What Are Platinum Resistance Temperature Detectors?
Look, if you're reading this, you probably already know temperature matters in batteries. But here's the thing most people don't tell you: getting temperature RIGHT is way harder than it sounds, and getting it wrong will cost you. A lot.
We learned this the hard way back in 2019. One of our forklift fleet customers in Northern China had 40+ units running our battery packs. Winter hit, temps dropped to -25°C, and suddenly we're getting calls at 2 AM. Trucks won't charge. BMS keeps throwing over-temp warnings. In minus twenty-five. Yeah.
Turned out the NTC thermistors we were using drifted almost 8°C in cold weather. The BMS thought the cells were overheating when they were actually freezing. Chargers shut down. Entire fleet dead in the warehouse. The penalties, the replacement costs, the flights to go apologize in person... let's just say someone at our company still has that contract framed on their wall as a reminder. Cost us over 2 million RMB.
That's when our boss literally slammed his hand on the table and said "from now on, platinum RTD on everything, and if anyone suggests NTC again to save money, it's coming out of their bonus."
So yeah. That's why we're writing this. Not because platinum RTDs are fancy. Because we've been burned (ironically, by bad temperature readings) and we're not doing that again.
The Basics (But Actually Useful)
Platinum RTD stands for Platinum Resistance Temperature Detector. The principle is dead simple: platinum wire gets more resistant to electricity as it heats up. Measure the resistance, you know the temperature. That's it.

The industry standard is called Pt100, which just means 100 ohms at 0°C. There's also Pt1000 (1000 ohms at 0°C) which gives you a bigger signal to work with, handy when you've got long wire runs or noisy environments. We use Pt1000 on anything with cables longer than 2 meters now because we got tired of chasing down 0.5°C errors that turned out to be lead wire resistance we forgot to compensate for.
Why platinum specifically? We actually tried nickel and copper early on. Cheaper, sure. But nickel goes crazy above 300°C and copper oxidizes if you look at it wrong. Platinum just... works. -200°C to +850°C, doesn't corrode, doesn't drift. Boring and reliable, which is exactly what you want when your battery pack is sitting inside a 3-ton forklift moving pallets around a warehouse.
One of our engineers likes to say "platinum is expensive because it's lazy" - it doesn't react with anything, doesn't change its mind, just sits there being accurate. We'll take lazy.
Three Ways to Build Them (And Which Ones Actually Work)
There's wire-wound, coiled element, and thin film. Here's what we actually think about each:
Wire-wound
The old school way. Platinum wire wrapped around a ceramic core. Super accurate, super stable. Also super expensive and kind of fragile. We use these for calibration reference and that's about it. Had one customer who insisted on wire-wound for his whole fleet because "only the best." Six months later half of them had drifted because the vibration from his rough warehouse floor loosened the windings. Now he uses coiled elements like everyone else.
Thin film
Cheap, small, fast response. Great for consumer stuff or low-stakes monitoring. We use them on our 12V/24V small packs where if something goes wrong, nobody's getting hurt and nothing's catching fire. Would we put them in a 600V forklift pack? Absolutely not. The ceramic substrate and platinum film expand at different rates, and after a few thousand thermal cycles you start seeing drift. One imported brand we tried - expensive one too - drifted 1.5°C in under a year. Supplier blamed our installation. We blamed their QC. Either way, we stopped using them for anything critical.
Coiled element
This is what we spec for high-voltage packs. The platinum coils sit loosely in ceramic tubes packed with powder, so they can expand and contract without stressing the wire. They handle vibration, they handle thermal cycling, they don't complain. More expensive than thin film but we sleep better at night. Any lithium battery pack supplier doing industrial applications should seriously look at these.
2-Wire, 3-Wire, 4-Wire: The Debate That Almost Split Our Team

This one caused actual arguments in our office. Our hardware lead wanted 4-wire on everything. Our cost engineer nearly had a stroke calculating the extra connector pins and cable costs. They didn't speak to each other for like a week.
Here's the deal:
2-wire is fine if your cable is short and you don't care about half a degree here or there. The resistance of the wire itself adds to your reading. For a 10-meter cable run, you might be off by 1-2°C. Not great when your BMS is making decisions based on that number.
3-wire compensates for lead resistance and is what most industrial setups use. Good enough for 90% of applications. This is our default now.
4-wire eliminates lead resistance completely. We use it for cell-level monitoring during R&D and validation, and for any customer who's paying for premium accuracy. Also mandatory for any lithium battery OEM solutions where the customer is doing their own safety certification - auditors love seeing 4-wire.
The argument ended when our hardware guy said "I can run two extra wires or I can explain to a customer why their thermal runaway protection triggered 3°C late." We run 4-wire on all high-voltage packs now. The extra cost is like 0.3% of the pack price. Not worth losing sleep over.
