Help me select a thermistor?

[amanda-marie]

Hi,

I truly hope I might get some direction here.  I am no an EE.  I only know the rudiments of thermistors.  I need to replace a couple of them in a water cooling circuit.  Two LED panels register F.  from approx 60 to 200 degrees.   I have a sample thermistor from a sending unit matched to the LED panel.    I am trying to find replacements so that a pair of identical thermistors will feed the panels.   One panel is for cool/out the other is for hot/in.    I developed a graph base on temp rise using a multimeter and digital thermometer in my own workspace.

http://img381.imageshack.us/my.php?image=finalrtcurveforthermistja3.jpg

And I sent the thermistor into a manufacturer who tested it and determined that it was approx 1006 ohm at 25C and 142ohm at 70C.

Everywhere I look, I am told that this cannot be replaced with an off the shelf part.  The LED panels are only for visual confirmation of relative temps to monitor efficiency of the system.  The data does not control process.  I don't think I need high accuracy.  A 'bead' type thermistor would be best suited to me for implementing a thermistor into this cooling circuit.   All of the tables, curves, charts and graphs accompanying thermistors really make my head start spinning. Is this really such an insurmountable quest?    

How close would I actually get (in the real world, without being too technical) with a part such as this

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=P10537CT-ND

An SMD device would work ok for me.

Thank you so much for any advice at all.   This device is turning into a major stumbling block in this project. 

Amanda

[lacy]

The only thing I can suggest is getting one and putting it in the circuit and see how the displays react. As long as it poses no personal safety risk I do not see where any harm can be done by doing this. What you will probably need to do is to  heat it up someway to find out what the results are across the temperature ranges to see if the displays give an accurate readout accorcding to what you need. If you test it out of circuit and get close to the same resistance values in the temperature range you stated, then I would tend to believe that it would work without a problem. I would suggest testing it both ways out of circuit and in circuit just to make sure.

This is about all I can suggest as it seem you have done a fabulous job at researching this. If you contacted the manufacturer of the device you are working on then they will always say you can't replace parts with off the shelf components and that you will have to order a part from them. I have run into that numerous times when I was serviciing televisions. Usually I ignored them and did my own thing and it worked 99.5% of the time. They just want to sell you their part and most of the time they get it from the same place you can get it like Digi-Key, Newark, or Mouser.

I hope this helps.

[Ridis]

Digikey part with R = 1k and B = 4500 K is a good replacement. 
If you look at page 3 of datasheet you'll see that at 70 °C the column of 4500K (this means B=4500K) give you a resistance ratio of 0.138. This means that a thermistor with a resistance of 1 k at 25 °C will have 130 ohms at 70°C. This is very close to your 142 ohm at 70 °C.

[amanda-marie]

I went ahead and ordered those. About fifty cents apiece.  I was really hoping to save a lot of work by finding a 'bead' thermistor, but I'll make due. I ordered four of these (two spares)  SMDs  USPS.  They arrived in a 'postage paid' USPS box from Digikey.   They paid the shipping!  Very cool.

I just spent a couple of hours doing some microscopic soldering and epoxy potting to get them into a couple of probes.   Hot air seems to work good with these since you can ramp up the temp to get the solder to flow then stop on a dime without risking a lot of temp differential to  the chip. 

Hope they work.  

And thank you Ridis for looking into the specs for me.   

Message Edited by amanda-marie on 04-16-2008 09:59 PM