I am working on designing a circuit using an lm339 and an analog temperature sensor to activate a fan on either high or low speed, I have pretty much all of the details worked out to my satisfaction based on circuits I have found on here and other places on the internet, but when I started putting together my parts list to make an order I noticed that the LM339 is only rated from 0-70*C. Since this device will be installed in a car it will frequently be exposed to temperatures below this, and maybe above this as well (although it will be located in the cabin so it won't get as hot). Searching around I see the datasheet for the 139/239/339 indicates the LM239 will meet the temperature range requirements I have but I can't find any difference in the specifications between the LM239 and LM339 except the temperature range. Is there any other difference between the 2 parts except for this?

Thanks in advance for any help!

 

I am working on designing a circuit using an lm339 and an analog temperature sensor to activate a fan on either high or low speed, I have pretty much all of the details worked out to my satisfaction based on circuits I have found on here and other places on the internet, but when I started putting together my parts list to make an order I noticed that the LM339 is only rated from 0-70*C. Since this device will be installed in a car it will frequently be exposed to temperatures below this, and maybe above this as well (although it will be located in the cabin so it won't get as hot). Searching around I see the datasheet for the 139/239/339 indicates the LM239 will meet the temperature range requirements I have but I can't find any difference in the specifications between the LM239 and LM339 except the temperature range. Is there any other difference between the 2 parts except for this?

Thanks in advance for any help!

The die and processing are identical, it's the testing that separates them.


The IC makers do what is called correlation testing so they can predict performance at temp extremes. The higher temp versions hold their specs tighter at room temp so they know they will meet the data sheet specs at the wider temp limits. They test the parts in sequence: run the tester with the tight limits and bin those out for marking as the higher grade parts, then retest what failed to see if they make the looser limits.

NOTE: the demand for the higher grade (EXPENSIVE) parts is always low so they may not need them sometimes. In those cases they just screen to the wide limits and label everything 339 so that in a lot of 339 parts, there may be a bunch of 239 parts. It's complicated.

 

I noticed that the LM339 is only rated from 0-70*C. Since this device will be installed in a car it will frequently be exposed to temperatures below this, and maybe above this as well

Note those are typically ambient temp specs. I doubt you could go higher, but maybe a shade lower.

I would be very surprised if the performance changed much in the small drop below 0C.

 

Thanks so much for the insight, I was worried there was some subtle difference I didn't catch in the data sheets between the two and it would affect my application. Since winters here can get to -40*C on occasion I think it is imperative that I move to the LM239A instead of the LM339A. The reason for this being that if the LM339A misbehaved in a way that triggered the fans to run they may experience a mechanical failure at -40*C or draw too much current and blow fuses, or who knows. Better safe than sorry

Interesting thought that there might be some LM239 mixed in with the LM339. The scientist inside of me wants to buy some LM339 and subject them to very cold weather testing right beside some LM239... but then I think my wife would get upset if she found out I was going to do these things inside our freezer, so maybe I should not do that

 

If your specifications is to operate to -40C there is no alternative but to test down to at least -45C or -55C.

btw, you may want to consider getting LM139, not LM239.

 

I don't know what specs you are worried about at low temp, so my answer is limited. I would build a small test fixture that would easily and quickly let me evaluate the parameters of interest, put the fixture (with an IC) in a freezer, let it soak for a while, pull it out and test it. If it tests good you have one IC you need; repeat for more ICs.

Look carefully at the cost of LM1XX ICs because they were meant for military use, and although the temp specs are better, they require a lot of mil documentation you may not want to pay for.

 

Unfortunately I don't have any way to test at temperatures that low. I can get to about -35*C easily, but much past that is difficult. I also noticed the LM139 is significantly (6-7x) more expensive than the LM239. I am mostly worried about the comparator making mistakes and turning the fans on all the time or something like that. If the set points simply drift a little bit that probably isn't a big deal since there is a significant safety margin built into the set points and the other components.

