Root cause insight into the common BMW blower motor resistor failures

I don't think Fluke makes much in the way of current probes for DC with appreciable sensitivity. The i1010 is sensitive enough for this sort of job, but not for a lot of other things you might want. They do make some fancy intelligent probes but they're all more expensive than just buying a Fluke clamp meter like the 365.

That said, I have a cheapo Extech 380947 and it's not built like the Fluke but it's sensitive enough to detect small ground leakage currents.

--scott

No, there are two transistors bolted to the heat sink. That's why the heat sink is there, to cool those two power transistors.

No, the spacing is consistent with a TO-220 transistor pair, and if it's a linear pass regulator like it appears to be, there needs to be a big transistor somewhere. Also, of course, there is the heatsink.

That IC is only control logic, it just takes some mystery input signal and produces a variable voltage for the transistor base. Those two transistors are doing all the hard work. BUT, if you want to replace the device with a retrofit one, you need to know what that mysterious input signal is.

--scott

It's not a switcher. And those transistors may not be MOSFETs. But there is no reason not to use mosfets in linear mode, other than the fact that no two off the line have the same gain or transconductance.

Agreed.

--scott

i understand that - and variable voltage is the problem. the secondary [bordering on primary in some applications] advantage of pwm is low speed start and torque. if a motor starts at low dc voltage, not only is the start speed inconsistent, it has little torque. pwm can start a motor slower and at much higher torque. it's a big deal.

way to understate - those things are insanely expensive. otoh, there's almost no competing product, so they have the ability to leverage pricing.

that's why i settled for the esi unit. not as good as the fluke, but has similar sensitivity in a package robust enough for automotive use.

>

if you want your purchase to be useful for automotive, you want sensitivity. and ability to hook it up to a scope is a huge plus.

there's been another post - the output has been scoped. it's linear.

besides, to-220 isn't reserved for fet's.

now you're guessing.

you're not telling the difference from a simple visual. sorry.

One key question in all this is if the interface from car to "resistor block" is some kind of simple digital interface, ie it sends some bits that get interpreted as "go to speed 3" or does it send a PWM signal. My guess is the latter. That's my understanding of what BMW does with the aux radiator fan in the X5. Another German electronic miracle that fails and in doing so, mysteriously drains the battery.

Someone should put an oscilloscope on this and find out what the signal looks like. If it sends a digital code, then making a replacement from scratch is a big hurdle. If it's sending a PWM signal, they you could build an equivalent from Radio Shack parts. It still seems like more work than it's worth.

How fast are these things failing for those that want to make their own? 2002 X5 here and it's only had this problem once, about 2 years ago and replacement one is still working. And another data point. The failure on that X5 resulted in the blower draining the battery when the car was off. Blower ran fine. Only odd thing in retrospect was that when the car was off, a couple times I heard a faint noise. In retrospect, it was probably the blower getting just enough current to start to turn then stop. And only noticed it a couple times. The bad thing with the failure mode of this and the AC fan is that both were draining the battery and both were very hard to pinpoint, resulting in huge labor charges.

The semi-conductors circled are not the MOSFET you are looking for information on. The Mosfet is 14 pin and what is circled is something different.

I think if these Mosfets were in 1000 BMW owner's hands, maybe one person out of 1000, if that, would have the skills to install in the carcass we have left after removing all the potting material. How are we determining the MOSFET is bad in any situation? Again, even with a data sheet which we will never get because these Mosfets are supposedly "special" to BMW (just like they try to hide their MAF sensor). Few would have the skills to test a 14 pin Mosfet. I believe the FSUs that have been "repaired" are again the usual "open" circuit BMW seems to like so much (can we say ABS module).

Ah, so it's a plain old voltage divider?

No. The power transistors (which might be mosfets) are those three pin TO-220 things bolted to the heatsink.

The 14 pin thing is an ASIC that is used to control the power transistors, and it's proprietary and undocumented but may not even be related to the failure.

You're getting ahead of things again. We don't even know what is failing.

--scott

All I know of the HVAC input signal is that it's a 2.0 to 7.8 VDC signal from the HVAC controller, presumably to correspond to the various levels of the fan blower motor speed.

All I know is that the HVAC controller sends a 2.0 VDC to 7.8 VDC signal to one of the five pins of the FSU. I presume that directly corresponds to the desired blower motor speed - but that is conjecture on my part.

The originals fail within about 5 years. I've had my second one fail in

3 years. I think we can safely say about 3 to 5 years is the lifecycle but nobody really knows for sure (least of all me).

This is one of the classic failure indications! Very very very common! However, another classic failure indication, other than the dead battery in the morning, is a blower that has a "mind of its own".

Together, those two sets of symptoms account for 99% of the failures.

Of all the anecdotal evidence presented, I don't think I've ever heard of a failure being that the system was totally dead.

What that tells us, I don't know.

Exactly!

I will try to unpot mine today, at lunchtime (if the boss will allow me).

I don't see any evidence that the 16-pin surface mount ELMOS 10901D (aka E109.01D) is the problem, yet. However it needs to be a suspect, just as anything else would - especially since the failure mode is not a go:nogo situation. However, the chip appears to be designed to be extremely robust, with a HUUUUGE voltage range up to hundreds of volts.

AFAICT, it is apparently a temperature compensated automotive voltage regulator, at least according to a google translation of this Russian web site:

A google reverse image search finds these potential sources of the chip (based on photos of the same part and part numbers): But, no datasheet.

Here is what Elmos Semiconductor had to say about it reputedly: [quote] Thank you very much for your interest in our products. Unfortunately we are strictly not allowed to provide any information concerning the E109.01D to you as this product is customer specific. Best regards, Sina Staufenberg staatl. geprüfte Betriebswirtin (Recht) [/quote]

But is that a clean DC level or a pulse width modulated signal? If it was measured with a multimeter, you won't know.

Has anyone probed around the module with a scope? Is that possible?

That and the question about the TO-220 devices. I think it would be possible to make a better replacement if those questions were answered.

tm

Your idea about testing it is correct, however a bad one can work fine until it gets hot enough for the IC chip to fail. Hvac control voltage is

0-8 VDC.

The load resistor should have a metal case mounted on a heatsink to dissipate the heat if you are going to run the test more than a few minutes.

My original 13 spike FSU was replaced twice under warranty, so the car had 3 FSU's. When the 3rd old style failed i replaced it with the new design FSU. That was 8-9 years ago, it is still working. One thing not considered in your post is how many did not use OEM units, the $75 eBay ones are junk. The $175 units at the stealer seem to hold up a little better.

Bear in mind that the $75 Sitronic Ebay FSU is known to be even more faulty than the $175 Valeo FSU from the stealer.

I can think of nice test equipment to buy instead of a $175 FSU every few years!

I don't even own a BMW. After this thread, I don't think I ever will.