Temperature 'sender' (sensor) Custom Replacement?

[808Dave]

I've got a never-meant-for-USA Caterpillar excavator that, just as they told me would happen, is more than a small challenge to repair, due to lack of both English documentation and USA-friendly part-numbers. The electrical temp gauge definitely falls under this description.

I've utterly given up on finding a compatible or the correct original temp sender (old one reads "open", so gauge is always pegged on 'Hot'). Nowhere can I find a listing of these devices by their physical or electrical characteristics - it's always "make, model, year," and not "ohms, thread-size, length."

What's inside these senders? I'd assumed thermistors, chosen per the OEM's design. But when I look at available thermistors from a random online sampling, seems like the lowest resistance is about 3K ohms. This makes no sense for a replacement in my application, because, using test-resistors to replace the sender, I find I need something along these lines:
Cold (25C): 43 ohms
Hot (110C-ish): 0 ohms.

My plan was to (really) drill out the existing one, and epoxy a new thermistor in place. And I'd do it in a heartbeat, gladly, to get past all the screwing around I've done with trying to find the "right" original sender replacement.

It could also be something "went wrong with" the gauge, but it's built into a cluster ('cluster' is right!) and so that wouldn't be a simple swap to replace.

And it's consistent, if that matters, in the 43ohm - zero-ohms range testing, months apart.

So is there some different variable-resistance sensor of some kind in these senders, then? Any suggestions for a replacement component, along the "drill the old one out" lines?

 

[Bluejets]

(old one reads "open", so gauge is always pegged on 'Hot'). Nowhere can I find a listing of these devices by their physical or electrical characteristics - it's always "make, model, year," and not "ohms, thread-size, length."

What's inside these senders? I'd assumed thermistors, chosen per the OEM's design. But when I look at available thermistors from a random online sampling, seems like the lowest resistance is about 3K ohms. This makes no sense for a replacement in my application, because, using test-resistors to replace the sender, I find I need something along these lines:
Cold (25C): 43 ohms
Hot (110C-ish): 0 ohms.

My plan was to (really) drill out the existing one, and epoxy a new thermistor in place. And I'd do it in a heartbeat, gladly, to get past all the screwing around I've done with trying to find the "right" original sender replacement.

Seems something in your readings is not correct.
Usually meters go "full scale" at zero (or near zero) resistance but your initial quote says otherwise.
Then you supply hot and cold resistances completely opposite.
Drilling out the sender is not a good idea, could end up with a giant blow-out on a "hacked" sender.

 

[Alec_t]

"Cold (25C): 43 ohms"
Plenty of thermistors available with resistances in that ball-park.

 

[davenn]

"Cold (25C): 43 ohms"
Plenty of thermistors available with resistances in that ball-park.

can you be a little more specific please, which one were you looking at in particular ?
99% of those were for 10k and up

 

[808Dave]

Seems something in your readings is not correct.
Usually meters go "full scale" at zero (or near zero) resistance but your initial quote says otherwise.
Then you supply hot and cold resistances completely opposite.
Drilling out the sender is not a good idea, could end up with a giant blow-out on a "hacked" sender.

I've seen many that act as I describe. No question my tests are accurate - repeated on more than two occasions with same results.

Also no worries about drilling out sender. Water-jackets typically run 15PSI max, well within the realm of "easy to seal" with an epoxy-filled threaded-brass plug, like the original - which wouldn't present the epoxy to the water side of that interface anyway.

 

[808Dave]

can you be a little more specific please, which one were you looking at in particular ?
99% of those were for 10k and up

Thanks for jumping in, davenn - but what's the 10K thing?

I've just been scratching the surface of digi-key's index and mindlessly applying Beta calculations without immersing myself in manual calculations ( using 'NTC Thermistor Resistance Calculator V2.0' at http://www.electro-tech-online.com/tools/thermistor-resistance-calculator.php ). At Digi-key's site, I found 26 options that I guess have a 25C resistance of *47* ohms (presumably close enough to my ideal 43 ohms). Then, if I plug in the Beta of 3500 (BTW - I made a big assumption here - is the "Beta" in the equation taken from what is listed in DigiKey's specs as "B25/50?"), it puts me in the right ball-park: the NTC thermistor calculator seems to be indicating I'd get 47ohms at 25C, and 3 ohms at 110C, which is certainly close enough for a scale that's only calibrated in terms of "H" and "C." That is, it's close enough IF my mindless use of the calculator was a good guess.

If my assumption about the Beta value was OK, then I presume it would just be a matter of me picking out one of the 26 options that doesn't require me to buy the minimum 4000pcs...?

 

[Kiwi]

"(old one reads "open", so gauge is always pegged on 'Hot')"
"Cold (25C): 43 ohms. Hot (110C-ish): 0 ohms."

Like Blujets, I think that there is a contradiction here.

What is the model etc of the digger?
Is this a Mitsubishi digger with Cat badges?
What is thread of sender?

 

[808Dave]

Contradiction? oh, wait, now I see...yes, substitute "dead short" for "open" - my bad. Better way to put it would be that when the sender lead is grounded, gauge pegs hot.

Yes, please, find me a sender if you're up for it - do those machines exist in NZ? It's a CAT307SSR, indeed part Mitsubishi, part CAT, 1993. But I should say I've so far had three senders that fit, with wildly different resistance readings, eg 5.5K ohm, 3K ohm @ 25C-ish, so they don't come close to lifting the gauge off of cold, even at well over boiling. I believe the thread is 16mm but I'd have to verify when it's light out & our current tropical storm has moved on...

 

[davenn]

Thanks for jumping in, davenn - but what's the 10K thing?

that's its resistance @25C 10kΩ (10,000 Ohms)

 

[Alec_t]

can you be a little more specific please, which one were you looking at in particular ?

Use the 'Resistance in Ohms' filter on that site to select the resistance value of your choice. I used the 47 Ohm parameter and, like 808dave, got a hit list of 26 items.

 

[808Dave]

Yes, as Alec wrote, it requires a primary filtering to focus on 47-ohm rated items. http://www.digikey.co.uk/product-detail/en/epcos-tdk/B57311V2470K60/B57311V2470K60-ND/652323 would be one of those options, though Digikey shows zero in stock. I'm assuming I can shop around a little and find something with a beta of at least 3500 (seems to be a standard) because a higher beta (apparently) gets me a wider/finer temp response.

But IS Digikey's "B25/50" really the same "B" (beta) used in the calculator?

 

[Sunnysky]

B is the slope at two temps so a part with B25/50 does not have the same value as B25/85
So yes if calculator allows a value for T and 25'C, it works.

But ensure application does not exceed sensor current rating for self heating reasons and the installation is not some thick thermal insulator like epoxy.

 

[808Dave]

Thanks - point taken on epoxy installation (I'm thinking I'd embed the device in some silicone-grease at the closed end of the threaded plug, then epoxy above that level to pot it). Pretty sure the little meter's impedance is very high, so little self-heating.

But I don't understand your reply about beta. I see that Digikey's 25/50 and 25/85 values are different, but don't know whether one of these, or another un-listed "beta" is SUPPOSED to be used in the calculator at http://www.electro-tech-online.com/tools/thermistor-resistance-calculator.php . The field calls simply for "Thermistor Beta Value."

 

[Sunnysky]

charge target temp to 50 or 85
B is the log_e slope of R at two temps. hence B25/85

I suggest you calibrate it with a pot first