Valvestate AVT 150 o/p chips failure

[oc disorder]

Hi there

Yes I'm stumped at the moment with this Marshall Valvestate 2000 AVT 150.
Came in with stories of someone attempting to plug the output into a mixer -
speaker out instead of emulated out so basically a short on the output.
I ordered in two new TDA 7293v and set about removing the old chips and cleaning
the little pcb's ready for the new chips.
While I was waiting checked the power supply via a light bulb limiter - was approx
plus and minus 43volts and inspected the board with a jewellers eyeglass for any damage.

Having already repaired one of those mode 4 Marshalls with a similar fault (all 4 o/p chips replaced)
and observing that the power supply has to be discharged before connecting them up I thought
it would be a straight foward repair but no the light bulb limiter started glowing a bit more and
I hit the off button pronto.

Post mortum revealed major internal shorts on both chips (eg positive supply pin and negative supply pin
shorted) no signs of smoke or any signs of damage anywhere !

So ordered 2 more chips and same again !

I have been very careful to make sure there are no signs of solder splashes on the board.
Cleaned it with isopropyl alcohol and a small tooth brush then checked it with a jewellers
eyeglass again for shorts ... both times.

Checked for DC on input at the connector the little output boards plug into and using a sig gen and scope
saw that there was signal at the emulated output and at pin 2.

Read somewhere another tech had a problem with the bootstrap loader pin not connected so checked
the two caps with a meter C216 & C191 22uF 50v that connect between the master and slave.
They both are not shorted and charged/discharged in a similar fashion with my meter.

I did notice a tiny bit of ripple on one of the supply rails and at the moment am stumped !

Reading stories on line about replacing that setup with a more traditional bipolar o/p stage almost make sense !!!

Any help greatly appreciated.

Attaching a butchered um altered schematic with the parts for the AVT275 in a light blue (not fitted) showing the
complete output stage but only showing one module of the identical pair.

[Alex R]

Connecting something grounded, like the ground lead of your scope probe, or a mixer input, to the speaker jack, of a mode 4, will cause the output chips to blow. Maybe the same with this amp?

[oc disorder]

Connecting something grounded, like the ground lead of your scope probe, or a mixer input, to the speaker jack, of a mode 4, will cause the output chips to blow. Maybe the same with this amp?

Hi Alex no the IC's are not in bridge mode thus the output ground is chassis ground.
Both times turned on with no load.
I've ordered 4 more...

[Alex R]

Ah ok, best of luck then. Watching for dc on the outputs whilst bringing it up REALLY slow on the variac with a light bulb limiter might enable you to look for shorts unsacrificially, worked for me anyhow.

[oc disorder]

Ah ok, best of luck then. Watching for dc on the outputs whilst bringing it up REALLY slow on the variac with a light bulb limiter might enable you to look for shorts unsacrificially, worked for me anyhow.

Thanks Alex I guess I will be doing that and btw I missed the .22ohm 10watt wire wound that exists between audio output ground and chassis
but I did have the ground of the probe on the chassis and not the output jack ground.

[woodyc]

I had to buy some of those 7293 boards yesterday too. The price doubled from the last time I bought them -- which was only about a month ago.

[oc disorder]

After destroying 3 pairs of TDA 7293's (helping the economy recover?) and reading comments
about replacing the output stage with a more traditional design this went in the "too hard basket"
until recently.
I phoned a friend who looked at it and said without a ballast resistor it's doomed to fail so I got
on with other things.
I had been particularly pedantic about cleaning the boards after removing the ruined 7293's but
was wondering how many more times I and the pcb could cope with the desoldering and resoldering!

I also was aware of the necessity to completely discharge the power supply before plugging them back in.
I also noted in the data sheet for the TDA7293 that a ballast resistor wasn't necessary.

I have the luxury of an esr meter and had checked the main power supply capacitors and the surrounding
caps to the IC's which in my opinion were still in the ballpark.

However my friend suggested just plugging one in and see what happens rather than the two in parallel.

I decided to replace the main filter caps which incidentally had a better esr reading than the ones I was
replacing them with ... and the two decoupling caps on both boards (47uF) and finally the 22uF 63v caps
which I noted were 105 deg type (as distinct from all the others) on the main board (bootstrap) near the
multipin sockets where the o/p chips plugged in.

I made a mental note that Marshall had deliberately put 105 deg caps only in this position.

Then I reassembled the amp to try it with one chip and it worked!!!!!!

So I then discharged any cap that had any indication of a stored charge and plugged in the other one
and tried it... and success!!

My conclusion was that a cap was not performing properly when power was applied but I'm not going to swap
them back to find out which one !

I noticed the other day a schematic for a revised output stage for a AVT 150x ? which is a more traditional
mosfet o/p amp so I guess I am not the only one to have trouble !

Perhaps sometimes its best not to think too much and "just replace the caps man" !

Am posting my doodles which could be inaccurate but give a better picture of how things are connected
as some modern circuits require a GPS to navigate them !

Also posting some pics for reference.