best sensors...

Discussion of the physical aspects
clearwaterbrewer
Posts: 383
Joined: Wed Feb 09, 2011 3:43 pm
Bot?: No
Location: Clearwater, FL
Contact:

best sensors...

Post by clearwaterbrewer »

I have looked a LOT, and have a bunch of ideas for the best sensors, I think they should be very small, have a quick connect, be sturdy, and waterproof... not asking for much, huh?

the closest brewers hardware has is their stainless compression fitting with a huge 1/2" NPT fitting (that ECC has also)
Image

and the probe with M8 fitting
Image


the above may work with an 1/8" NPT instead of the 1/2", but 1/8" NPT is still big for using in lines, fittings, and valves...


here are RTD ones Auber sells that are pretty close, but a bit big(and RTD, so we cannot use, but that little block that the connector screws to and the sensor housing is somehow connected to is really interesting):
http://www.auberins.com/index.php?main_ ... ucts_id=84
Image

I ordered a couple of these stainless 1.25" long 1/8" npt thermistor probes for $19.17
http://search.digikey.com/scripts/DkSea ... 15-1084-ND
Image

and a couple of these 1/4" smooth stainless ones for $4.33
http://search.digikey.com/scripts/DkSea ... 15-1081-ND
Image

here is a small one for $2.20, I think it is still 1/4" diameter:
Image
http://www.temcocontrols.com/product.ph ... egoryID=13

US Sensor has some other nice ones, but they are special order, $250 minimum..
if they put a o-ring groove in this one, it would be really close..
H3432 S/S Hex 10/32 Thread
Image

H3134 S/S Set Screw 8/32
Image


What I also would really like is basically a tiny weldless thermowell to glue or whatever a tiny bead thermistor in..., how hard would it be to have a 1/4", 5/16", 6mm, or 8mm stainless bolt machined to be nearly hollow, have an o-ring groove in underside of the head for sealing and ability to really tighen, and all but a bit of the threads machined off? (that would not handle the quick disconnect part, but close...) I can put a 6mm threaded hole in a LOT of different places without adding a bunch of complexity, such as in a valve... I wish I had one of those little lathes...

thoughts?
User avatar
mpilchuk
Posts: 45
Joined: Sun Jan 04, 2009 7:31 pm

Re: best sensors...

Post by mpilchuk »

Clearwaterbrewer,

Look what I discovered.

Nice housing, just need the right type of probe in it.
Attachments
IMG_2690.JPG
IMG_2690.JPG (14.97 KiB) Viewed 12799 times
IMG_2689.JPG
IMG_2689.JPG (43.88 KiB) Viewed 12799 times
IMG_2688.JPG
IMG_2688.JPG (39.9 KiB) Viewed 12799 times
User avatar
mpilchuk
Posts: 45
Joined: Sun Jan 04, 2009 7:31 pm

Re: best sensors...

Post by mpilchuk »

I picked up two 10k NTC thermistors for 9.50 off ebay and I have installed one in this housing. I now need the proper cable to test it out. I may also put a conductive jel in the channel the thermistor is in (it didn't seem to have any in it when I took it apart). The device I installed into it is ATH10K1R25
Last edited by mpilchuk on Thu Jul 12, 2012 12:41 pm, edited 1 time in total.
User avatar
mpilchuk
Posts: 45
Joined: Sun Jan 04, 2009 7:31 pm

Re: best sensors...

Post by mpilchuk »

I have tested the above modification and the temperature readings are ok, off by a degree or two of my fluke digital thermometer. I tried to calibrate the sensor with the data sheet from the manufacturer, but I can't seem to run the resistance numbers through the online calculator and get anything logical that gives me a calibrated temp probe. In fact the temp range goes off the scale. I have to plug back in the default setting to get it to work. I use the resistance of 32.738, 5.3302 and 1.7610 for the temps 32,104, 158 F and I get .0027724567, 2.5058596E-4 and 3.3779412E-7 these bury my temp dial to the right.
User avatar
bbrally
Posts: 210
Joined: Sat Mar 27, 2010 3:59 am
Bot?: No
Location: Vancouver, BC
Contact:

Re: best sensors...

Post by bbrally »

The resistance values you are using are in KΩ. Multiply the values by a thousand and I think you'll get the right coefficients.

The thermistors from Analog technologies look alot nicer than the ones I've bought on Ebay from China. I'm thinking of getting some and trying them with these M8 pico connectors (RSMEDGN 3). http://lumberg-automationusa.com/produc ... eceptacles

They seem like they might be small enough to glue into the end of a standard 1/4" probe. You'd then have a really compact detachable probe with standard thermowell dimensions.

