# Stiffening of thermistors

#### Thomas Holm

Joined Jul 31, 2018
8
Hello,

#### Attachments

• 61.6 KB Views: 12

Joined Jan 15, 2015
6,204
Something you can try is either a rubber dip or a Plasti Dip, I have used it to coat some very small thermocouple junctions. I don't know your location but here in the US you can find the Plasti Dip at Lowes or Home Depot, here is an example from Lowes:
I have also used a rubber dip I bought from McMaster Carr Supply. The more or less dips control how rigid your wires will become. Cost about $7.00 USD to see if it will work for you. Wipe off with isopropanol alcohol after using them. Ron Thread Starter #### Thomas Holm Joined Jul 31, 2018 8 Something you can try is either a rubber dip or a Plasti Dip, I have used it to coat some very small thermocouple junctions. I don't know your location but here in the US you can find the Plasti Dip at Lowes or Home Depot, here is an example from Lowes: I have also used a rubber dip I bought from McMaster Carr Supply. The more or less dips control how rigid your wires will become. Cost about$7.00 USD to see if it will work for you. Wipe off with isopropanol alcohol after using them.

Ron
Thanks for your suggestion - I'm located in Chicago downtown. The dip you suggest looks interesting. The product information says that the product "Remains flexible and stretchy". Perhaps combined with some kind of rigid rod attached to the lead wires, it would become stiff enough. Again my concern is that I do not want to exceed the 2mm diameter. Secondly I need to have a contact area where I can apply tape to secure the probe. So would probably have to glue a steel sleeve to the wire also. - I don't know if I'm explaining this correctly.

Joined Jan 15, 2015
6,204
I get what you mean. Yes, the rubber and plastic coating stuff remains pretty flexible but no much so that it can't be inserted. It also cleans pretty well as I mentioned, swab with some isopropanol. It may be worth a try. While I never tried doing it the stuff can likely be thinned down if necessary with maybe some acetone when applied. Maybe you have seen the rubber coatings on basic hand tools? Figure for seven bucks it will give you something to look at and see if it does what you want and need.

Truth be known the thermocouple picture you referenced looks to be a relatively common Type T thermocouple. They are actually pretty easy to make. The problem comes in when they say "for veterinary use". Then the base price is multiplied X 20.

There are some inexpensive data loggers out there which will record temperature over a period of time, including for use with thermistors. Matter of if you want single or multi-channel; and how many data channels. Simple stand alone units for single channel are not overly expensive.

Ron

#### ebp

Joined Feb 8, 2018
2,332
What do you need in terms of biocompatibility of the materials? Most of the materials I've seen mentioned so far will be somewhere between really really bad and abominable.
Do the probes need to withstand sterilization? If so, by what method? Absorption/adsorption of and general compatibility with chemical sanitizers/disinfectants/sterilants may be a significant concern. You certainly won't be able to toss the stuff mentioned so far into an autoclave except for sterilization prior to disposal.
What is your target for unit to unit tolerance? Standard tolerance for most thermistors is sloppy, making them non-interchangable. If you don't mind having to calibrate each individual probe and always use it on the same channel (because that's where the corresponding cal resides), this may not be an issue.

Thermocouples are generally a better choice for small probes and unit-to-unit consistency. They do require good conditioning circuitry.

#### Thomas Holm

Joined Jul 31, 2018
8
I get what you mean. Yes, the rubber and plastic coating stuff remains pretty flexible but no much so that it can't be inserted. It also cleans pretty well as I mentioned, swab with some isopropanol. It may be worth a try. While I never tried doing it the stuff can likely be thinned down if necessary with maybe some acetone when applied. Maybe you have seen the rubber coatings on basic hand tools? Figure for seven bucks it will give you something to look at and see if it does what you want and need.

Truth be known the thermocouple picture you referenced looks to be a relatively common Type T thermocouple. They are actually pretty easy to make. The problem comes in when they say "for veterinary use". Then the base price is multiplied X 20.

There are some inexpensive data loggers out there which will record temperature over a period of time, including for use with thermistors. Matter of if you want single or multi-channel; and how many data channels. Simple stand alone units for single channel are not overly expensive.

Ron
I get what you mean. Yes, the rubber and plastic coating stuff remains pretty flexible but no much so that it can't be inserted. It also cleans pretty well as I mentioned, swab with some isopropanol. It may be worth a try. While I never tried doing it the stuff can likely be thinned down if necessary with maybe some acetone when applied. Maybe you have seen the rubber coatings on basic hand tools? Figure for seven bucks it will give you something to look at and see if it does what you want and need.

Truth be known the thermocouple picture you referenced looks to be a relatively common Type T thermocouple. They are actually pretty easy to make. The problem comes in when they say "for veterinary use". Then the base price is multiplied X 20.

