Heater Rheostat

[KevR]

You know, the wossname that lives in the fan blower body with the aluminium bit in the airstream.

What does it actually do...?

Just had my blower apart to clean the motor brushes and commutator, and changed the rheostat because a: I had a spare new one on the shelf, and b: it would save me taking it all out again if it turned the rheostat was actually the problem. Fan is now whizzing away happily but it always bugs me when I don't know how/why something works, or in this case, what it's even for!

[mat_fenwick]

It's a power transistor (IIRC a Darlington type) - the dial in the car controls a small current and allows a much larger current to flow through the transistor to the motor. More efficient electrically speaking than dissipating power in a resistor pack. The collector of the transmitter is connected to its metal case, which it relies on for contact. So it can often be simply corrosion around its securing rivets causing a poor connection - don't bin the module as it would be handy to have a refurbished one on the shelf!

[KevR]

What is this word, 'bin'.....? Thanks for the explanation - so it could indeed have been the rheostat that was at fault. I'll have a look at the old one and see what it tells me.

[Vanny]

Is it not a PMOS rather than a FET? I've never googled the component, but if it is a FET, then a PMOS might be a good upgrade?

[mat_fenwick]

The only one I matched up part number wise (I've just checked) was a p type Darlington transistor - bigger gain but slower response and higher base/emitter voltage. It would need someone cleverer than I to say whether you could substitute a suitable pMOS in the circuit but would think it would be significantly more efficient. Too late at night and too much wine has been consumed to start circuit design now!

[rutter123]

it controls the speed of the heater blower motor by wizardry and witchcraft, or goblins.
tho known to be a bit troublesome and quite pricey to replace, not always the problem motor brushes are quite commonly found burnt away or just the fan jammed with leaves and crap.
can these old duff transistor units be repaired in any way?

[electrokid]

Goblins it be...

Previous topic...

http://www.bxclub.co.uk/forum/viewtopic ... 79#p124507

2N5879 datasheet...

http://www.datasheetcatalog.org/datashe ... 2N5882.pdf

MJ11012 datasheet...

h[url]ttp://www.onsemi.com/pub_link/Collateral/MJ11012-D.PDF[/url]

2N5879 from China... (I've used this seller - ok)

http://www.ebay.co.uk/itm/2N5879-Manu-O ... 19d1d413a2

MJ11015 from UK - good seller...

http://www.ebay.co.uk/itm/MJ11015-Trans ... _930wt_725

MJ11015 from the UK...

http://www.ebay.co.uk/itm/MJ11015-PNP-1 ... 1593wt_962

MJ11015 from the UK - excellent seller...

http://www.ebay.co.uk/itm/MJ11015G-PNP- ... _246wt_962

[mat_fenwick]

I'm going to write a guide, with pictures on how to test and repair. Only trouble is that the 'back to back diode' test on the transistor isn't showing what I'd expect, which I now understand is typical of a Darlington pair. Brian, do you know any way to test a Darlington transistor with a multimeter?

[electrokid]

Well if I ever knew then I didn't commit it to memory, However, checking a PMD11K80 I happen to have around shows, as no doubt you have found, something I didn't expect but at least one way round the results are logical.

I'd expect one measurement (base to emitter) to show st least 2 diode drops - so around 1.2 volts - it actually shows 1.76 volts. Leaving the same test probe on the base and moving the other to the collector shows 'OL' which is IMHO to be expected.

Measuring the other way round with the other probe on the base it measures like a normal transistor - one diode drop each way - around 0.63 volts.

[mat_fenwick]

Hmmm, well I'm getting 0.64 V with -ve on base, +ve on emitter, and open circuit the other way round to start with. But after a few seconds getting a base to emitter voltage of 1.92 V! The base to collector junction I'm getting what you'd expect from a normal transistor, with a voltage of around 0.6 V. But collector (+ve) to emitter (-ve) I'm getting a voltage of 0.5 V, when it should read open circuit both ways.

When I get back home I'll compare it with a good one to see what differences there are...

[electrokid]

I'm getting something similar - even tried a PMD10K80 in the hope that a NPN device would be easier on the brain than those upside-downy PNP thingys but it was not to be.

I think I may have found the reason for the wierd values - built in bias resistors and / or protection diodes across CE as in...

http://www.cougarv6.de/imrc_transistor_check_en.pdf

I don't think there's any need to use a darlington though since the straightforward PNP device works ok but as long as the gain and max collector current is sufficient then anything should be ok. Cooling of the transistor is because it's in the airflow created by the fan of course.

[electrokid]

But after a few seconds getting a base to emitter voltage of 1.92 V!

I think that was because the voltage was varying slightly and 1.92 is very close to the max reading on your meter of 1.999 to confirm you could connect another meter to the same test probes set to read volts.

Because all these readings seemed quite wierd I decided to do some more digging so today I wired up 2 transistors in darlington configuration and did some measurement.

Because there were no other components to confuse the readings they were entirely logical. I used NPN devices so probes would need to be reversed for a PNP device.

With +ve to the base there were 2 diode drops to the emitter (1,114v) and 1 diode drop to the collector (0.581v). All other measurements were open circuit as they should be - C to E measurement is only used in faultfinding since that should be open either way.

To replicate the protection diode I used a zener which would not only protect from reverse volts but provide some protection for overvoltage spikes as well - connected the same as a normal diode - anode to emitter / ground, cathode to collector.

Measurements again as expected - Vbe and Vbc were the same, Vce (+ to collector) was o/c but Vec (+ to emitter) was 1 diode drop - 0.6v.

The only strange reading I got was when I disconnected the collector of the smaller drive transistor (which is likely to be the most vulnerable device and so more likely to be blown open circuit) - Vbe was unchanged but Vbc was now slightly lower at 1.09v (measuring to the collector of the larger transistor).

I think checking a specific device needs comparison with a known good device and / or checking the datasheet for any additional components that could confuse the measurement.

For my test I used devices that could form a useful combination for future use - the small drive transistor was a 2N3439 and the big jobby was a BU208A. The diode was a 1N5388B. BU208A specifications vary between manufacturers - this was an ST device spec'd at 8A continuous / 15A pulse. The diode is a 5 watt / 200v device but a diode up to 350v could be used without changing the other devices (2N3439 is good for 350v and the BU is happy up to 700v). The current gain was quite good - 1390 - which means it could be driven from any logic source - minimum 7mW drive with control of 1600W with a 200 volt rail - power gain of over 200,000 - might come in handy

[mat_fenwick]

Interesting! As it happened, the fault with my new arrival was simply a poor connection to the transistor case. I couldn't find another one to compare it too anyway, although I know I have a FW26925A1 somewhere! I removed it from a working module (to upgrade to a 30 amp transistor) when I did the AC conversion, as I binned the apparently failed module from the donor car. The data sheet for the FW26925A1 shows a diode collector to emitter of the package, and a couple of internal resistors...

[electrokid]

That's good - glad you've got it working. IMHO it's quite a good design - if the transistor has current flowing through it then the fan is very likely to be running so there will be airflow to keep the transistor cool. Everything under that grille gets a bit damp so the electrical connections all suffer.