For extra thermal conductivity I thought I'd try a dab of Arctic Alumina Thermal Adhesive (http://www.arcticsilver.com/arctic_alum ... hesive.htm) to the end of the probe to keep the sensor in contact with the probe walls rather than filling the probe with a conductive paste/gel which would then have a greater thermal mass and dampen any temp changes.
User avatar
mpilchuk
Posts: 45
Joined: Sun Jan 04, 2009 7:31 pm

Re: best sensors...

Post by mpilchuk »

Thanks, I did it with x1000 and my numbers are
A-.0011038767
B..3822582E-4
C. 7.159652E-8

This shows up as
A. 0.0011038767
B.0.0002382256
C. 7.14e-8

A and B seem like they are in the ballpark but C is way out there.

I just re-ran the number and got entirely different numbers!

0.047627714
-0.0072069354
2.755162E-5

Man, something is screwy.

I did get the rest of the items I ordered and tested the probe in my RIMS tube and it works well. Nice metal braid cable and the quick disconnects on each end along with the female to mount in my box. 40ish bucks but I have a number of extra thermistors so when it goes flakey, I can fix it.

The ATH10K1R25 I got were from Analog Technologies as well. Damn small critters they are.
User avatar
bbrally
Posts: 210
Joined: Sat Mar 27, 2010 3:59 am
Bot?: No
Location: Vancouver, BC
Contact:

Re: best sensors...

Post by bbrally »

I ran the numbers from the analog tech datasheet (http://www.analogtechnologies.com/docum ... 0K1R25.pdf) for resistances at 0c, 40c, and 70c and got the following results:

0c = 32,728Ω
40c = 5,330Ω
70c = 1,761Ω

from http://www.capgo.com/Resources/Temperat ... rCalc.html:

A = 0.0011038767
B = 2.3822582E-4
C = 7.159652E-8

So your first set of numbers look fine to me. I've seen C values much higher and lower than this example in the past.

The Steinhart equations don't compute a prefect match to the thermistor, only a close one. It will only be truly accurate for the three temperatures used in the calculation, thus using resistance values from different temperature points will provide different coefficients.

I find the best way to be sure the coefficients are as close as possible to the actual temps are to calibrate the thermistors yourself as there are also manufacturing tolerances in the thermistors that would give differing temps between thermistors. Ideally I think you'd like the three temps that you calibrate from to be the temps you are most interested in being accurate, like your mash temp (0c, 65c, 100c).

http://www.embeddedcontrolconcepts.com/ ... =10&t=1121
User avatar
mpilchuk
Posts: 45
Joined: Sun Jan 04, 2009 7:31 pm

Re: best sensors...

Post by mpilchuk »

bbrally,

I got the same numbers. When I plugged them in the other day, that temp probe went off the charts. Today, it seems to be working. I just ran a test with some water and there is about a degree difference when at 90F. I'll soon use your other thread to get the numbers closer

Thanks.

Good stuff to try.

I will be doing this as soon as time permits.
User avatar
bbrally
Posts: 210
Joined: Sat Mar 27, 2010 3:59 am
Bot?: No
Location: Vancouver, BC
Contact:

Re: best sensors...

Post by bbrally »

I've been wanting to make a new temp probe for a while now and have finally got all the pieces together to finish it.

Here's the parts list:

4" 316 probe from Brewers Hardware ($3.75)
10K Thermistor PN:MF51E103F3950 from Digi-Key ($2.22)
M8 Pico receptacle PN:RSMED3 from Sandtron Automation ($3.52)
M8 Pico connector PN:RKMCW3 from Sandtron Automation ($16.37)
Arctic Alumina thermal adhesive from ebay ($9.99)

The probe's inside diameter is slightly too small for the receptacle, so I filed the threads off.
The probe has a slight inward taper, so I flared the end a bit with a tapered punch and a small hammer.
I used a small amount of thermal adhesive to bond the thermistor bead to the end of the probe, and some epoxy to bond the probe to the receptacle and waterproof the probe.

The result is a very clean inexpensive probe. The wire and connector are the expensive part. A straight connector would have cut the cost a bit, but I liked the 90 degreee connector.
The parts before modification and assembly
The parts before modification and assembly
IMG_0818.JPG (87.55 KiB) Viewed 12477 times
Heat shrink on the contacts and a twist in the wire to provide the proper length
Heat shrink on the contacts and a twist in the wire to provide the proper length
IMG_0831.JPG (51.47 KiB) Viewed 12477 times
Completed Probe
Completed Probe
IMG_0834.JPG (61.75 KiB) Viewed 12477 times
JonW
Site Admin
Posts: 1726
Joined: Sun Jul 18, 2010 7:51 am
Bot?: No
Location: Huntington Beach, CA
Contact:

Re: best sensors...

Post by JonW »

Very nice job. This is simply the best temp probe setup I've seen.
Post Reply