Marshall AVT

[Valvehead]

2 questions about the Marshall AVT series amps-

In these amps , does the 7293 chip need a spacer/washer on the back side of it? or is it just thermal paste and screwed to the metal heatsink?


Has anyone come up with a "protection mod " , that will help these 7293 chips last longer ??

[Jazz P Bass]

The 7293 backplane is at B- potential.

If the heatsink is isolated from the chassis, then there is no need for a mica isolater.

[Valvehead]

ok thanks

[elipsey]

Pretty sure the one I repaired had spacers, and the heatsink was bolted to the chassis. I think the supporting circuitry for the tda7293 is inadequate in this amp, and this causes it to have reliability problems. I have read about others installing custom daughter boards, but I can't seem to find the link now. You might need to replace more than just the amplifier module or they can blow again. Also, watch out for pulling traces off that crappy board. I almost screwed myself the first try, and ended up cutting the chips off at the legs with a dremel to make desoldering easier. Here's another MEF thread about one of these (not mine): http://music-electronics-forum.com/t28444/

Sorry if I'm telling you stuff you already know.

[Valvehead]

so do i need and insulating washer with a shoulder to keep the screw from touching the chip?

i cant remember if this on had them , much less the isolator---i will have to look tonight

i was able to get 1 chip back in the 1st pcb but i hurt several traces, the other pcb is destroyed --so ill be ordering 2 of these guys =

Amazon.com: Marshall, Avt Power Board Kit Avt275x-: Musical Instruments




thanks

[elipsey]

so do i need and insulating washer with a shoulder to keep the screw from touching the chip?

I think so.

i was able to get 1 chip back in the 1st pcb but i hurt several traces, the other pcb is destroyed --so ill be ordering 2 of these guys =

Ouch, sorry to hear that.

[g1]

Is the heatsink mounted directly to the chassis, or is it insulated? Check resistance from chassis to heatsink.

[drewl]

The easiest way I've found for replacing the IC is to clip the leads with cutters leaving enough the grab onto with needlenose pliars as you heat the legs and remove them.

Then clear the holes with solder wick, open them slightly with a pick so the new chip goes in easy and solder the legs.

It doesn't hurt to replace the two filter caps on that amp board and check the other components while there.

I usually reflow the connector pins also.

[elipsey]

Is the heatsink mounted directly to the chassis, or is it insulated? Check resistance from chassis to heatsink.

I think he took it apart a while ago, and can't remember. It was a year or so ago that I worked on that amp, so I can't remember for sure either. The datasheet for tda7293 says TAB CONNECTED TO PIN 8, which is -V S (SIGNAL). Maybe if it's signal ground, it doesn't matter?

[g1]

That's a little confusing as they show pin 8 connecting to -V supply on their circuit diagram pg.13 of datasheet you posted.

[52 Bill]

The datasheet for tda7293 says TAB CONNECTED TO PIN 8, which is -V S (SIGNAL). Maybe if it's signal ground, it doesn't matter?

Pin 8 of the chip connects directly to the negative power supply. The tab is live and if there is no insulator the heatsink will become live as well.

If the heatsink is isolated from chassis ground like by the mounting fan, then having the heatsink be live doesn't matter. If the heatsink is directly, electronically connected to the chassis ground, then there has to be isolation hardware used when mounting the chip.

[elipsey]

Pin 8 of the chip connects directly to the negative power supply. The tab is live and if there is no insulator the heatsink will become live as well.

If the heatsink is isolated from chassis ground like by the mounting fan, then having the heatsink be live doesn't matter. If the heatsink is directly, electronically connected to the chassis ground, then there has to be isolation hardware used when mounting the chip.

From datasheet:

15 -Vs (POWER)
14 OUT
13 +Vs (POWER)
12 BOOTSTRAP LOADER
11 BUFFER DRIVER
10 MUTE
9 STAND-BY
8 -Vs (SIGNAL)
7 +Vs (SIGNAL)
6 BOOTSTRAP
5 CLIP AND SHORT CIRCUIT DETECTOR
4 SIGNAL GROUND
3 NON INVERTING INPUT
2 INVERTING INPUT
1 STAND-BY GND

My recollection is that mine had the tabs isolated. The heatsink was definately bolted to the chassis; you can see it in the pictures.

If "-Vs (SIGNAL)" is the negative power supply rail, what's "-Vs (POWER)"?

[elipsey]

That's a little confusing as they show pin 8 connecting to -V supply on their circuit diagram pg.13 of datasheet you posted.

I am confused also.

[g1]

"signal" here is not what we are accustomed to. Actual inputs are pins 2&3.
I think signal as used on the data sheet means "sense" as in "this is the pin that senses whether power is present". If it does not get connected to circuit power, the chip will not work. So pin 8 will be a -V rail and the tab or heatsink must be insulated from ground.

[J M Fahey]

FWIW TDA729x chips , whose body is the same as 2 x TO220 soldered edge to edge is NOT capable of dissipating what it takes to put out 100W RMS , period.

Yet many try and then amps are "un reliable"
Not only them, generally well considered and conservative Fishman has one mid sized Loudbox model putting out 100W RMS and powered by one TDA729x , this is also a regulat Service Bench visitor, and if you don't see more, it's because there are not that many to begin with.

Same some Crates, Behringer, varuous powered speakers/monitors , so it's not just a Marshall problem.

They are not *bad* in my book, just too tightly specified.

So now you know.

Mounting them straight to heatsinks, with grease but not mica, adding (CPU chip sized) coolers, etc. , simply is not enough.

While same body LM3886 has a well deserved reputation for reliability, ... but nobody pulls more than 50W from them, a few try 56W ... way less stressed.

I can only shudder at Marshall MG100 ... with a lonely TDA725x carrying the load.