G-7700 near-basket-case where to start?
- Thread starter willray
- Start date
Right - I meant that it was possible that the same component in both channels had gone off-spec, making the fact that everything measured as fairly similar on both sides, not particularly helpful.
I was in a rush to finish up and didn't read the resistor codes to make sure that what I was measuring anywhere was actually the correct value.
With the original pots in place, I thought I was seeing 0mV on the Sansui-provided test points for bias. It might have shown 1mV, but I don't trust small readings, and I didn't want to touch the original pots.
Right now, I'm only working on the right side of the F2980 board -
After replacing the pots, I see about 1.2mV across the "top" (where Sansui's test points are) emitter resistor with 7 Ohms on the bias pot, and 1.6-1.8mV with 95 Ohms on the bias pot. Similar, though not identical values across the "bottom" emitter resistor. *SEE NOTE
Waveforms across the emitter resistors aren't quite identical, there's a bit more noise on the "bottom" resistor, and the amplitude on the "top" resistor is about 20% larger than on the bottom resistor, at least with small signals.
At 1.2mV bias on the emitter resistors, with a 1kHz tone injected, the final output at the speaker terminals has some high-frequency ringing at about (assuming a sine waveform) 45deg. This does decrease and starts to become invisible around 90 Ohms on the bias pots/1.7mV on the emitter resistors.
So yes, voltages are changing, but only a very, very little bit.
My current best guess is that either there's a leakage path across the bias pots, or, something in the TR08/TR16 and/or TR10/TR18 region that's preventing adequate voltage being applied to the output transistor bases.
*NOTE : Possibly relevant measurement - with the 100 Ohm bias pots in place, I only see 34 Ohms end-to-end across the pot, and a fraction of that from the wiper to either end. Pulled one back out to make sure it was really ok, and out of the F2980 board, it's a clean 100 Ohms across the pot, and the expected (sums to 100) values from the wiper to either end. "ohms on the pot" values measured after setting, then removing the pot to get a real reading.
I don't see an obvious parallel path across the pot - maybe I'm just blind, but should I be looking at C24 and C26 for leakage current?
super98lsc
Well-Known Member
x2 had same issue when I had a case of transistor dyslexia when rebuilding my BA-2k. Pin orders like to change on the new components vs the old. I always check the datasheets and write the Pin Orders on my packages. (I flipped the board over to solder it and inverted the pin order) brilliant...
I'll certainly check on this the next time the receiver and I meet, but I'm not holding my breath on this being the issue here. I'm 99% certain that this one was factory original and had never been opened before I dove in. If it's wrong, it came from Sansui that way (which, of course, isn't impossible).
For lack of a better culprit, and unless someone else verifies the measurement in a working 7700, my money is on the apparent low-impedance (roughly 50 Ohms) parallel path around the bias-adjustment pots. I need to get my wide-carriage printer running again, so that I can print the schematic out large enough to be useful…
Woohoo - firing on all cylinders!
Ok, who wants to make a guess at the dumbest reason possible for not being able to adjust the bias?
hint - certain assumptions shouldn't be made based on the seller's "it's in perfect working condition", because the seller is a lying sack of ****.
Regardless, I think we're back to fully-functional, at least on the Aux preamp and main amplifier circuits. Dialed in to a stable 0mV +/- 1mV offset on both channels, and right now I've got it idling at 3.3mV across each emitter resistor and it seems pretty solid there too.
As Kevzep said, it's definitely not the same on the F2980 board as the earlier G7700. When I have time, I'll map the values that exist on this version into the schematic. Some are quite un-subtle changes, like R617 going from the marked 1.8k, down to 2.2ohms, which completely changes all the voltages around the DC offset trimmers.
(update, just in case anyone follows this rabbit hole in the future - the preceding is incorrect, I was misremembering what component I'd been looking at - R617 doesn't even exist on this version of the board)
I suppose I'll eventually have to hook it up to a distortion analyzer and see if I can identify an actually appropriate offset for this version.
Now I need to pull all the sheet metal off, and flatten it back out again. Overall, there's been a lot less drama with this unit than I anticipated!
Ok, who wants to make a guess at the dumbest reason possible for not being able to adjust the bias?
hint - certain assumptions shouldn't be made based on the seller's "it's in perfect working condition", because the seller is a lying sack of ****.
Regardless, I think we're back to fully-functional, at least on the Aux preamp and main amplifier circuits. Dialed in to a stable 0mV +/- 1mV offset on both channels, and right now I've got it idling at 3.3mV across each emitter resistor and it seems pretty solid there too.
As Kevzep said, it's definitely not the same on the F2980 board as the earlier G7700. When I have time, I'll map the values that exist on this version into the schematic. Some are quite un-subtle changes, like R617 going from the marked 1.8k, down to 2.2ohms, which completely changes all the voltages around the DC offset trimmers.
(update, just in case anyone follows this rabbit hole in the future - the preceding is incorrect, I was misremembering what component I'd been looking at - R617 doesn't even exist on this version of the board)
I suppose I'll eventually have to hook it up to a distortion analyzer and see if I can identify an actually appropriate offset for this version.
Now I need to pull all the sheet metal off, and flatten it back out again. Overall, there's been a lot less drama with this unit than I anticipated!
