped the ziptie and probe as in the OPs procedure. The zip tie prevents the wire from moving in the shrink wrap. So far(crosses fingers) the shrink wrap has not slid on the probe either.

 

[pompeiisneaks]

Ooh Beerthirty! that's a great idea, i need to do that... thanks for the pointer

 

[bull8042]

Another thing you can do is find some "adhesive-lined" shrink tubing. That is what I am using and you can't hardly cut that stuff off. Panduit makes it, but there are several manufacturers out there as well.

 

[beerthirty]

I have never seen it but I will look. My major concern with the shrink tubing was to get it sized so when the probe got hot, the tubing wouldn't try to shrink more and split on the probe. That is why I resorted to the zip tie, to make the wire closer in size to the probe. The adhesive tubing would probably solve that issue.