Mark Bass Combo Head II SMPS Woes

[tdlunsfo]

As the title suggests, I was begged to try my hand at fixing a Mark Bass and I reluctantly agreed even though I have only worked on a handful of class D, SMPS designs. Apparently nobody else in town would touch it and he was desperate. So... I'm the sucker...lol.
Findings: Blown mains fuse and shorted Mosfets in the power supply (T22/T23). It seems this is pretty common when these amps fail and simple enough to replace but here is the rub... I can't find anything else wrong with it. While in circuit, the Final Mosfets test good as well as the dual diodes in the secondary D24-27. I've checked all around this board and everything checks out. No other shorts. So, I replaced the mosfets as well as the driver chip, fired it up and pop... back to square one.

So, obviously this amp still has another failure somewhere that I am missing. Any thoughts/experiences? My next step was to start removing the finals and dual diode and test them out of circuit just to be sure. I guess I could reorder the parts, disconnect the F1 fuse and the +/-80V supplies and fire up the SMPS on my variac to see what happens (probably what I should had done the first time but we live and learn). I stole the schematic below from another thread so please ignore the X's, or at least the X on D25... for now.

[Jon Snell]

IC5 is a High and Low side driver, designed for stepper motor driving.
There is an internal oscillator controled by R74 and C72. They are not likely to have failed.
Replace the IR2153D and before replacing the output FETS, check the waveform on both High and Low side drives.
If the do not switch with at least a 5nS dead time, find out why not and then fit the new output FETS.

[tdlunsfo]

I guess my point is when I replaced these parts and fired it up, it lasted about 2 seconds. I didn't have time to check anything.

[Jon Snell]

The idea of fault finding is to find out what the cause of the failure is before replacing parts. A good solid understanding of how the SMPSU works is vital in attempting such repairs.

If you remove C68, that will disconnect any external faults, bearing in mind it won't work for long like that as R78 can only supply enough current to start the oscillator and drive the FETs for a short period before D40 takes over the supply. Use a stabilised power supply to check the time constant of the output drives. You require at least a 100MHZ dual channel storage oscilloscope, to view the drive wave form.
As you have found out, there is very little time to fault find.

Mark Bass will probably not accept it for repair if a repair has been attempted by other than their service agent.

PS. Check R76/77 if one has gone high value, the FET is not turning off fast enough.
If there is a short on the secondary side, it will run into over current and limit the drive. It should not go bang.
Good luck.

[tdlunsfo]

Jon thanks for the reply. R76/R77 are dead on, in fact it looks like they have been replaced previous, along with other parts. Although the owner wasn't aware of any previous work, it likely would've been rejected by the factory anyways. I actually brought this up to him and he sidestepped the idea. I have a scope and I have been searching for faults for hours to no avail, hence my post. I have a basic understanding of SMPS but lack the experience. A couple of other notes, R78 was found blown the first go around (it survived this one) and it looks like C69 is measuring 100u instead of 10u (esr still tests fine), but I'm having trouble seeing how that would obliterate a 7A fuse.

[tdlunsfo]

Use a stabilised power supply to check the time constant of the output drives.

Jon Snell , just so I'm clear, are you referring to hooking up a seperate power supply to finals?

[Jon Snell]

Put a 10 volt supply on pin 1 of the IR21531 but ONLY if the MB is isolated from your mains.
That will power up the oscillator and driver stage safely, as it will think there is mains power on it, without blowing it up again.
Then measure the signals on pins 7 High side and 5 Low side. As the High side switches on the Low side must be off and vica versa.
C69 measuring 100u off the board? That won't take out the mains fuse. Short circuit FETs will.

[The Dude]

Agree with Jon's assessment, I'll just add: Also check resistance from supplies to ground to make sure there isn't an external short on one of the supply lines.

[tdlunsfo]

When I started the amp up last time I had a pretty high pitched squeal before it failed so not sure how to break that down. I pulled all the finals and dual diodes to test them out of circuit and they are fine. I tested all the resistors and caps surrounding the IR2153 and they were fine, including the one cap that measures 100uf in circuit. I'm at a loss.. there has to be something that has failed to make this amp pull this much current but I can't find it. I am starting to think it is the main transformer. I called the service center in California and they said these boards are available for $300 shipped but backordered from Italy and they haven't received a shipment since summer. I've gone down the shorts route as well but that all seems fine both to circuit ground and chassis. I'm tempted to spend another $30 on parts just to run Jon's test and learn more about this setup but also tempted to send it on it's way and move on.

[Jon Snell]

When I started the amp up last time I had a pretty high pitched squeal before it failed so not sure how to break that down. I pulled all the finals and dual diodes to test them out of circuit and they are fine. I tested all the resistors and caps surrounding the IR2153 and they were fine, including the one cap that measures 100uf in circuit. I'm at a loss.. there has to be something that has failed to make this amp pull this much current but I can't find it. I am starting to think it is the main transformer. I called the service center in California and they said these boards are available for $300 shipped but backordered from Italy and they haven't received a shipment since summer. I've gone down the shorts route as well but that all seems fine both to circuit ground and chassis. I'm tempted to spend another $30 on parts just to run Jon's test and learn more about this setup but also tempted to send it on it's way and move on.

You should not 'test' components in sittu, the measurements are skewed as you have found out as I expect you are measuring both capacitors at the same time and everything else connected to it.

[tdlunsfo]

Main electrolytics were tested out of circuit. Primary side filters are a little low..900uf, but rest are good. All have low ESR. I am aware of the issues with testing caps in situ.

[x-pro]

Why was the D25 broken?
Check the -80V circuit elements for short circuits.
Check for a short between the transformer windings on pins 7 and 14.
Check BR1.

[tdlunsfo]

D25 tests good. I lifted the schematic from another post.

[tdlunsfo]

BR1 checks good with diode test. I don't see any shorts to ground on the 80v rails and no shorts on the transformer primary to secondary.

[x-pro]

It is necessary to unsolder and check the capacitors C39 and C40 on a suitable device.
Check RT2 2.5 ohm, C68. Temporarily unsolder diodes D24, D25, D26, D27.
Install the burned out elements to the new ones, of another batch (of another store).
Check the operation of 21531 from an external power supply, by controlling the outputs of this microcircuit by an oscilloscope. To do this, supply 12V power to C70 according to the polarity.
If everything is normal, connect the whole circuit to the network, in the gap of one of the network outputs (it is possible instead of RT2), through the incandescent lamp of 60-100W in series