To be your first schematic, is very well drawn.

thank you i've only replaced components on boards before so everything here is completely new to me haha.

 

I suggest that you figure the power dissipated in the transistor, and then use a heat sink rated for that much power. It will probably be rather big.

 

I suggest that you figure the power dissipated in the transistor, and then use a heat sink rated for that much power. It will probably be rather big.

will do so. theres alot of space to go around so hopefully that wont be a problem. do you think two IXTQ60N20L2 (together with resistors and a diode) would be a suitable replacement?

 

I have not investigated the specifications of those, nor of the darlington device currently used in the module. The values to check, aside from the current ratings and the temperature derating.
And what is puzzling me is the voltages on the CA-4 pin for the different speeds. or are those the average of the PWM signal as seen by your meter in the DC volts range. Using the transistor in the linear mode will result in huge amounts of heat being generated.
So switching speed for the replacement device will also matter a great deal. It needs to go from cutoff to fully saturated rapidly, and switch off just as fast, to avoid time spent generating heat while it is in the linear mode.

 

MisterBill2, As there is a 22uF capacitor between base and emitter the transistor can't be working in switching mode so as you say it will generate a lot of heat. The way the voltages change on AC6-4 and AC6-2 with speed seem to be the wrong way round.
The voltage on AC6-4 would need to increase to increase the speed. (making the transistor pass more collector current.)
The feedback voltage on AC6-2 would decrease for higher speeds. (I am assumings these voltage readings are with respect to ground. IF theywere with respect to +12 volts they would make sense.) From memory the maximum collector current is 20 amps average. (Peak 40 amps.) The IXTQ60N20L2 is an N channel mosfet with a continuous max drain current of 60 amps (150 peak.)
The fact the the gate is votage driven and the transistor base is current driven may require modification to the circuit that drives signal AC6-4.
samili, using a mosfet would only need one device as almost no current is required to drive the gate.

Les.

 

Consider the possibility that at a lower speed the transistor is dropping six volts, and the current is ten amps. THAT is SIXTY watts of heat produced. Considering that one of my hotter soldering pens is rated at 30 watts, 60 watts is a lot of heat energy and will tend to cause a problem with temperature rise. So what makes a lot of sense is investigating what you can do to change to a PWM controller, still interfacing with the same control voltages for the lower speeds.

 

As MisterBill2 suggested in Post #7, determine the cause of the darlington failure. Keep in mind that the design has been successful in tens of thousands (or more) cars, so the design is at least 'not bad'. My money is the failure mode is excessive temperature for the transistor.

Knowing how "cost conscious" the auto industry is, the transistor may well be mounted in the air stream to avoid the cost of adding a heatsink. It probably works OK when everything is new and the air stream is free and clear. Add some dirt and a bit of corrosion after several years and the transistor may now run too hot.

Rather than re-designing the entire system, consider keeping the original darlington type and perhaps adding a slightly larger heatsink in the air stream. The "forced air" cooling will make any heatsink much more effective. Perhaps start with a simple 1/16" thick U-shaped piece of aluminum with 1" square fins (use thermal grease on both sides of the heatsink).

If you try to move the transistor out of the air stream, you will wind up needing a HUGE heatsink that has enough capacity to keep the transistor within its limits on a hot, sunny day when the car interior temperature can reach 150 degrees.

 

Certainly adequate heat removal is vital! AND, given the comment about corrosion, very early in the thread, the application of a good conformal protective coating to both sides of the circuit board is what I suggest. That being in addition to possibly a better heat sink in the air stream. AND, if there is a similar transistor with a better power handling rating, or a military temperature range, or both, that could be a longer lasting choice.
AND still a check to see if something has changed in the system. Faults are some times rather subtle, and not at all obvious.