Rectal thermometer

[Nelg]

Thanks! Now that I have your attention...I am serious, I need to record rectal temperatures for a University assisgnment and I'm trying to find a suitable project to burn and build... data logger, 2mtrs of cable, and you can guess the rest

My education is in Sport Science, and I have been know to 'tinker' with the odd pic chip. I don't have the skills to program an entire project such as this, but I am certainly capable of building and burning.

If anybody knows of a suitable web based project to suite (with source code to burn), I would really appreciate the heads-up

I have done quite some searching and I am assuming a thermal couple (due to it's size) would be most appropriate, but I am a little confussed about accuracey and a non-linea voltage relationship??? (as I understand it)...any thoughts greatly appreciated.

Cheers

Nelg.

 

[Diver300]

Don't use a thermocouple. They have a very small output and the reading depends on the temperature of the cold junction.

Better would be a RTD, which is a resistance that changes with temperature.

However, I would recommend a temperature sensing IC, such as an LM73
**broken link removed**

If you want to have fewer wires and you don't mind a slightly worse resolution, you could use one of the Maxim 1-wire range, such as the DS1825
**broken link removed**

These ICs measure temperature directly, and so there is no other data conversion to do. They interface directly to microcontrollers.

If you use a microcontroller such as the 16F876, there is an I2C port for connecting to the LM73, and a serial port that connects to a computer. You can use a smaller microcontroller like a 12F629 but the code to connect to the temperature sensor and the PC is a bit more complicated. This is because the 12F629 doesn't have the specialised ports, so the routines have to programmed in code.

I don't think that there are any microcontrollers that have ports specifically for connecting to the 1-wire sensors, so the interface to them has to be programmed in code for all microcontrollers.

The voltage levels that a serial port on a computer works at is bigger than microcontrollers work at, so a converter such as a MAX232 or similar is needed.

To log the data, you could quite easily program a microcontroller to read the temperature and output it on a serial line to the serial port on a computer. On the computer, there are lots of terminal emulator programs that can record to a file what arrives on a serial port.

The overall scheme is:-
LM73 converts temperature to serial data
PIC initiates reading the temperature and outputs data in a different format
MAX232 converts the voltage level to RS232 levels
Terminal emulator program. Reads the serial port and stores the data to a file.

 

[Nelg]

Thanks Diver300,

I have read about the RTD in the book "Microcontroller Based Temperature Monitoring & Control" and wasn't sure if this would be the best option...what you write makes sense. (I'm also an electrician by trade, but this doesn't make me a programmer )

I did find a site recently (but cannot find it again) that sold a kit where you made all the bits and they sold you the code...if anybody knows of similar sites, I'd appreciate the link

I am about to head down to my local electronics shop to see how big the 'insertable' pieces are in physical size...I'm also off to do another major goo*le search on "LM37 microcontroller temperature"...

Would really appreciate any links

Nelg

 

[Diver300]

It is an LM73 so don't go searching for an LM37.....

has quite a few examples of i2c code that would let you talk to the temperature sensor.

is examples of serial code to talk to the PC.

You will need something like a PICKIT2 to do the programming. Part Number: DV164120 - PICkit 2 Starter Kit (https://www.microchipdirect.com/productsearch.aspx?Keywords=DV164120) comes with a demo board that would give you somewhere to mount the MAX232 and its associated components. You could power the whole lot from the PICKIT2.

You can use DIP devices for everything outside the "probe". The LM73 is only 3x3 mm but you only need 4 connections to it.

As a safety item, I would suggest a resistor of about 1k in series with the 5V supply to the LM73, and a 10 nF capacitor to ground after the resistor. The resistor will limit the current to less than 5mA if there is a short, and the capacitor will keep noise off the line during normal operation.

 

[Torben]

Just one more point, which you might already have considered:

Unless you really know what you're doing I would think it's a good idea to power the whole device from batteries so that there is no possible way that mains voltage could appear on the probe if something failed.

Ow.


Torben

 

[philba]

Just one more point, which you might already have considered:

Unless you really know what you're doing I would think it's a good idea to power the whole device from batteries so that there is no possible way that mains voltage could appear on the probe if something failed.

Ow.


Torben

After the old joke about how you can tell the difference between oral and rectal thermometers, that was my next thought. Not a good idea to have line voltage even remotely close to, uh, conductive tissue...

 

[Nelg]

thanks

Thanks for all the input people

in breif:
* no wonder I couldn't find anything on LM37
* re: using PICkit 2 Starter Kit; I do have Nigles tutorial boards all set up...
* I'm pretty sure Isolation transformers are law for medical equipment here in Aus...thanks for the heads up.

I'm told the AD590 bases it's result on µ amps thus a non-linear reading (like occurs in voltage reading devices) won't happen and will provide a more accurate reading...? don't know enough about it...where would the world be without go*g*e?

went down to my local Ja*car yesterday and they barely knew what I was talking about... did find this in my searching yesterday
https://www.electro-tech-online.com/custompdfs/2008/09/k145.pdf

has anybody attempted to use this with accuracy...?

Thanks again for all the input

Nelg

 

[Diver300]

I'm told the AD590 bases it's result on µ amps thus a non-linear reading (like occurs in voltage reading devices) won't happen and will provide a more accurate reading...?

The AD590 is a linear output temperature sensor. It isn't any more accurate than a digital output sensor, and you still have to convert to digital to record the readings.

I suppose you could just wire one directly to a power supply and a chart recorder, or if you have an analog input logger you could use the AD590 with that, and save yourself any programming or wiring.

did find this in my searching yesterday
https://www.electro-tech-online.com/custompdfs/2008/09/k145-1.pdf

has anybody attempted to use this with accuracy...?

I haven't used that one, but it is basically the same idea as I suggested in my first post.

That design uses DS1820, and the DS1825 that I suggested has a better resolution but it very similar. The DS1820 has now been replaced by the DS18S20. Either the DS18S20 or the DS1825 can be powered from the signal line so can be run with just signal and ground, but the software in the microcontoller needs to be changed slightly for that mode.

The LM73 has even better resolution but you do need more wires to it.

They have also not added the level converter to the serial port. That doesn't meet the RS232 specification, but will almost certainly work fine. I suggested using a MAX232 to do the level conversion but it is probably worth trying without one as you certainly won't do any harm.

 

[Nelg]

Thanks Diver300

I've ordered 3 of the k145's...will see how they go. The only issue now is the isolation transformer between laptop and the eehhmmm

I think I will try to disable the power from the serial port and use a battery...

still searching for the project...