Resistor values change when overheated

We would assume that you lifted or removed the resistor from the circuit and verified the resistance. I've found that generally most types of resistors (carbon eg) when subjected to that point that is above their wattage rating WILL go down in value to a point only to reverse and go to a point where they will basically burn open.

Sounds like there is something wrong elsewhere in the circuit. Hope this helps.

Yes, I removed the resistor to measure it.The banks of 7 identical resistors in a number of boards had all overheated,the designer must have assumed that a darlington took a minute base current even though it was switching 20A.The resistors were dissipating about 1/2 watt.Problem solved by 1/2 watt resistors and standing them off.

As in many circuits if a resistors value is reduced they take more current and this is a runaway situation,more current,more heat, causing lower ohms,more current etc.

The old carbon resistors always went high,and of course without the runaway characteristc.

But I'm curious to know if all metal film resistors go low when roasted,and what the exact mechanism is?

oldeng, yes I believe that all metal film resistors will have that property. I have seen the same fail for no apparent reason and along the same lines as I think you have described.

Not knowing your actual application, but the designer could have made the wrong call and maybe only needed carbon film 5% types.

In design, I tend to use about a 100% wattage factor (use a 1/2 W if the power dissapation is 1/4W). Also, I call for MF types only if the application needs 1% tolerance. Otherwise (5% and up) carbon film is fine.

Maybe we are in queue now . . .

Hi Bill ,yes,we are on the same wavelength now.Thanks for your helpful comments.

I agree that it would be better to use carbon resistors. I came across another snag with metal film where the spiral track was of uneven width,being very narrow at one point,thus causing a hot spot as nearly all the resistance was concentrated at that point.The laser etching had cut short having achieved the 'correct' resistance value. There were 72 of these and after the first few were 'non carbon' copies I just changed the lot.

Just another lesson learned and hopefully remembered when diagnosing and repairing.

If any equipment is operated outside its rating parameters then expect odd things to happen. Try flying a Boeing 747 at 50kph and 1500kph, for example!

This is manufacturing defect of Metal Film Resistors. During manufacturing of MFR ceramic subtract is rotated in a barrel and resistance material like Nickle Chromium, Chrome Silicon, Chrome Silicon Tungsten, etc is sputtered (or by Chemical Vapour Deposition). All metals have a positive temperature coefficient. To bring-in Semiconducting properties 6 to 18 SCCM of oxygen is feed in to the chamber so that Chrome Oxides or Silicon Oxide's from along with the metal. These Oxide's have negative temperature coefficient and overall resistance will remain constant over a temperature range.

In your case the manufacturer during process must have passed more oxygen making it a semiconductor rather than a resistor.Degraded MFRs are sold in market as SFRs, that is Slandered Film Resistors, that got positive temperature coefficient. You may try some or get components from a reputed manufacturer. Also the ceramic subtract that is use as base metal, baking, laqour, etc also play major roll in keeping resistance value constant over a temperature range. Also you can try mounting resistor lead lengths more and then soldering on to PCB.

I found this to be a very interesting forum that I will never forget!! Many thanks.

Problem noted and understood.

If I had been asked before, I would not have stated the problem correctly......!!!

Well done in even just taking the time to do the job properly and informing all of us.....

I agree with Andy Germany. This was a most interesting and informative thread. I have certaintly learned a lot about metal film resistors from it.

If you don`t need low tolerances then coal resistors mostly are the better choice in applications where higher pulse ratings can occur. The tolerable rating is 5-20 times higher than with MFRs.
The lowered resistance when extremely owerheated is explainable when you look at the parabolic resistance over temperature diagram. Resistors are manufactured/ trimmed in that way that the maximum resistance lies at the operating temperature. The operating point is selected so that sweeping over the whole temp-range the resistance is a little lower at low temperatures, a little higher at mid tems and lower again at high temps. So the overall error over temperature is minimised. Overheating significantly means, that you are on the descending slope wide outside your specs which leads to a non-revertable change of value.

Well put and completely true, I was just unaware before of the possible problems that could be encountered with metal film if I went slightly outside of the design envelope.

I am just VERY grateful to be a member of CR4!!

Right then kvsubramanyan,you've really educated me and probably the rest of us about mfrs. I'm going to use good old carbon in future 'cos they're simpler and have simple failure mechanisms.

I'm off to try to fix an old telequip 'scope uneconomic but a challenge,so, as someone famously said, 'I may be gone some time'.

One more thing about resistors is their noise aspect, which determines which type to use for different circuits. Wirewound resistors are the quietest, having only thermal noise, followed by metal film, metal oxide, carbon film, and lastly, carbon composition. Resistor noise is made up of three main types: thermal, contact, and shot noise. Thermal noise is mainly dependent on temperature, bandwidth, and resistance, while shot noise is dependent on bandwidth and average DC current, and contact noise is dependent upon average DC current, bandwidth, material geometry and type. Another consideration for resistors, which is sometimes overlooked, that resistors have a max voltage rating. The 1/2 watt and some 1 watt units usually are only rated for 250-350V. Be sure to get a resistor rated for the appropriate voltage.

The problem for me with wire wound is that they have often a tiny inductance.....that in 99% of circuits does not matter, but one should not forget it!!

I have worked as Maintenance manager in MFR manufacturing unit for few years. Also I have visited one big unit in Germany (Berlin) that makes MFRs. Production and QC guys concentrate on Positive temperature coefficient of devices manufactured than Negative parameters. That is why all this non-sense happens. Manufacturing Process is critical but price they get is very low. Hence for experimental and other purposes try to procure from best manufacturer by paying a little more. After metal film coating and before Trimming the components they are baked at high temperature to stabilize the value, but some manufacturers give less importance to step. For ensuring that you got good components you can heat the device with multimeter connected (Heat with soldering iron or hire dryer)