How to measure water temperature?

This just came to mind, as I was thinking about measuring water temperature.

I have a temperature sensor (one of the long, flat, ones used for measuring GPU, socket, etc. temperatures) hooked up to a little screen which shows the temperature being read.

I was planning on running it down my T-line, so that it was positioned right at the top of the upwards facing branch of the 'T' fitting. The top of the T-line would be sealed air tight with a Vinyl test plug (you turn a thumbscrew at the top, and it enlarges an O-Ring seal a very great deal), which would allow for the very thin cable being run past the plug.

Potential problems that have come up:

Is this type of sensor accurate for measuring temperatures of water? I can try and do a primitive test on it, but am also worried about the following,

The (I'm assuming it's soldered) joint from the actual sensor (which has a thin plastic film over the probe, and any wiring) to the cable which runs to the screen is heatshrinked, but probably isn't waterproof, so wouldn't live voltage from the sensor cable be running into the water, shorting the sensor, and causing other bad things?

Can anyone think of any means by which I could waterproof the sensor, in a manner which wouldn't compromise the water flowing through my water circuit?

Explanatory Diagram: (ignore the periods, they are spacers)

T-Line:

Sensor cable
............|
............|
...........\/

.......|...|.|
.......|...|.|
.......|...|.|
.......|...|.|
.......|...|.|
.......|...|.|
.......|...|.|
.......|...|.|
.......|...|.|
.......|...|.|
.......|...|.|
.......|...|.|
-----|..+.|-----
...<----Water flow.....
----------------

+ = sensor head

One thing to note is that the T-Line is composed of 3/4" ID tubing, so the actual line is very wide.

I have thought that the T line is a good place to measure temperatures, because it is the easiest sealable opening to run a probe down, and is at the end of the circuit where the water is theoretically hottest.

Any comments or suggestions are appreciated!

Does anyone have answers to my questions? Noone?

If anyone else is measuring their water temperatures, how did you do it?

well i had the same issue as you, i wan tto know what my water temp in and out of my radiator was. THe answere i got was there are no good way to do it. Adding the probe only reduces flow rate, so i want to excited about that. On my older "ghetto" WC setup, i bought on of those cheap stick on temperatures that they have for fish tanks. Yes it was useless as I didnt feel like shoving my head in my case to see the watertemps.

Flow rate isn't a problem for me, and in any case the sensor would be right at the top of the upper T branch - out of the flow path of water entering the pump (hopefully ).

I was thinking - what if I made a special fitting, something like this:

(ignore the periods, they are spacers)

.....\........\
........\........\ <--------- Upper Branch angled sharply
...........\........\ <------- so water won't flow up (easily)
..............\........\
------------\........\----------
..Water Flow ------->
--------------------------------

And ran a probe down the upper branch? The flow would be going opposite the upper branch's angle, and very little water would be pushed up the branch, allowing for a very short tube line atop the branch, and no risk of air catching (run it upside down?) long term. The only issue being waterproofing, and getting the probe in ther safely, without risking leaks.

I played around with my thumb-screw plug, and although it expands a great deal, forming a seal, the cord of the probe still leaves a pocket for air to escape, meaning I need another method of getting the probe into the water.

I read that some people use laser temperature readers - but I have no access to such a thing, and want along term soloution

How about just putting a copper fitting somewhere in the system and then fixing your temp probe to the outside of the copper. The copper fitting should be close to the actual water temp that way your probe temp should be pretty accurate.

I have an inline water temp probe I made that works really well. Basically I took a 1/2" ID coupling, drilled a small hole close to the middle on one side of it, and routed out a channel for the probe and wire. Then I sealed it with some silicone. The tubing has to stretch over the fitting and wire, but I can't get it off once it is on. VERY TIGHT!



BTY, the water comes out of this hose and into the rad. This way I have the hot temp directly from the cpu!

In practice it worked great, but it gives me serious concern about the quality of the onboard temp monitoring abilities of this NF-7. I know they are really off anyway, but when I was testing out this thermocouple, my room temp was 25C, cpu temp was 40C but the water temp was only 27C. I don't know if this is saying my waterblock sucks (Maze4) or that the onboard thermal monitoring is just THAT bad.

