Crossover distortion in Yamaha G100-212

[Nyqusit5]

I have noticed that clipping off the 56Ohm resistor in parallel with the Idling current trimpot allows to greatly reduce the crossover distortion (but not remove it) if the trimpot is turned to maximum value.
This obviously allows for a greater idling current setting (as in this amp version 2) but I still have OmV across the .22Ohm ballast resistors (no load, no signal).
At max settings, I have 46.3V at T5 51's anodes and 45.8V at its cathode (with 46.7V at point 5 and 95V at point 6 or V+)

[J M Fahey]

I have noticed that clipping off the 56Ohm resistor in parallel with the Idling current trimpot allows to greatly reduce the crossover distortion (but not remove it) if the trimpot is turned to maximum value.

You are saying that even with *greatly* increased resistance and going wildly away(3X higher ) from circuit requirements you still can not properly bias the power transistors

This obviously allows for a greater idling current setting (as in this amp version 2) but I still have OmV across the .22Ohm ballast resistors (no load, no signal).

That

At max settings, I have 46.3V at T5 51's anodes and 45.8V at its cathode (with 46.7V at point 5 and 95V at point 6 or V+)

Ok, you are showing that you have about 0.5V = "1 diode drop" across T551 (or TS 51) which is clearly not enough.

You have to overcome "3 diodes" to bias Vbe of: 2SC783 , 2SC1586Y (top) and 2SA483 (bottom) which would require 3 diodes in series, or, to make it adjustable, 2 diodes plus a series resistor.

I am quite certain that TS 51 is NOT a regular (say 1N4002) diode but a double one (2 diodes in series) and sometime along the way some un-enlightened Tech (or a shotgunner) replaced it with a plain vanilla one.
So you will NEVER be able to bias amp.

Please post a closeup picture of TS 51, mainly to show how is it mounted, it must track heat sink temperature so it may either be clamped to heat sink, epoxied to it or going through a hole drilled in it (which is often filled with thermal grease).

In principle and just for testing replace current "TS 51" with two 1N400x in series, of course replace the 56 ohm resistor where it belongs and try to rebias.

You **might** need 3 x 1N400x in series for biasing so if 2 are not enough, experiment with 3, but I find that unlikely.
In any case, measurements rule.

If confirmed, search for double or triple diodes in a single package, as shown by measurements, to be mounted there; I suspect you have space for "1 diode only" although I might be mistaken.

Of course, you have that amp on your bench and we do not, so those closeup pictures will help a lot.

FWIW Peavey (and others) sometimes use such a double diode for same duty, maybe Enzo or another Peavey expert can suggest one.

My feeble memory seems to remember MZ2361 might be one of them:

http://pdf.datasheetcatalog.com/data...360-MZ2361.pdf

http://pdf.datasheetcatalog.com/data...emi/SA7-29.pdf

PS: next time measure voltage drop directly across TS 51 instead of 2 separete measurements relative to ground, you´ll have more precision.
So please do and post results; both current one and double (triple?) diode one, as well as final base to base voltage 2SC783 to 2SA483, as soon as you can notch distortion out and get steady voltage across emitter resistors.

[glebert]

Please post a closeup picture of TS 51, mainly to show how is it mounted, it must track heat sink temperature so it may either be clamped to heat sink, epoxied to it or going through a hole drilled in it (which is often filled with thermal grease).

Maybe looks something like this (or should), which is on my Fifty 112.

[J M Fahey]

Thanks.
Well, that one does not look like a multipurpose 1N400x diode or similar but a component specially made to be bolted to a heat sink for thermal purposes.

Maybe you can measure diode/voltage drop across yours?

I would expect close to 1.2 - 1.4 V.

[glebert]

Thanks.
Well, that one does not look like a multipurpose 1N400x diode or similar but a component specially made to be bolted to a heat sink for thermal purposes.

Maybe you can measure diode/voltage drop across yours?

I would expect close to 1.2 - 1.4 V.

In circuit I measured 0.8V for one polarity and 1.7V in the other, then again my amp has questionable driver transistors (power transistors are removed). Looking at the schematic that seems closest to my amp they call out a STV-3H for that part.

[Enzo]

So look specifically for an STV-3H then. I had a few of those, and I sold the last half dozen for $50 a couple years ago. They were used a lot in older stereo amplifiers and receivers. And they are indeed rare these days. Three diode drops or about 1.8v, you saw 1.7, sounds good that way.


Google just that part number for a variety of discussions, including this one:
https://music-electronics-forum.com/...ad.php?t=24991

[J M Fahey]

Yes, looks like 3 diodes in series.

Funky radiomuseum says:

Material/Aufbau: Silizium Stabilisator-Diode / Forward Reference Diode;
Form/Case/Outline: TO-92 Anschlussfolge _-_;
Daten/electr.data: V F: 1.6 ... 1.8 V; I sp: <10 ľA; U sp: 50 V; N: __ mW; Imax: 7 mA; t stg: -60...+100 °C;
Manufachtured by Sanken (J)

Personally I would solder together 3 side by side 1N4002 (easier to handle than smaller glass cased 1N914/1N4148) , in series, up-down-up so only 2 end leads leave the "pack" and epoxy them to a heatsink, scratching and degreasing surface first for good adhesion.