Last edited:
ghazzer
Senior Member
That surprises me. Was the installed value of R616 also 2.2Ω? Are they both 2 watts?
My schematics show the value of R616&R617 to be 1.2KΩ, 1.5KΩ and 1.8KΩ in the G-5700, 6700 & 7700, each lowering the B+ of 54v, 61v & 71v to ~18v. I wouldn't think that a 2.2Ω resistor could handle that. A 53v drop across a 2.2Ω resistor calculates out to almost 24 AMPS!
Did you measure the voltage wrt ground between R617 and R16?
My schematics show the value of R616&R617 to be 1.2KΩ, 1.5KΩ and 1.8KΩ in the G-5700, 6700 & 7700, each lowering the B+ of 54v, 61v & 71v to ~18v. I wouldn't think that a 2.2Ω resistor could handle that. A 53v drop across a 2.2Ω resistor calculates out to almost 24 AMPS!
Did you measure the voltage wrt ground between R617 and R16?
Sorry - brain fart, I was misremembering what component I was looking at - however, it's even weirder than that. R616 and R617 aren't even populated on this version of the F2980 board. The DC offset region of the circuit is powered completely differently. C628/629 are vacant, as well as ZD614/605.
[update -- Doh! can't type either, ZD604/605 are vacant, not ZD614...]
Last edited:
ghazzer
Senior Member
Interesting! Perhaps we can create a schematic of your version of the F-2980 bd and post it as a .PDF file. Are there any markings on it other than "F-2980"?
I'll see if I can see anything in my reference photos - there are no obvious revision-numbers/etc near the main F-2980 marking, but I haven't closely compared to any other F-2980 photos from online.
I'll check more carefully once I'm back with the beast. At first glance it looks like the same PCB - it has markings for, and is drilled for the missing components.
My board picks up what would be the anode side of ZD605 and cathode of ZD604 with jumpers that run off to the right side of the board. If I'm seeing this right, the jumper from the anode point for ZD605 enters a thru-hole just to the right of R610, and the jumper from the cathode point of ZD604 enters a thru-hole just right of, and between points 24 and 25 on the edge of the F-2980 board.
If the stencil on the component side is accurate, it looks like those would be tied to the traces on which points 24, and 25 are located, respectively, at - and +33V (which checks out with the voltage that I found on the jumper between R15 and R16).
Will
Next up - 60-cycle hum - likely culprits?
Another weekend evening with the beastie, and a bit more new information:
The 7700 now seems to be running stably with both channels almost perfectly balanced, holding at 0mv offset, and 3.3mV bias (remember, this is the later version, and the service manual may not apply).
For reasons I don't understand, I don't see anything resembling crossover distortion reaching the speaker terminals, even with the bias turned way down. I don't recall how low I took it, but, an input sine wave looks like a clean sine wave on the output, essentially invariant of the bias. (Yes, I realize it's awfully difficult to see small distortions of a sine wave - just saying that no setting of the bias produced an obvious crossover dead-time gap in the signal).
To see if I could learn anything more about the correct bias, I took a look at the THD, and as it sits, I'm seeing THD+N way over spec, at about .3% at 10W output. However, this is with a huge amount of 60Hz hum leaking into the output, so this is the next quest. Turning the bias up to 8mV didn't change the THD+N at all, at any frequency.
I know - inject a signal and work forward from the inputs - and that's my next step, but, does anyone know of any canonical culprits in these to which I should pay extra attention? Of possible note, the hum does _not_ increase appreciably with increasing volume.
Another weekend evening with the beastie, and a bit more new information:
The 7700 now seems to be running stably with both channels almost perfectly balanced, holding at 0mv offset, and 3.3mV bias (remember, this is the later version, and the service manual may not apply).
For reasons I don't understand, I don't see anything resembling crossover distortion reaching the speaker terminals, even with the bias turned way down. I don't recall how low I took it, but, an input sine wave looks like a clean sine wave on the output, essentially invariant of the bias. (Yes, I realize it's awfully difficult to see small distortions of a sine wave - just saying that no setting of the bias produced an obvious crossover dead-time gap in the signal).
To see if I could learn anything more about the correct bias, I took a look at the THD, and as it sits, I'm seeing THD+N way over spec, at about .3% at 10W output. However, this is with a huge amount of 60Hz hum leaking into the output, so this is the next quest. Turning the bias up to 8mV didn't change the THD+N at all, at any frequency.
I know - inject a signal and work forward from the inputs - and that's my next step, but, does anyone know of any canonical culprits in these to which I should pay extra attention? Of possible note, the hum does _not_ increase appreciably with increasing volume.
So, just for posterioriarity's sake, should anyone else find themselves down this rabbit hole:
Should you find yourself with a G-7700 that "works" just fine, including the tuner display, reasonable output at the speaker terminals, etc, but, output bias that just won't budge above 1.5 to 2mV -- look up and to your left at that voltage-selector doohickey, and make sure that some dumb*ss hasn't configured the unit for 220V, with a 110V plug.
I probably never would have noticed this screwup, if I hadn't been forced to start backtracking supply voltages when I discovered that the bias was handled differently on this version of the system, than is documented in the schematic.
Oh baloney. You know infinitely more about these things than I do.
I know a fair bit about the doing, but precious little about what needs to be done or why. Makes me a pretty damned dangerous substitute for a competent technician!
Will