# How do final stage resistors fail?

I replaced the FSU on my BMW today to restore normal fan speeds to my HVAC blower (having just 20% fan speed in 40+ °C isn't always comfortable).

After removing the old resistor pack, curiosity got the better of me and I decided to measure the resistances across all possible pin combinations.

The results are logged in the table below. By asking this question, I'm looking to understand what happens to the final stage resistor over its lifetime that eventually makes it unfit for service.

Pin legend

``````A   B   C
o   o   o
o   o
1   2
``````

Measured data

``````Pins    Old         New
A-1     107.1 kΩ    0.634 MΩ
A-2     106.1 kΩ    23.22 kΩ
B-1       0.2 Ω       0.2 Ω
B-2     1.008 kΩ    0.614 MΩ
C-1      7.94 kΩ    1.212 MΩ
C-2      6.94 kΩ    0.601 MΩ
A-B     106.9 kΩ    0.634 MΩ
A-C     112.8 kΩ    0.607 MΩ
B-C      7.96 kΩ    1.212 MΩ
1-2     1.008 kΩ    0.611 MΩ
``````
• Your situation is strange to me. Most resistor packs fail the other way. Most fail on the low side and only allow full power on the top side. This is because there isn't a resistor on the top side so the fan gets full voltage. Commented Jun 13, 2015 at 23:29
• @Paulster2 : I'm not interested in the symptoms of a bad resistor here as that may vary across vehicles. With this question I was hoping to understand what happens to the resistor over time and why that happens.
– Zaid
Commented Jun 14, 2015 at 12:55
• I think your best bet is to google how resistors fail. Final stage resistors are resistors just like any other, and will suffer the same failure modes. The EE stack exchange might also be a great place for this question. Commented Jun 14, 2015 at 15:10
• In a word heat. They burn out over time. Commented Jun 14, 2015 at 15:31
• WOw, that answered my question perfectly. In short, it may not be the FSR, but in an older vehicle, say 20 years old, it maybe the blower motor brush contacts. Thus a new FSR will only temporarily fix the problem and require a new b.motor instead. Thanks. PS. This seems to be a popular issue with BMW products. Commented Mar 7, 2020 at 0:19

There is no real point in doing a resistance measurement.

BWM, like with most other stuff, decided to put more in there than you need.

They have put active elements in there, that control the current through the motor based on a control input signal, this is nothing to do with a bad resistor, but with some failure in the transistor-drive somewhere inside it.

A resistor pack would likely be designed to allow at least off and full speed as options, but with a transistor in the final current path it is much more likely to offer no more high-current flow.

A first hint that your measurements might not be meaningful is the fact all, except the internal ground connection between B and 1, are in the kilo-Ohm range. at 12V a single 1kOhm resistor would only deliver into a short circuit a current of 12mA. And it would then only dissipate 0.144W, which requires no heat-sink at all, let alone such a huge block of metal.

That's not to mention the motor will likely still want to have about 6V, so that would reduce it all to: 6V across the resistor divided by 1kOhm, is 6mA. Giving the motor only 6mA multiplied by 6V, is 36mW. If you can refresh anything more than an ant with that, I'd be amazed.

Once you pass the 10's of Ohms in a 40A-fused motor circuit, you can stop measuring resistance, because that's not going to be the thing you need to know.

You can't measure anything useful with a simple multi-meter, because the control input would need to be actively controlled to turn on the internal transistor stages, which doesn't happen when you randomly measure signal pairs.

Why a transistor stage fails like this, to be honest, is bad design. What often happens when the substrates in a transistor start to overheat is that they diffuse lightly into each other. Some transistors fuse "on", but most fuse to "nearly no current gain", as their thin internal barrier gets merged with the upper N-layer, and the thin barrier effect gets nullified.

So the thing to do if you design something like this is design it for a continuous load larger than what the fuse would allow. And include some output stage protection for excessive sparking in the motor. The latter may have been what killed it, in which case: Check your motor's carbon brushes, or it might happen again soon.

• @Zaid No problems. Procrastination for the win :-) Commented Jun 16, 2015 at 21:11