Position on heatsink is not that critical, can vary 1 inch any way as long as it´s on the thicker part, not on a thin fin.

Way faster than searching for a hard to find part.

Just 1 hour ago I epoxied a thermal sensing transistor to a 300W Bass amp backplate/heatsink I have to deliver tomorrow.
I added a small improvised PCB to act as a sort of terminal strip so wires to main PCB can´t tear transistor legs off if pulled.

As always, I am a minimalist searching for the fastest cheapest solution.

Everything epoxied to a sanded degreased sheet of aluminum (homemade back panel):



You can make a similar "terminal strip" out of junkbox PCB clippings to hold 2 or 3 diodes , epoxying everything to heatsink.

Notice 2k2 gate stoppers soldered straight to Gates and attached to IRFP250 body with a dab of Epoxy.

If gate stoppers are used so close, then wires to PCB can be relatively long (say 4") without danger of self oscillation, allowing easy pcb removal for servicing before unsoldering anything.

[Jazz P Bass]

"I added a small improvised PCB to act as a sort of terminal strip so wires to main PCB can´t tear transistor legs off if pulled."

Sweet!

[Nyqusit5]

I tried replacing TS 51 with two 1N4007 and the amp still had the crossover notch. I then tried with three and the crossover distortion disappeared... but the voltage accross the ballast .22Ohm resistors keeps on creeping up: from 35mV to 60mv and then on (no load, no signal) even with the idling current trimpot set to minimum. I stop the amp when it reaches this value. Maybe two 1n4007 and a low voltage drop diode would do it. I will try.

[J M Fahey]

Some use 3 diodes in series but with a lowish value resistor in parallel with *one* of them.

Evans LV500



Peavey 260



Try it, a popular value is 100 ohm, but of course you adjust it with a potentiometer and once found put a fixed resistor instead.
If not exact resistor available, choose slightly lower, a tiny little barely visible crossover notch is better than a thermally unstable amp.

Or slightly mod circuit to an "amplified diode" or "Vbe multiplier" circuit which is more adjustable.

Diodes come in discrete values, often (as you found) "2 diodes are too little, 3 diodes too much" , and a series resistor as you have now is somewhat crude, since resistor portion of total bias voltage does not track temperature at all.

EDIT:

I then tried with three and the crossover distortion disappeared... but the voltage accross the ballast .22Ohm resistors keeps on creeping up: from 35mV to 60mv and then on (no load, no signal) even with the idling current trimpot set to minimum. I stop the amp when it reaches this value.

I HOPE your diodes were touching heatsink; if separated they will track sh*t .

Also remember to wait a couple minutes for diodes to reach ambient or heat sink temperature; if just soldered they are HOT and adjusting bias in that state is useless.

[vintagekiki]

Yamaha G100 schematics

Thank you Glebert but this service manual is for version II. The PA circuit is completely different.
The schematic I gave the link to at Elektrotanya is the right one.

Accurately
The Yamaha G100 amplifiers are manufactured in three versions.

G100 version I (transistor preamp; two channel)
https://elektrotanya.com/yamaha_g100_preamp_schematic_sch.pdf/download.html
https://www.owner-manuals.com/G100-service-manual-YAMAHA.html (with amp adjustmens)

G100 version II (IC preamp; parametric EQ)
https://www.owner-manuals.com/G100115II-service-manual-YAMAHA.html

G100 version III
https://www.owner-manuals.com/G100112III-service-manual-YAMAHA.html

Yamaha Combos, Heads and Cabinets
https://yamahamusicians.com/forum/viewtopic.php?f=134&t=7006


https://www.owner-manuals.com/-manuals-1-20a.html?alpha_filter_id=71

[Nyqusit5]

Using one 1N4007 and two MUR120 in series seems to do the trick. I can set the idling current to 2.4mV accross the ballast with the trimpot at minimum and the value only increases to 4.2mV after 15mn (with load but no signal).

I tried one double diode salvaged from a Peavey studio Pro but it wasn't enough.

[J M Fahey]

Using one 1N4007 and two MUR120 in series seems to do the trick. I can set the idling current to 2.4mV accross the ballast with the trimpot at minimum and the value only increases to 4.2mV after 15mn (with load but no signal).

I tried one double diode salvaged from a Peavey studio Pro but it wasn't enough.

Cool.
Are diodes epoxied/attached to heatsinks?

Only way to guarantee tracking.

I see MUR120 boast of "low forward voltage" and that seems to do the trick.

In the old days some amps used "special sauce" diodes, famous RCA 1N3754 comes to mind.

Absolutely unavailable here in Argentina, everybody replaced them with 1N4002 / 1N914 / 1N4148 and their amps overbiased and burnt, not kidding.

In the early 70´s, when "RCA amplifiers" became very popular, specially the 70W one with (also unavailable here) 40411 power transistors, because they were revolutionary at the time, "everybody" built a couple ... and was deeply dissappointed because they thermally overrun, oscillated like crazy, or both.

I became sort of a "whiz-kid" because (after much frustration and experimenting) was able to tame both, so got a ton of them through my "shop" ... which way back then was a plain table at my Grandma´s appartment "servants room".
Oh the good old days

[vintagekiki]

Oh the good old days

Good old days