There are some inexpensive data loggers out there which will record temperature over a period of time, including for use with thermistors. Matter of if you want single or multi-channel; and how many data channels. Simple stand alone units for single channel are not overly expensive.

Ron
Ron, yes I believe that the ones we have are Type T thermocouplers. As I understand, for those to work, I would need to have amplifiers installed to read them whereas thermistors are pretty straight forward. I already have made the PCB to record from 8 probes onto a SD card. So the only thing that I'm struggling with is the probe design...

#### Thomas Holm

Joined Jul 31, 2018
8
SMD thermistors are available in 0201 package, 0.6 x 0.03mm, US\$11.35/100pcs.

https://www.digikey.com/product-det...america/NCP03XH103J05RL/490-4644-1-ND/1203985

Create your own probe with fine stainless-steel tubing and seal the SMD at the tip with epoxy.

MrChips,

As I read the datasheet, the SMDs measure 0.3x0.6mm... I'm not sure how I would be able to attach any wires to such a small component. Could you perhaps send me some suggestions on how you'd approach this?

Thank you

#### Thomas Holm

Joined Jul 31, 2018
8
What do you need in terms of biocompatibility of the materials? Most of the materials I've seen mentioned so far will be somewhere between really really bad and abominable.
Do the probes need to withstand sterilization? If so, by what method? Absorption/adsorption of and general compatibility with chemical sanitizers/disinfectants/sterilants may be a significant concern. You certainly won't be able to toss the stuff mentioned so far into an autoclave except for sterilization prior to disposal.
What is your target for unit to unit tolerance? Standard tolerance for most thermistors is sloppy, making them non-interchangable. If you don't mind having to calibrate each individual probe and always use it on the same channel (because that's where the corresponding cal resides), this may not be an issue.

Thermocouples are generally a better choice for small probes and unit-to-unit consistency. They do require good conditioning circuitry.
epb,

I completely agree, that my probe design needs improvement. My main priorities are 1. size: should be no more than 2mm in diameter, 2. easy to clean: I would prefer to have the probe completely enclosed, so that I can wipe it with e.g. ethanol. it does not need to be autoclaved. 3. for my current setup, I would prefer to stick to thermistors. I already have set up my "box" to calibrate the probes that I've built so far and get approx 0.1C deviation between the 8 of them, hooked up to individual channels. 4. the probes need to be rigid enough to withstand daily use and have a surface that I can secure with tape... does that answer your questions and do you think it possible to find a solution?

Thank you,

Thomas

#### MrChips

Joined Oct 2, 2009
24,612
MrChips,

As I read the datasheet, the SMDs measure 0.3x0.6mm... I'm not sure how I would be able to attach any wires to such a small component. Could you perhaps send me some suggestions on how you'd approach this?

Thank you
The main difficulty would be how to insulate the SMD from the walls of the SS tube.
I would align the SMD length-wise and solder enameled copper wire to the two ends of the SMD.
Then I would cover the SMD with thermally conducting epoxy such as Wakefield Deltabond 155.

#### ebp

Joined Feb 8, 2018
2,332
You might try experimenting with fluoropolymer heatshrinkable tubing. "Kynar" is a brand name for one such material. Though it isn't rigid, it is quite stiff, depending on how much it shrinks. You could include a stiffening device (e.g. some sort of high-tech toothpick) inside the tubing.

I haven't used this type of heatshrink much and it has been some time since I did anything with it. I can't remember if it can be effectively heat-sealed by simply getting it hot enough and squeezing the end closed (something like small smooth-jawed surgical pliers works pretty well with other types). Careful trimming may be necessary to shape the flattened end.

Regular heatshrink tubing is mostly "polyolefin" (same class as polyethylene, polypropylene & others). It tends to be fairly flexible in the common thin-walled form, but again a high tech toothpick might do the trick. It is very easily heatsealed as described above. It is resistant to a wide range of solvents, so agents such as ethanol, propanol, quaternary ammonium compounds, or chlorhexidine won't harm it, though the latter two do tend to adsorb to it to some extent - which is either good or bad.

Of course the drawback to any polymer covering is that they all are thermally insulating to some extent, resulting in time lag and temperature error when heat is put in on one side and free to exit down the wire or on the other side.

Stainless steel tubing can be closed by TIG welding, but that means finding someone who can do very fine precision work. As someone else mentioned, "hypodermic tubing" is readily available are fairly low cost. Alternatively, there are (or were - haven't looked for some time) 14 gauge 3 inch disposable needles in the veterinary product line from Monoject (Medtronic), though the diameter may be too small for your thermistors. The trick with steel tubing will be to get the connections to the thermistor electrically insulated so the tubing doesn't short them.

