Back in 1980, Hitachi briefly experimented with a "Super linear ckt" design". I'm only aware of its use in two low-power integrated amp models, HA4700 (50 w/ch) and HA3700 (35 w/ch). (FYI: I am currently servicing an HA4700 which, with a prev. owner, suffered some damage in the R channel pwr amp stage. The L. ch is okay.)

The design goal seems to have been to "super linearize" the top (+) and bottom (-) half of the waveform to minimize switching (crossover) distortion. The serv manual goes into some of the theory.
It is a unique design that Hitachi didn't use long because of difficulty in controlling/stabilizing oscillation and biasing issues.

The full 27-page service manual is here:
https://archive.org/details/manual_HA4700_HITACHI/page/n1/mode/2up

This attached screen capture is from the HA4700 serv. manual schematic. It shows just the pwr amp section of one (L) channel.
Note Q715L (2SC1775A) after the two output pwr transistors (Q714/715) . Q715L is in the signal path (right-pointing arrow) and is in ckt just before the protection relay.
Anyone know what Q715L (2SC1775A) does?

 

Q715L's B-E samples the current thru R723L, and feedback (negatively) to the input of the driving stage (Q710L+Q714L). At least it offers certain level of overload protection. Q715L is symmetrical in functionality to Q712L/Q711L on the bottom half, which samples R724L.

 

Q715L's B-E samples the current thru R723L, and feedback (negatively) to the input of the driving stage (Q710L+Q714L). At least it offers certain level of overload protection. Q715L is symmetrical in functionality to Q712L/Q711L on the bottom half, which samples R724L.

Thx for the quick respose!
The touchy nature of this design is sometimes referred to as pwr amp ckt "bathing in negative feedback) to minimize that "nasty" crossover distortion.
Back in 2013, this UK vlogger created two extensive videos dedicated to servicing the HA4700:

Yes, this design can be very difficult to get behaving. No wonder Hitachi gave up on it quickly and in the follow-up HA4800 model, switched to STK power-amp modules.

 

As I noted in the OP, the R-ch had suffered damage in some prev. owner's mishap, and both pwr trans (Q713/Q714) were blown as well as some other transistors, including Q702/703 (2SB648, TO-126) , and burned out (charred) R's, way back early in the ckt (R703/704 both 2.2 k). I noted much of this a while back here on Reddit:
You can see many photos there or orig. damage and cond just before I began to repair.
As I noted in Reddit:

Not sure what happened to this amp? Evil spirits. Lightning strike. Or multiple attempts by former user (abuser!!) using shorted loudspeakers???

Both output transistors in Right channel were shorted, several related driver transistors also bad. Not sure what happened. It did power up and all the "lights came on" and main fuse okay.

The pwr-amp board has several metal-film fusible resistors (raised legs on PCB), and some are in signal path. Not sure why they are metal film (so they open faster for fuse protection???) . All are 1/4 W.

Many were open or very high in value (from 82 ohm, heat damaged to above 2k-ohm, etc). And several in Left channel, also, which still had output and driver transistors intact!

This amp has been on the back burner for a while. One issue was that replacement output transistors from China were bad (they actually cracked on power up). I did manage to finally acquire some good, new-old stock Sanken substitutes (see photo). Even sub-sub-substitutes are now "Obsolete" at DK/Mouser!
While the L-ch is very nicely tracking -- it was never damaged, just a few R's that were out of tolerance -- I can't power up the full amp w/o disconnecting the base fusible resistors (R738/739) for right-ch. Q713/Q714 because the both the output trans go into thermal runaway. With R738/738 still in, I can bring the variac up to to about 40vac and before Q713/Q714 start getting warm and sucking in the watts! (I also use a thermal camera for troubleshooting).
Not sure what the issue is other than some of the other blown/bad transistors I had to replace are closest substitutes I could find (because the originals are no longer avail).

Notably, these are:

Part on schematic | Orig. | What I used as replacement

Q702/Q703 | 2SB648, TO-126 | NTE2520
Q705, Q708 | 2SC1775 | 2n4401
Q710 | 2SD667 | NTE382
Q706, Q707 | 2SA872 | NTE91
Q709 | 2SB647A | NTE383

Note that for researching substitutions I use:

Google (AI results)
ChatGPT
NTE Source Research, SRI : sourceresearch.com
alltransistors.com

I'm still pretty new to troubleshooting, so I'm guessing -- other than broken (hidden) PCB traces, cold solder joints, etc -- that the above transistors are what is causing the issue. Those nine (9) transistors are the only asymmetry with the working L-ch section.
The schematic does a good labeling voltages, and I have gone thru some, but not most of these. Again, as noted earlier, I have to remove the base fusible resistors (R738/739) for right-ch. Q713/Q714 because the both the output trans go into thermal runaway. So voltage measurements might be tough because this design relies on so much feedback (as in the case of Q713/Q714 being on).

 

Consider checking the diodes used in the bias part of the output stage. At least in MARANTZ brand amps, those would tend to fail open, taking out the output transistors.
The reality is that some circuit designs are just plain junk, no matter how well they work when they have not yet failed. They are much too dependent on component properties that tend to drift.

 

I've put the HA-4700 aside for a while.
L channel , with many orig. components, is fine. The R-ch, with many new components, notably transistors, continues to act "possessed" . I do have the R-ch mostly connected, with only the two metal fuse R's (10 ohm, R738/R739L) disconnected, so I can test the whole amp at full power with it consuming "normal" idle AC mains pwr.

I have another equally mysterious vintage 1976 Int. amp (Kenwood KA-3500) with similar "possessed" issue in only one channel.
Will post on that in a dedicated thread.
Maybe we can put that project to bed before returning to the HA-4700!

 

I really applaud your dedication in fixing these vintage transistorized amplifiers.

Back when Japanese consumer electronics were at its absolute zenith. Different vendors each would come up with very unique circuit configurations in an effort to squeeze the very last performance drop. The downside which you have aptly described, is that they are a pain in the rear end to set up correctly, and in many instances require specialized semiconductors to operate correctly.

Sorry that I cannot provide any helpful information, but at least I can cheer up for you.

 

in the rear end to set up correctly, and in many instances require specialized semiconductors to operate correctly.

Sorry that I cannot provide any helpful information, but at least I can cheer up for you.

Well maybe you or someone can thru on the Kenwood I just posted on.
The KA-3500 sold very well and is still quite revered for its sonics. It uses a more standard topology, but lacks speaker protection...
https://forum.allaboutcircuits.com/...fset-and-bias-trim-stuck-around-0-4mv.210777/

 

I repaired a Kenwood amplifier for a co-worker, at one job, quite a few years back. It was not a kit, but a standard product. Since it was "bump sensitive" the actual fix was resoldering every soldered connection in the amp, which consumed a half hour lunch break one day. THAT repaired it very well. I did that because I spotted at least two poor solder connections. That did fix all of it's problems.

 

I only repair or mod items as a hobby, not a profession (service tech). So ... buy stuff on ebay or FBM ... and get it up and running. Amps are tough. A lotta art and science. Problem-solving them is addictive, almost narcotic.
CD players and phono gear is much easier and results are more reliable and predictable.
I have acquired quite a few tools of the trade (see my Profile page) and plan on getting some more, most notably a decent spectrum analyzer .
All that said, when it comes to electronics troubleshooting (and not being able to quickly resolve the issue), I can't use the excuse of "just a hobby" ... because my academic/educational/professional background is EE.