Power supply problem on Marshall amp

[Electron Tornado]

I'm repairing a Marshall MG100DFX solid state amp. The TDA7293 chip had blown and damaged a couple of the caps on the daughter board with it. The amp has a headphone output, so I removed the board with the TDA 7293 and tested the amp. Everything works on headphones, so everything works up to the master volume control. After the master volume, I have good continuity between all components all the way to the speaker output.

Here is a link to the Marshall schematic:  http://www.amprepairparts.com/mg100dfx.pdf

Here is a link to the TDA7293 datasheet: http://www.st.com/web/en/resource/technical/document/datasheet/CD00001887.pdf

I replaced the TDA7293 and all the caps on the board with it. I have correct voltages to all the pins on the amp chip  - except for pin 9 - and I get no output to the speaker.

The TDA7293 pin 9 gets a voltage from the output of TR2 (C3198) in the power supply section. TR2 is in the power supply section at the bottom of the page with the power supply. The supply voltage in that section is +12 volts. However, the voltage I read on the base is -0.5V. On the collector it's -0.2V. Can't be right. I think the voltage on the collector should be at least 3.5V.

The diodes D14 and D17 were tested OK in the circuit. Is this just a failure of TR2, or is there something else I should check?

Here is a link to a datasheet for TR2, (C3198):  http://www.ic72.com/pdf_file/c/c3198.pdf

[PRR]

That schematic does not make sense, probably has errors.

And is probably over-fancified.

Does the fan run?? Does the digital effects thingie do its thing?

If so, hard-wire the chip Standby always-on.



This may thump at turn-on.

[Electron Tornado]

The fan runs, digital effects all work, all other channels and features work.

Paul. that might be a workable "quick and dirty" solution, but I'm hoping to learn more about what I'm seeing around TR2, and whether it's the ttransistor itself that failed or something else. TR2 is ni the power supply section on the first page of the pdf file, in the section on the bottom of the page.

Avoiding thumps when powering on would also be a good thing.

[PRR]

TR2 can be "any" Si NPN jellybean.

I'd like to understand it too. My first reading was wrong. TR2 apparently pulls STBY to ground when power comes through D10-D13... but the STBY pin must be high (>+3.5V) to play.

No, that's not right neither. Maybe if I stare at it a while it will make sense.

Replace C76 just for luck. Observe polarity. Value not hyper critical.

[Electron Tornado]

TR2 can be "any" Si NPN jellybean.

I'd like to understand it too. My first reading was wrong. TR2 apparently pulls STBY to ground when power comes through D10-D13... but the STBY pin must be high (>+3.5V) to play.

No, that's not right neither. Maybe if I stare at it a while it will make sense.

Replace C76 just for luck. Observe polarity. Value not hyper critical.

If I'm reading the TDA7293 datasheet correctly, the standby pin does need to be above +3.5V. Right now it's not, and that's why there's no output from the chip.

I'll try putting in a different transistor and change the cap. I wonder - if TR2 can be any NPN, why did they use a C3198? Trying to source a replacement, it doesn't seem to be very common.

[PRR]

> any NPN, why did they use a C3198

The part-number is surely 2SC3198.... most Japanese transistors start "2SC", often abbreviated on the package and apparently even in documentation.

Shorted Zener is another possibility.

Cut out the transistor. It may un-standby. If it isn't rude, Marshall over-designed it. If thumping is an issue, put in whatever NPN is handy (watch the pinout).

[teemuk]

Marshall over-designed it.

Sorry, but that just sounded so comical. Marshall amps using this chip have a reputation of great unreliability and since hundreds of other mnaufacturers seem to get them working just perfectly it likely isn't a case of "over-designing". For Marshall's defense, I have understood that the chip is very, very sensitive and easy to destroy with many not-too-obvious quirks (e.g. asymmetric power up of supply rails). The high failure rates may well be due to using a standby/muting scheme not recommended by the chip datasheet.

Anyway, the chip failure is very common problem of these low to medium power Marshall solid-state amps. Usually its fixed by simply replacing the chip but in rarer cases the failure also takes down some of the mute circuitry. It's not many components so why not just test all of them thoroughly and replace what needed.

[teemuk]

I'd like to understand it too. My first reading was wrong. TR2 apparently pulls STBY to ground when power comes through D10-D13... but the STBY pin must be high (>+3.5V) to play.

The transistor in the circuit does not "drive" the mute circuit in any conventional way. Delayed startup in mute mode is actually introduced only by the simple RC filter at the transistor's collector and the transistor circuit itself has no effect on delayed startup.

The transistor's function is quick drainup of capacitor C98: Whenever supply voltage drops enough (e.g. amp gets powered off) the voltage decrease is highly asymmetric per rail (2200uF draining vs. 4.7uF draining). This asymmetry turns transistor's base voltage potential positive. The transistor shunts, the capacitors sees a very low impedance to ground and bleeds off its charge very fast. The collector voltage quickly drops below 3.5V (or whatever that threshold was) and the amp goes on mute mode, making no inconvenient noises when powered down. Without the quick transistor shunt, draining the charge would likely take even longer than what it takes for the amp to get off from mute mode when powered on.

If the transistor is shorted, it will constantly pull the mute pin close to zero volts. If the voltage limiting zener is shorted it will have the same effect.

[Electron Tornado]

The 2n3904 looked like it had similar enough specs, so I installed that in place of the bad C3198. Powered on and got signal to the speaker. The voltage going to the standby pin is +9V. Noodled around a bit with a guitar and everything worked OK.


"....and there was much rejoicing."



Thank you for your help, Paul, and thanks to teemuk for explaining how that part of the circuit works.

[PRR]

> thanks to teemuk

+1.