Where to Stick Them (Placement Actually Matters)
We've seen some wild stuff from other industrial battery manufacturers. Sensors zip-tied to random spots. Sensors pressed against the case with a piece of foam. One memorable audit where we found a sensor that had fallen off completely and was just dangling inside the enclosure, reading air temperature while the cells cooked.
Our rules now:
On cells
Direct contact with the case, thermal paste or 3M VHB tape (we like 9888T), then potted over with epoxy. Yes it's annoying to rework. That's the point. If someone has to chisel through epoxy to move a sensor, they'll think twice about whether they actually need to.
Between cells
Catches hot spots from cells that are degrading faster than their neighbors. This is how we found a batch of cells from a supplier that had internal defects - one cell was running 8°C hotter than its neighbors. Caught it before anything bad happened, swapped the module, traced it back to a manufacturing lot, rejected the whole batch. The sensor paid for itself about 10,000 times over that day.
Busbars and connections
Loose bolts heat up. A lot. We had a pack where the main contactor terminal wasn't torqued properly, and under high current it was hitting 85°C while the cells nearby were at 35°C. Customer complained about "random" power derating. Wasn't random at all - the BMS was doing exactly what it should do. Fixed the connection, problem gone. Now every pack has a sensor on the main current path, non-negotiable.
Coolant in/out
If you have active cooling, you need to know if it's actually working. Seen too many systems where the pump died and nobody noticed until cells started swelling.
The Stuff They Don't Put in Datasheets
Here's some things we learned by screwing up:
Self-heating is real
The current you push through the RTD to measure it also heats it up. Standard is 1mA for Pt100. We had a junior engineer once who cranked it to 5mA because "bigger signal, better accuracy, right?" Wrong. The sensor was reading 2°C high. In a 48V golf cart pack it just meant slightly early fan turn-on. In a 400V industrial pack it would've meant nuisance shutdowns. Use 1mA. Or less for Pt1000.
Wire routing matters
Keep RTD wires away from power cables, motor drives, inverters - anything that's switching big currents. We use shielded twisted pair now after spending three days debugging a "noisy temperature signal" that turned out to be EMI from a VFD mounted 20cm away. Lithium battery wholesale distributors take note: if you're reselling packs and the integrator puts the BMS right next to the motor controller, temperature readings will be garbage.
Tolerance class is not accuracy
A Class B Pt100 can be off by ±0.8°C at 100°C per the spec. That's the ALLOWABLE error. If your BMS assumes it's dead accurate and sets thermal limits at 60°C, you might actually be limiting at 59.2°C or 60.8°C. Design your limits with tolerance in mind. We build in 2°C margin on all our protection thresholds.
Cheap sensors drift. Good sensors don't
We tracked sensors from three different suppliers over 2 years of field operation. The budget ones (like $0.80 each) drifted an average of 0.6°C. The mid-range Japanese ones ($2.50) drifted 0.1°C. The premium Swiss ones ($8) drifted basically nothing but c'mon, $8 per sensor? We use the Japanese ones. Best value for industrial stuff.
Why We Standardized on Pt100 Across Everything
We used to have a mix: NTC thermistors on low-cost packs, Pt100 on premium ones, some weird KTY sensors on legacy products. It was a nightmare. Different calibration curves, different BMS firmware branches, different spare parts inventory.
Now it's Pt100 everywhere. Same sensor, same supplier, same handling procedures. When something goes wrong, we don't have to ask "wait, which type of sensor is in this pack?" Battery system integration gets way simpler when you're not juggling five different sensor types.
The cost difference between NTC and Pt100 is like $1.50 per sensor. On a $3000 forklift battery pack, that's noise. On reliability and support costs, it's everything.
For Anyone Speccing Sensors Right Now
If you're a design engineer reading this, here's the short version:
Use Pt100 or Pt1000. Coiled element for anything over 100V or anything that vibrates. Thin film is fine for bench testing and low-stakes monitoring. 3-wire minimum, 4-wire if you're doing certification or precision work.
Buy from established sensor makers with actual traceability. We've seen counterfeit Pt100s that were actually NTCs with faked markings. If the price seems too good to be true, it is.
Pot your sensors. Use thermal interface material. Strain-relief your cables. Document your sensor locations so the poor technician 5 years from now knows where to look.
And for anyone sourcing packs: ask your supplier what sensors they use and why. If they can't give you a straight answer, that tells you something. If they say "we use platinum RTDs because we learned the hard way what happens when you don't," that's a supplier who's been around long enough to know what matters.
We've been doing lithium battery packs for over a decade now - forklifts, golf carts, AGVs, floor scrubbers, you name it. Temperature sensing is maybe 0.5% of the BOM cost and 50% of the field failure root causes we've diagnosed. It's just not worth cheaping out on.
If you're looking for a team that actually sweats the details on stuff like this, you know where to find us. We're always happy to talk thermal management with anyone who's serious about getting it right.
And if someone on your team suggests saving $1.50 per pack by switching to NTCs, feel free to send them this article. Or the invoice from that 2019 incident. Whatever works.

