a bit OT:Repairing brushless DC motor?

Just thinking aloud, the following may or may not be relevant ....
1. I can see 2 off TIP137 which are PNP darlington, are the other two also TIP 137, or the complementary NPN TIP 132? (I'm assuming that for a brushless DC motor, the field coils will be commutated by an H bridge)
2. Is it a Stepper motor, which might also be suggested by the number of leads?
3. You mention corrosion; my experience is that many electronic faults can be cured by touching all joints with a soldering iron, to resolve dry joints.
4. If one of the power darlingtons was short circuit and holding the field coil in one polarity, or open circuit, and not permitting the dirction of the field to change, this might produce the effect you describe.
5. Look at the collector voltage of each power transistor, this might give you a first indication of the state of the field driving .
6. This is a bit like a GP diagnosing your illness by you describing over the phone the colour of your urine :-)
A customer has a failed 24V heating circ pump on his boat. It's based on a standard domestic ('Grundfos type') central heating pump. The mains motor has been changed for a (presumably) brushless DC motor He's enquired about getting it repaired, and has had answers varying between 'No' and 'We can sell you a new pump for £500'.
The spindle turns freely with the power off, putting on the power holds it in one position.
Here are some pics of the electronics, looks pretty basic. To my untutored (read 'behind the times'!) eye appears to be four small power transistors, six smaller transistors(?) and a few standard resistors & capacitors. There are three sets of leads going from control board and disappearing into the motor case apart from the red & black pair (power in). Five leads which look to be the main motor leads. Six smaller leads which I'm guessing must come from a tacho/ encoder. Three small leads, maybe speed control pot.
Pics of the innards:-
I'm looking for suggestions as to what checks I could make to track down faulty components, also is there a more or less 'standard' circuit for this sort of controller? If the motor windings are OK I reckon it should be within my capabilities to repair the circuit board as it's all discrete components. Nothing is obviously cooked, there's a little bit of corrosion on the PCB tracks. I have decent multimeters and even an old oscilloscope hanging on the wall, not been used for a while but probably still works. Alternatively, a pointer to someone who could repair for sensible money, preferably in the NW, would be useful.
Thanks for any help Tim

Can't really fault Phil's analysis. Just to add that something should be driving the H bridge in 4 sequential states to make the motor turn. If all the drivers look OK then the sequencer is stuck for some reason. it might be a 555 timer driving a counter or flip flops or these day just as easy to for the designer to use a PIC micro controller to run it.
I see what might be a shaft of a pot in the pictures, is this a variable speed control? if so that might be shagged??
If I was given this job I would be charging a fixed investigation fee with the option to give up if analysing it gets too time consuming.
Good luck
Bob

The power transistors are all TIP 137. There's no evidence of anything more complex than transistors & diodes in the circuit. It's not a speed pot, it's a 3-position speed switch. just found this on the web, this is the same pump:-
I've emailed them to ask if they can recommend a repairer in the UK.
I'm just 'playing' at the moment out of curiosity, there'll certainly be a bill if I'm successful though.
Tim

i probably know less than you Tim ..
but aren't those transistors supposed to have insulating washers and heat conducting plastic sheet behind them...i cant tell in the pics ..but it don't look right .
all the best.markj

I think you probably mean Mica washers.
As Tim has informed us that they are all the PNP TIP137, then it is possible that they're being driven inverted from a +ve rail, in which case all the Collectors (the tabs) will connect directly to ground without any insulation required.
Having looked at the web site that Tim found, what I now guesstimate to be the case is that there are 4 four field coils to be energised in sequence, each one being pulled to ground in turn by the darlingtons. This would account for the high number of motor leads.
(These Darlingtons tend to have inbuilt suppression diodes, thus accounting for the very few diodes to be seen on the PCB).
My remote diagnosys still tends toward one of the darlingtons having failed short circuit; or else one of the (presumed) open collector outputs of the encoder being shorted to ground, thus generating a permanent drive signal. Unfortunately without further info on the motor windings, or on the nature of the commutating encoder, it is difficult to suggest further.
but aren't those transistors supposed to have insulating washers and heat conducting plastic sheet behind them...i cant tell in the pics ..but it don't look right .
all the best.markj

I've done a bit more prodding.
The collectors/tabs are screwed directly to a common heatsink, which is deliberately insulated from everything else. There is what appears to be a star point for the motor windings which is permanently connected to supply +ve. Then the other end of each phase goes to the individual emitters on the power transistors. Without disconnecting anything, I've checked (with DMM) resistances across all the various combinations of power transistor connections. All are within 10% of the other transistors, some much closer. Is there any mileage in checking the junctions with the diode test function, on the basis that the voltage used may be higher and enough to force forward conduction? It's all a bit of a black art to me, I got a bit lost after electronics involved glowing glass envelopes. The winding resistances, again still connected, are all within 2% of one another and the same in both directions.. The central 'driver' transistor is marked CG39 w12. I can't see any markings on the others.
Many thanks Tim

