ds18b20 used for fluid?

I have a van that I'd like to make a temperature gauge for the transmission fluid temp. Could I use a DS18B20 sensor and electrical tape it to the metal line?(I would probably solder the wires to the leads of the sensor and then hot glue around the exposed leads/wires to insulate them)
Or is there a better type of sensor to use in this type of application? I was going to use 3 - 7 segment displays in the dash area to display the temp.
 

BeanieBots

Moderator
As far as protecting the DS18B20 from the elements are concerned, there are quite a few options. The simplest, is simply make your connections and then dip it resin.
The method I use to make them very robust is to slip one into a short length of stainless tube.
Cut a length of tube, plug one end with two part appoxy, coat sensor with silicon rubber and slip into tube. Once that has set, slip some silicon tubing over the leads and into the stainless tube securing with more two part appoxy.
I've made several of these which have been immersed under several meters of water for many years with no ill affects.
What sort of temperature range are you looking for? From memory, I think it only goes up to 85C.
 
The datasheet for the DS18B20 states a max temp of 125C, (of course me, being in the USA will be converting that to Fahrenheit - which would be a max temp of 257) Any temp above 200 Fahrenheit will start shortening the life of the transmission, so the temp range of the sensor is not a problem.
I was more concerned with the mounting of the sensor, if that would give a good indication of the temp of the fluid in the line.
 

Michael 2727

Senior Member
Like BeanieB suggested mount the DS18B20 inside
a small metal tube. If you can't find a
stainless tube (some pen ink refils use them)
you could try a piece of telescopic car antenna
they are brass or S/Steel and with a hot iron
you could solder one end to plug it up.
Probably a good idea to clamp a sample piece
of the hookup wire you intend to use to the
line first and see that it won't melt from
the heat also.
 

BeanieBots

Moderator
You're quite right, they will go up to 125C. 85C is the limit for 0.5C accuracy.
To increase accuracy measuring the fluid temp, use heat transfer paste between the "ruggedised" sensor and the pipe. Then wrap heat insulating material around the whole area going about an inch past each side of the sensor. That will help stop heat loss causing a temperature gradient between the fluid and the sensor.
 

bobrayner

Member
Stainless is a fairly poor heat conductor so the lag between fluid and sensor temperature may be too great.
cheers BobR
 

ylp88

Senior Member
Is stainless steel such a bad conductor of heat? One would assume that it would be quite a good conductor of heat, akin to it's relatively good conduction of electricity. Not as good as something like gold but that would be out of the budget, surely!

I have measued the temperature of a liquid (water) though a square aluminium tube by using a LM92 (analog) sensor attached to the tube using thermal compound and some epoxy. I don't know the "response time" but the reading changed almost as soon as liquid flowed past.

<b><i>ylp88 </b> </i>
 

Dippy

Moderator
I thought it might be a brand/trade-name for something fancy.
Whilst looking at epoxy adhesive (and 'potting' compounds) I noticed that some have a poor high temp performance (e.g. 65oC) whilst others can get pretty durned hot before breakdown. It was just something to keep an eye out for.
 

rfs

Member
Dippy.
A genuine thanks for the reminder.
I was about to pot a ds18b20 in a 7mm brass pen tube.
Araldite (r) precision =65 C
Z-POXY (tm) states 80 C

Could I use Servisol (r) Silicone?
It states 230 C on the tube.

It's not for engines, but just a general-purpose thermometer.

Rob S
 

Dippy

Moderator
Have a look at Permabond and Devcon epoxy products. Some of those seem to have a higher temp performance (claimed!!).

Also, do a search on 'potting compounds' or 'encapsulents' as these might be an option.

Note that any of these things will reduce the thermal conductivity between your metal tube and sensor. This will lead to a non-accurate reading of the thing you are trying to measure AND a lag (hysteresis) too. Some potting compounds have better thermal conductivity than others - I'm afraid its a matter of looking round and reading the data sheets.

Will thermal expansion of a 'hard' encapsulent damage your tube or the sensor? Should you try a 'flexible' compound? Maybe.

Silicon? Dunno. But I'd guess that any (one part) RTV type silicone would take yonks to cure fully inside a tube.

R&amp;D can be quite expensive eh? Good luck.

(Perhaps someone has tried all these things and can give you an answer straight away.)
 

BeanieBots

Moderator
The two part &quot;Epoxy&quot; is purely for mechanically sealing the end of the tube in a stronger manner than silicone. It does NOT go between the sensor and the tube. That gap is filled with heatsink compound. The silicone is to prevent water ingress.

Yes, stainless is not the world's best conductor, but the resistance will be small compared to that of the tube carrying the fluid. Aluminium or copper could be used but they would need to be coated to prevent corrosion.
 

ArnieW

Senior Member
Another idea that has worked for me.

I've used copper pipe (standard water pipe) and soldered a cap on the end. This gives a good thermal contact and it is watertight and good for my purposes which is boiling water (and malt and hops ...)

cheers, ARnie
 

bobrayner

Member
&quot;Araldite&quot; is a 2 part epoxy where you have a resin and hardener which remain pretty inert until mixed.
As for the thermal conductivity of stainless I have actually held a piece of stainless pipe while it was welded some 400 mm from my hand. Note Stainless saucepans have a copper (or in the case of cheaper ones aluminium) bases for this reason.
cheers BobR
 

Dippy

Moderator
Thermal Conductivity?
This may help:-
http://www.engineeringtoolbox.com/thermal-conductivity-d_429.html
 
Top