What do you all think?

Hey that's a good idea. Better than my copper idea above.
You can check to see how off your on-board thermal monitoring is by taking the probe and putting it in some water along with a proper thermometer and see how big the difference is. I did this with my on-board thermal probe and there was only a 0.6 degree difference between the two readings (probe versus thermometer) with temps ranging from 11 degrees all the way up to 47 degrees.

I use a 40-channel temperature monitor =p with variable sensor logging ^_^. It's a sort-of work in progress, but... I have sensors for...

CPU t°, CPU water t°, NB t°, NB Water t°, GPU t°, GPU Water t°, pre-rad t°, post-rad t°, pre-chiller t°, post-chiller t°, pre-pump t°, post-pump t°, 5 different sensors for case t°, one sensor that's 6 feet from my system for a nice 'ambient room' t°, PSU t°, chiller PSU t°... pretty well losing ideas for what else I can monitor =/. The accuracy is only to the tenths though, a better thermocoupler would give me hundredths.

That's not to say that I'm keeping everything monitored 24/7 and will do so forever... it's just a nice tool to work out a few kinks in the system. For instance, with a pump-in and pump-out monitor, I can see that the 40rz was actually adding nearly 1° to the water (at worst case) over the 20rz (iwaki pumps). The total performance increase from the higher output wasn't even picked up by the sensors, and I wasn't able to break any ground on the overclock. So... I sold the pump ^_^.

Just the same, I was able to mess around with 2 different GPU blocks and found that in my system, the innovatek GPU block was complete donkey-***. It worked GREAT in a previous loop, but switching to a DangerDen block actually reduced the GPU-water t° by almost 1.5°... which was a big surprise because I usually hate DD =/. It's about *now* when I'd like a Cather or Jfettig custom GPU block ^_^. *hint hint*

edit: Oh yeah... the pic:

very nice.

Here is what i use in mine for temp i use a T and when to radio shack and got one of there $10 indoor/outdoor temp moniters, and i cut the outdoor sensor off (about 6' long) and added it to my Doc 5 monitor. The sensor is waterproof. just drill a small hole in T cap add sensor and seal up..

Cool ideas

I think I'll seal a drilled hole in my T cap for this, as it looks safest, and easiest

vonkaar, that's a crazy nice setup there - I can see how it would be useful for working out wrinkles in your system setup

felinusz said:

Cool ideas

I think I'll seal a drilled hole in my T cap for this, as it looks safest, and easiest

vonkaar, that's a crazy nice setup there - I can see how it would be useful for working out wrinkles in your system setup

Click to expand...

Yeah, leave it to Vonkaar to use an earth mover where a shovel would do.

You will get the best sensor performance by placing the sensor in the flow stream of the water. Having it in the T will give you slow temp change response and could actually never give you an accurate reading. Water in the T will be essentially dormant and will only heat up or cool down by convection.

depends on where you put it. If it is pretty close to the flow, it should work. If the temp probe is near the top of the t line, then you are correct.

DTek sells these. They would do exactly what you want and can be hooked up to a digital readout.

That's pretty expensive for $3 in parts =/

true dat

JimmyG said:

You will get the best sensor performance by placing the sensor in the flow stream of the water. Having it in the T will give you slow temp change response and could actually never give you an accurate reading. Water in the T will be essentially dormant and will only heat up or cool down by convection.

Click to expand...

My interpretation of his first post was that he was going to feed the thermistor all the way down into the stream.

Yep, thats still the game plan

I'm also going to play with making a fitting with a probe inside - although I'm worried about leaking due to the very high pressure that will be in my system. I have several spare fittings to play with, so I'm sure I'll come up with something for long term use that's better than using my T line.

felinusz said:

Yep, thats still the game plan

I'm also going to play with making a fitting with a probe inside - although I'm worried about leaking due to the very high pressure that will be in my system. I have several spare fittings to play with, so I'm sure I'll come up with something for long term use that's better than using my T line.

Click to expand...

Trust me, talking about "very high" pressures with these aquarium pumps is just silly