I'll pick up a couple LM239s and test them in my freezer at -38*C and see how it performs, that is already 13*C outside of the specified range so it should give a valid indication of any troubles I might find.

Thanks again for all the advice

 

Compare:

LM339N 0 - 70C $0.53
LM239N -40 - 105C $0.49
LM139N -55 - 125C $2.31

Which would you choose if your low temperature range is -40C ?

 

I wouldn't be adverse to choosing a component rated for -40 if -40 is my low point... self heating would help keep me on the safe side.

Thus I would choose a LM2901N which I can get for .52 in single quantities. It's a pin-for-pin replacement for the '339, rated -40 to +125.

It even appears in the 139 data sheet: http://www.ti.com/lit/ds/symlink/lm139a.pdf

 

Interesting thought that there might be some LM239 mixed in with the LM339. The scientist inside of me wants to buy some LM339 and subject them to very cold weather testing right beside some LM239... but then I think my wife would get upset if she found out I was going to do these things inside our freezer, so maybe I should not do that

It's almost 100% certain all parts are functional at -40C, you might find some that have DC specs drift out of min/max limits slightly in the 339 group.

 

Since you mentioned that the device will be installed in a car, I am compelled to ask you what will be the purpose of the said device.

Please describe the device you want to build in a few words.

 

Since you mentioned that the device will be installed in a car, I am compelled to ask you what will be the purpose of the said device.

Please describe the device you want to build in a few words.

Wasn't that explained in his first post?

I am working on designing a circuit using an lm339 and an analog temperature sensor to activate a fan on either high or low speed...

 

I sort of glossed over the purpose of the device because my question was mainly about parts suitability and not so much a "how do I build this" type question. Since you are curious I'd be happy to share but I want to mention in advance I don't have time to start on the project for another month or so yet, I was just trying to combine shipping with a project I am working on right now so that I can have some parts in hand and do some experimenting

I am building a radiator fan controller for my 2nd gen RX-7. It has a mechanical fan which does a great job in daily driving but when at the auto-x course and there are extended idling periods in the pits then very high rev/high load/low ground speed situations on course it cannot keep up. I am getting the temperature climbing well above where I would like it (and so are many other people) as the mechanical fan disconnects at about 2000RPM, so you have high revs, high load and very little airflow over the radiator. This means I have to keep an eye on the temperature gauge which distracts me from driving, and occasionally be less aggressive than I would like to be which hurts my times. The problem is even worse when I allow co-drivers in my car and we are running twice as often.

A common solution to this is to swap in the radiator fan from a 3.8L Ford Taurus, as it is a 2-speed fan which moves a lot of air, and the shroud from it is almost a direct fit on the RX-7 radiator. The problem here is the fan is controlled by the ECU on the Taurus so you can't just bring that over. Most people install a simple temperature switch to control a relay which runs the fan, but that restricts you to a single temperature for switching and a single speed. 2-speed fan controllers are somewhat expensive, and as an avid DIYer I felt this was a good opportunity to build my own solution.

I am going to install a GM temperature sensor in the outlet for the upper radiator hose (there is an undrilled and untapped spot there perfect for this) and use this for temperature sensing. I have found a few circuit diagrams using LM339 to switch a fan on or off at a set temperature so I am basically going to build 2 of these circuits side-by-side with different set points to control both speeds of the fan. Of course there needs to be a few extra components to prevent the low speed from being powered while the fan is on high speed, to bring the low speed fan on if the A/C is turned on regardless of engine temp, and to upgrade the power handling ability so that this box can switch the power relays for the fans. My goal was to build a circuit that does not require any programming or additional software configuration (even though really I prefer using microcontrollers) so that I could share it with other members on an RX7 forum so that they could build their own if they liked. Using potentiometers or a line of resistors and jumpers the temperature set points and hysterisis amounts count be adjusted easily by somebody with no electronics experience.

Sorry if I rambled here, I always explain these things poorly