Yamaha G100 - crossover distortion

[bigchief]

Hi, I'm working on a Yamaha G100 410. The owner had put 5amp fuses in it (it calls for 3amp) and I'm told by the owner that the amp's output was weak and intermittent until smoke started pouring out of it. Inside I found both of the 0.47ohm emitter resistors hanging free from the board and one of the 150ohm emitter resistors connected to the base of the 2SC1586 (identified as TR23 in the schem) went nuclear and burnt a hole through the board, destroying some traces. Elsewhere a 10ohm resistor was also smoked and the Toshiba 2SC483 tested bad in my transistor tester, and had continuity between each leg. I searched for an alternative as I couldn't find a good source and it was suggested in a forums post I found that an STM 2N3440 would be fine, so I went with that.

I replaced these parts, jumped traces with 22awg wire, cleaned all carbon deposits on the bottom side of the chassis and the surrounding areas on the MA board, removed and tested all transistors, and replaced the power transistors with some NTE MJ15015's and put on fresh thermal paste -- reasoning being that although the original Toshiba 2SC1586's tested good, I wasn't sure how much heat and abuse they had taken prior.

Unfortunately, due to the stacked construction of the MA module I failed to notice the NTE MJ15015's I had were thinner top-to-bottom than the original power transistors and the TO-3 socket mounting screws dug into surrounding traces of the board. Connected the amp to a light bulb limiter, flipped it on and was greeted with a loud, angry hum. I immediately flipped the amp off and dug back into the MA module to find the MJ15015's were shorted, along with the original Toshiba 2SA483 -- 2SC783's checked ok. I put the Toshiba 2SC1586's back in, ordered an NOS Toshiba 2SA483 that looked identical to the original, and tried again. Amp fired up, no hum, no blown fuses and everything seemed ok but I have crossover distortion and the amp seems weak for 100 watts.

I fiddled with VR3 but the crossover distortion got worse anywhere but where it was set originally and I adjusted V2 so that I got 45v at the +CO terminal of the DC board. What concerns me is that voltage at the test point of the DC power board is supposed to be set to 10mv but my meter registered 00.5mv with VR1 at its maximum. The output transistors are running quite cold at idle -- 70 degrees fahrenheit. I'm relatively inexperienced with SS gear but I'd expect more?

Suspecting there might be an issue with the electrolytics in the power supply, given the amp's age, I broke out my Anatek Blue ESR meter and checked the four electrolytics there (2x 1000/25v, 2x470/50v), as well as the 1000/100v and 1000/160v cans mounted to the chassis; all read around .08 which checks out as good. I also checked the MS1 rectifier packages, and S5151's & S5151R's diodes in-circuit and they appear fine.

Schematic I'm using is here: http://schems.com/bmampscom/yamaha/Y...ice-manual.pdf

[Jon Snell]

The first thing I would do is check the voltage on the collector of TR23. It should be a 50% of the total rail voltage, about 70volts.
Then check the voltage across D1 for about 0.6volts. Switch the amplifier off and ensure TR21s emitter to TR23s collector reads 10R when power is not applied to the amplified, (as a cold check).
That should give you enough information to work out what the issue is.

[bigchief]

The first thing I would do is check the voltage on the collector of TR23. It should be a 50% of the total rail voltage, about 70volts.
Then check the voltage across D1 for about 0.6volts. Switch the amplifier off and ensure TR21s emitter to TR23s collector reads 10R when power is not applied to the amplified, (as a cold check).
That should give you enough information to work out what the issue is.

Howdy Jon,

Collector of TR23 to ground is 46v, same as my reading (+/- 1volt, per the schem) at the +CO terminal on the power supply board. Across D1 is 1.6v. With the amp off collector of TR23 to the emitter of TR21 is 2 megaohms.

On a hunch, I previously checked D1 out of circuit and it behaved as I would expect a diode to when checked with my DMM in diode mode. However, while googling prior to making my original post, I found this thread here, where the well-known Enzo specified a shorted, not open, STV3H would produce more crossover distortion than usual. I suppose that might be it then?

[Jon Snell]

If you think about it, there is a 10R resistor between Collector TR23 and Emitter TR21, that reads 2M ! There is your issue.
Replace the 10R , turn the bias pot to minimum resistance and power it up. If all goes well, set the bias current.
Hope that helped.

[bigchief]

Yep, that was one of the resistors I replaced initially as the original was burnt to cinders. I just checked the resistor in circuit and it measured 10ohms across. I removed the board from the heatsink, de-soldered one side of the 10ohm resistor and checked again; sure enough, 10ohms. I must be missing something?

[Jon Snell]

What is the voltage between the bases of TR20 and TR21? Should be enough to turn 3 junctions partially on.

Are the 10k and 5k6 resistors OK?

[bigchief]

Hey Jon, my apologies. Yes from the Collector of TR23 and Emitter of TR21 I get 10ohms. My brain registered "TR21" as TR22.

