A warning about cheap metal film resistors

[MrBungle]

I just used some metal film resistors from a batch I bought a few years ago for general repairs and projects. These were supposed to serve as current sense resistors in a switching regulator in this instance. Needed 0.25 ohms at 1W so I used 4x no brand 0.25W 1R 1% resistors in parallel. Had these on hand so "it'll do" while I was waiting for some nicer parts to arrive from RS.

Well my regulator was happily churning away until there was an overcurrent situation, part of the regulator testing. Turns out that when they get hot they have a massive negative temperature coefficient and my sense resistor dropped to a paltry 0.05 ohms. This caused a massive measurement error for the poor little MC34063 which understood that it wasn't in an overcurrent situation, which drove all the magic smoke from the outboard MOSFET and the inductor simultaneously. Fun fun fun. Windows open, smoke let out, back to sanity.

A quick check with a DVM, thermocouple and the gas cooker with a few random selected resistors and the things can pretty much drop by 60% in value at 100 degrees C and recover completely with a couple of seconds! On low resistance values that's disastrous.

So ParcelForce just dropped my order off containing some nice Vishay Pro units. Absolutely fine. 50ppm/degree C as described in the datasheet keeping them well in tolerance in overloads.

Moral of the story: buy quality parts. I've just thrown the whole bag of 1280 resistors in the bin.

[Boater Sam]

Negative temperature coefficient resistors = premium price.
You will want some, some day.

[boxdoctor]

Hi, Any chance of a pic. of the faulty types, please?
I keep numerous different types of resistors in stock, and would love to know which ones are giving this trouble, as simple ohmmeter testing is not going to reveal this defect. I can then check any I find in my stock,and if the same problem is present, bin them. Cheers, Tony

[dave cox]

Certainly not normal to have such a big negative coefficient for something used to make resistors ! It must turn around eventually and go back up but at what temperature ?

dc

[MrBungle]

I didn't get a photo of the original ones. Similar to the ones depicted below. Not very helpful as most of the metal film varieties look similar:



I've replaced them now with a mix of TE LR1F and Vishay MRS25 resistors from RS. These look like the following (Vishay at top, TE at bottom):



I spent 4 hours yesterday testing and cleaning out my component stock. I'm going to buy new from RS only in future. Cheap components are a false economy (unless you're avoiding Linear Technology...)

As for the question about at what point does the negative coefficient return back to positive: when they start glowing orange (I tried it )

[Boater Sam]

So when they go positive coefficient again they are all 33K Ohm?

Must check my stocks, lots of surplus resistors.

[MrBungle]

Haha

[Radio Wrangler]

They look very like the Corning ones I used to use, and those were superb. It's a begger when they start faking resistors.

David

[MrBungle]

Vishay acquired Corning in the 1980's AFAIK (along with Philips/BC/Sprague/Dale/Siliconix etc) so that would explain the similarity.

Some of the best quality parts I've had are from Vishay.

[Maarten]

Almost certainly carbon film resistors in disguise. This has been a standard practice in component forgery for years now