For all you Electronic guru's out there, Temp Probe design
 

Well I am no expert in electronic design by any means but I do know a thing or three about building things from schematics etc.

What I need is a discussion on how to best design a calibrated tempeture probe and what you would need to get parts wise to do this.


Also in this might be a good idea to discuss the best way to get the most accurate calibration of the probe.

I have searched the forums but havent found any real good info on the design of a hand built model.

You'll have to run a search to bring up pHaestus's post, because it's not an active thread, I think. Otherwise you can refer to his great article, with the Maxim AD circuit.

You need to elaborate on what you wanna do with this probe:

1) Integrate to MBM
2) display on separate LCD
3) just dump results via serial or parallel port somewhere
4) what is it you are monitoring (this will determine the size of probes needed)
5) what is the application (this will determine the accuracy required and therefore the cost).

Examples: You can get 4 probes into MBM via thermistors and a gameport on a PC. Expect pretty bad accuracy, but hey you can calibrate them at least and you can fit them most anywhere and it shouldn't cost you much money at all. I keep thinking that the way to improve www.benchtest.com 's design is to add a wheatstone bridge but I haven't done so yet.

You can make a Dallas 1-wire setup and daisy chain a ton of 0.5C accuracy, 0.125C resolution DOW sensors. No calibration required (never hurts tho). These are larger though and not so easy to place in some locations.

You can make a diode-based setup using Maxim ICs. Also possible to use lots of probes and you can even incorporate an AMD CPU's diode reading into the loop. Easy MBM and speedfan incorporation too. Diodes are a pita to deal with for many reasons, but in theory careful calibration deals with most of them.

I have never heard of a DIY thermocouple reader; I would be interested in one just from the standpoint of seeing how difficult it is to do.

The application and your patience really dictate which of the DIY solutions are better.

I am also interested in this as I want to wire up my gaming computer project with TC's that all go to a LCD display.

I found this so far:
http://www.chipcenter.com/circuitcel...99/c129cd6.htm

http://www.autospeed.com.sg/cms/article.html?&A=0376

I thought about the game port option but then realized it is my gaming computer and would be using that port. I don't know....

don't use USB controllers?

The joystick is, the wheel is not. Good point though. It is about time to get a new wheel anyway.

Well the different options you have stated are more interfaces than uses really.

But to explain. I want to know how to build the probe itself. as well as the calibration for it that allows it to give an acurate reading on a program like MBM or something similar.

The primary use would be for monitoring CPU, GPU NB etc temps with a more accurate than the in socket thrermistor that is so wildly inaccurate on most mobos.

Although it would be nice if we could have one that displays on MBM as well as an LCD display up front of some sort for when you are gaming and such.

Well for MBM AND LCD interface then the easiest is to do something like the gameport setup with 4 thermistors (get some from a digidoc if you need flat ones perhaps). You can push those into MBM5 using a program at www.benchtest.com and Jim also includes all the spreadsheets and info required to calibrate. Epoxy one thermistor under the CPU core and do a bit of modding to the socket to make a place for the wire.

Then once you have those 4 probes in place and running, just add a serial or parallel LCD and get the numbers from MBM to display on that. I really like Crystalfontz units, but even cheap ones from surplus work fine if you can track down software.

Well this site has the info on adding sensors for MBM using the SMBUS,
It works great and its simple
http://www.madhacker.org/mbmsensors.htm
I dont have much of anything as far as building the probe or the calibration of it.
I am thinking that you could use one of those IR thermal guns to compare your sensor readings with a know good reading,
then you could add a resister (or trim pot) in the line to adjust the sensors reading to match the known good temp from the gun ?

Calibrate using a range of water temps and a better thermometer. IR gun I dont think is going to be very accurate (how to get to exactly the same spot as probe?) Plot Temp1 vs. Temp2 with one of those being the new temp probe and another being your known good one. 0.1C accuracy readers (thermocouples or thermistors) arent too expensive on ebay for using as your standard.

Diodes won't adjust in temp properly by adding a pot or resistor. It just isn't how they work. Every resistance you put in the wire makes it deviate more from the true reading, in the region of 0.4-0.8C per ohm resistance. Solder points can muck them up too for this reason. Not to mention that wire length and choice and electrical interference can make them deviate from linearity. Oh and you can't fit them in tight spots easily.

Oh and with the 16QSOP parts in that link, you must be a god among men if you call that soldering "simple". Possible for sure but slow going and a pita for me.

To me the maxim/diode solution is a lot more complex and has a lot more potential sources of headache. I have a few around the house though that certainly do the job.