1 watt 10K and 5.6k resistors read to spec with one leg lifted from the board. Voltage between the bases of TR20 and TR21 with the board removed from the heatsink is .121v

[Jon Snell]

We are getting closer.
If you consider that we get a voltage drop because there is an amount of current flowing. (Kerckhoffs's Law).
We need HT on the collector of TR19. TR19 in turn will hold its emitter at 0.6v less than its base and then the emitter of TR20 then TR22, which according to the service manual has about 80volts on it and on the emitter of TR22 then pin 5 on the board.
Bias is created with the current drain set up by the two resistors, 10k and 5k6.
Is the bottom of the 5k6 at ground level or 0v Pin1
the top of the 5k6 will equal about 1/3rd of the voltage on the output Pin4.
The top of 10k feeding the diode shoul have about 1/2 HT and equal to pin5 at half HT voltage.
0v on Pin4
TR17/18 emitters should be about 120v

If the bootstrap capacitor 10u 80v is short circuit, there will be no bias voltage as the current will be directed away from the diode and pot.

[bigchief]

We are getting closer.
If you consider that we get a voltage drop because there is an amount of current flowing. (Kerckhoffs's Law).
We need HT on the collector of TR19. TR19 in turn will hold its emitter at 0.6v less than its base and then the emitter of TR20 then TR22, which according to the service manual has about 80volts on it and on the emitter of TR22 then pin 5 on the board.
Bias is created with the current drain set up by the two resistors, 10k and 5k6.
Is the bottom of the 5k6 at ground level or 0v Pin1
the top of the 5k6 will equal about 1/3rd of the voltage on the output Pin4.
The top of 10k feeding the diode shoul have about 1/2 HT and equal to pin5 at half HT voltage.
0v on Pin4
TR17/18 emitters should be about 120v

If the bootstrap capacitor 10u 80v is short circuit, there will be no bias voltage as the current will be directed away from the diode and pot.

Hey Jon, here's my measurements. Hopefully I got all of those that you advised:

TR19 Collector - 127v
TR19 Emitter - 38-51v (Voltages would cycle up and down)
TR19 Base - 46v
TR20 Emitter - 50v
TR22 Emitter - 46v
Pin 5 (MA board) - 51v
Top of 10k resistor - 50v
Bottom of 5k6 resistor - 0v
Top of 5k6 resistor - 18.5v
Pin 4 (MA board) - 0v (I checked continuity and pin 4 does link to chassis ground)
TR17 Emitter - 119v
TR18 Emitter - 119v

I removed the 10uf 80v capacitor and it tested 11uf, so not shorted.

I might add that the traces destroyed in the initial damage destroyed the point where one end of the 2w 0.47ohm resistor connectors to the emitter of TR23. I tack-soldered a wire directly from the emitter socket to Pin 4 of the board and soldered the other end of the 0.47ohm resistor directly to the socket pin of TR23's emitter. It also destroyed the trace connecting the positive ends of the 33uf 160v capacitors -- here I simply soldered a resistor leg to either solder joint. The replacement 1/2 w 150ohm sits over the hole.

[Jon Snell]

As TR19 base is 46voits emitter can be no lower than 45.3volts which will appear on TR20s base, so why is there 4volts across those two junctions?
There is also an issue with 5volts across the base emitter junction of TR22; base at 50v (same as TR20 emitter) and only 46volts on TR22s emitter.
The maximum junction voltage is 0.7volts for a silicon transistor, the same as a diode.

At this point I would replace TR17 - 23 inclusive plus check the function of the regulator TR24 and TR16 components.
Recheck peripheral components, when you are satisfied all is good, power it up with a current limiter to avoid more damage. 50mA is the maximum current to be safe.

[glebert]

Admit that I did not read everything in this thread, so feel free to ignore if this isn't relevant. Do we know that D1 is a regular diode? On other G50/100 amps in that general position they use a special part labeled as TS (for thermal sensor probably) but with diode symbol that has a drop of about 1.7V. Here is another thread where we discussed this. https://music-electronics-forum.com/...813#post538813

[Jon Snell]

D1 is a multi unit diode and it does have about 1.4 volts drop.

[bigchief]

As TR19 base is 46voits emitter can be no lower than 45.3volts which will appear on TR20s base, so why is there 4volts across those two junctions?
There is also an issue with 5volts across the base emitter junction of TR22; base at 50v (same as TR20 emitter) and only 46volts on TR22s emitter.
The maximum junction voltage is 0.7volts for a silicon transistor, the same as a diode.

At this point I would replace TR17 - 23 inclusive plus check the function of the regulator TR24 and TR16 components.
Recheck peripheral components, when you are satisfied all is good, power it up with a current limiter to avoid more damage. 50mA is the maximum current to be safe.

Potentially a dumb question, but would you concur that the 2n3440 I have in the place of the original 2sc484 (TR24) is an appropriate substitute?

I will remove and test all the transistors out of circuit and compile results, and start assembling parts.

[Jon Snell]

The 2N3440 would appear to be a good choice.
Still, it doesn't explain the excess junction voltage present across TR20/21 & TR22 (base emitter voltages must not exceed 0.7volts each which is 2.1volts in total).

[bigchief]

The 2N3440 would appear to be a good choice.
Still, it doesn't explain the excess junction voltage present across TR20/21 & TR22 (base emitter voltages must not exceed 0.7volts each which is 2.1volts in total).

Thanks. Just wanted to verify what I read elsewhere and cross tracking down a replacement 2SC484.

A discovery I made just now: I tack-soldered some scrap braided wire onto the emitter solders joints of TR20 and TR21 so that I could take readings while the MA module was put together and inserted into the amp. While removing these wires, I found the pad for the emitter of TR20 lifted right up off the board with the wire. Not sure if this existed beforehand or is just the result of fatigue from having removed the transistors to test out of circuit a few times now -- soon to be a fourth. It is an old PCB, afterall. Regardless, it has been fixed now.