Based on what has been discussed is it possible that the motor is seized? As there appears to be no electronic commutation it is valid to presume that the commutation is via the encoder. If the motor is seized then one phase on and nuffin goes nowhere
Mechanical you can do Tim
Richard

"The spindle turns freely when the power is off"

. No, not seized, it turns quite freely.
Regarding the encoder/Hall sensor or whatever it is, there are three pairs of wires - two red, two black, two white. The two reds are connected together, as are the two blacks, the two whites go to different places, one to somewhere within either set of three small transistors. The two reds appear to be just connected via a diode of some sort to the two blacks, nothing else, while the two blacks go to the common rail with the supply -ve/collectors. Does that make any sense, and is there some way of checking the encoder/hall sensors or whatever?
Thanks again Tim
Ta Tim

My apologies, I was following the recent posts and did not recheck the original.
Returning to lurk as I am intrigued
Richard

I regret that I am now at the stage where I'd require you to trace the circuit diagram from the PCB before attempting any more remote guesswork :-)

I've had a go, here is my attempt at one half of the circuit, the other half is almost identical:-
I've referred to Hall effect sensors. For all I know they might be something different.
I've assumed some components to be capacitors, and shown them as such, they may be something else. They're the orange ones with black bands as can be seen here, three of them roughly in the centre:-
And one here:-
Thanks for any help
Tim

Not had time to interpret fully your cicuit, but the components that you query are diodes.
What I can see is that there are two fairly similar driving circuits in the two halves, with one half having an extra inverter so that the two coils are driven in anti-phase (ie, when one is on, the other is off and vice -versa) to produce the commutating effect
How does the cct change if the diodes are drawn on it?
(The black bar marks the cathode of the diode, normally drawn as a transverse bar; the anode is drawn as an arrow pointing to the black bar)

I've altered it here, with a couple of corrections:-
If diodes, they must be Zeners or similar??
BTW the rail in the centre of the sketch, with the various diodes connected two it, is shared with the other half of the circuit.
Thanks Tim

1. The sensor is probably an opto, not a hall, with the black and red being the drive for the LED and the white being the phototransistor return.
2. Something is wrong because the "Speed Rail" is not connected to anything.
There should be a resistor from the Speed rail to the -ve rail, otherwise there is no source of emitter current for the c639's
3. Most of the diodes are to suppress the Ldi/dt voltage spikes when the inductors are switched off.
4. Can you read off the part numbers of the "zener diodes"?
5. I now understand (subject to cockiness :-) ) sufficient of the curcuit to confirm my original diagnoses. One of the TIP137 is either on permanently due to a faulty drive, or else it has failed short-circuit.
6. A faulty drive might be caused by the immediately preceding transistor also being short-circuit.
7. A faulty drive might be cause by the "extra" transistor in the sequence of 3 on the right side failing open circuit.
8. Finally, a faulty drive might be caused by the photo transistor in the opto failling short circuit.
9. If the problem does lie in the area of a dry joint, then it'd have to be somewhere around that "extra" transistor on the right.
10. What you need to do now is to disconnect either of the TIP137 to motor connections and connect an ammeter in series to determine which one is drawing current, and then to correlate it against the cct diagr

Not sure whther that would be possible, AIUI the motor rotor runs in water and is sealed from the windings. I may have that wrong, though.
If the diodes are Zeners, will it no be supplied via the bottom diode from the + rail (not a zener) and the third ?Zener from the left?
I'll try, might not be for a couple of days.
Thanks Tim

If optical, then there'd be some form of slotted disc attached to the rotor that would pass between the opto components. The opto components could then well be sealed and isolated from the water.
Thinking again on this, if a Hall sensor, then it would incorporate its own amplifier ahich would account for the red/black feed wires. So, it could either be a Hall effect or an opto.
No. Something has to be there to pull the speed rail up (in a graphical sense to the negative rail. The emitters of the c639's have to be pulled up towards the negative rail.

OK, you're absolutely right. Just checked, there is a resistor which I'd missed from my sketch, between -ve and the 'speed rail'.
Tim

It's just occurred to me that the problem may yet be a mechanical one, without there being any fault within the electronics, if the sensor mechanism (in whatever form it takes) has somehow moved circularly with respect to the coils, or if the shaft-mounted components of the sensor mechanism have moved circularly with respect to the rotor magnets.
Think about the steam engine - there has to be a phase offset between the driving piston and the valve position; in the same way, the sensor mechanism has to be out of phase with the rotor magnets.