#### ebp

Joined Feb 8, 2018
2,332
I once made thermistor probes for measuring the temperature of the cross-linked polystyrene "wedges" on which ultrasonic transducers were mounted (pipeline weld inspection).

I used polyurethane tubing. The tubing was boiled in water to take out any distortion from being on a reel. After allowing drying time, an inch or so from the end was thinned by turning it while lightly holding it against an abrasive wheel about 15 mm in OD in a rotary tool ("Dremel" type) - took a bit of practice to get the method down. I cut the tubing in the middle of the thinned part. Extension leads of fine stranded insulated wire were soldered to the the thermistor. I think I used bits of Kapton (polyamide) tape to insulated the leads from each other, but I can't remember for sure - this was 15 years ago. With a "piano wire" hook and using 99% isopropanol for a lubricant of sorts, I pulled the wire through the tubing (these were about 12 inch lengths) until the thermistor body came to rest at the thinned end of the tubing. The thermistor was then carefully pushed in just far enough to that the widest part aligned with the end of the tubing. In quite a short time the PU tubing developed a very firm grip on the thermistor. Thinning the tubing wall meant the end of the probe was no wider than the rest of the tubing. A bit of slack was left in the wires inside the tubing so they wouldn't pull on anything when the tubing was flexed. Time was allowed to be sure the IPA had evaporated completely before fitting a connector.

The probes stood up quite well with constant flooding with water which is used to couple the ultrasonic energy from the probe to the pipe and back.

PU tubing is moderately stiff and incredibly durable but doesn't have great chemical resistance.. I used 1/8" OD tubing which is easy to get. I have no idea if smaller sizes are easily had.

#### KeepItSimpleStupid

Joined Mar 4, 2014
5,090
Your "T" type selection isn't a bad choice, but I don;t think your going to get the precision your after. Thermisters basically suck. Only the RTD (Resistance Temperature Transducer)

https://www.omega.com/pptst/P-ULTRA_RTD.html

here: https://www.pyromation.com/TechInfo...ect_and_Use_the_Right_Temperature_Sensor.aspx

Type T has limits of error of 1.8 deg, so your already messed up.

Here's https://www.minco.com/products/sense/temperature another blurb about RTD's.
They are experienced in medical applications. http://www.temflexcontrols.com/pdf/SSG01(A)_web.pdf#page=111 has some 3.2 dia probes. It does look like for medical applications, you would need to call them.

Thermocouples have "extension wire" that can be used to extend the reach of the TC. It has the same electrical properties, but costs less. it's not to be used for the TC. You have to use the "T" connector.

Thermocouples are very much dependent on the "cold junction compensation" and most of what I;ve seen generally sucks. Only a Fluke and an Analog Devices block seems to do it right. You have to measure absolutely, the temperature of the terminals where the thermocouple attaches to the instrument. That connection has to be ISOTHERMAL meaning both connections need to be at the same temperature.

A data-logger, multi-sensor device from Measurement computing was the worst I've ever seen. It worked fine for diode and RTD sensors, but forget thermocouples. The block was totally exposed. Any air movement would muck up the temperature.

For TC's the instrument has to measure the voltage. Then it has to know the temperature of the terminals and the block has to be isothermal. Then the voltage for that temperature has to be added/subtracted (forget which) and then looked up by lookup table or polynomial. These tables would be the reference junction at 0C tables.

So, thermocouples will typically give you the smallest size. repeatability is good, but it's no where the accuracy you want.

RTD's are great if you can use them. I used them from 200C to -80C. The leads are usually tough to solder and the reccomend spot welding. The measuring system is a basically a wheatstone bridge.

The usual connection consists of three equal length and wire guage wires. Why 3? You can save on cost by assuming that the drop in the excitation wire is dropped in the sensing wire. You can also use 4-wire sensing.
So, you measure the voltage drop in the excitation wire and assume the same drop in the 3rd wire.

Diode sensors are usually used for cryogenics.

Thishttps://www.lakeshore.com/products/Cryogenic-Temperature-Sensors/Platinum-RTDs/Models/pages/Specifications.aspx PT-103 PT100 RTD sensor has a 1.6 mm diameter.

You might want to look at PTFE (Teflon) or Silicone insulation on the leads.

Your caught between a rock and a hard place, Cold junction compensation has to be really good. If the TC wire doesn't come from the same batch, I would think you would need a way of normalizing them. e.g. Place in a bath and note the deviations.

LabVIEW is the defacto standard for instrumentation automation. https://store.digilentinc.com/labview-home-bundle/ The commercial package is a lot more expensive. Educational pricing is nice too.
The programming is graphical.