[yachadm]

I have just completed a recap, new potis, and complete RF and Bias calibration.
The machine is working just fine, except for one unusual issue.

On shutdown, the Right channel +/- B rails drain from +/- 37VDC (when powered on) to a few fractions of mV.
That's the normal behavior of a DC-Coupled output amp on shutdown.

However, the Left channel +/- B rails drop to +/- 10VDC on shutdown, and never go to 0V.

See the attached picture.

After shutdown, I measured all the points with about +/- 10VDC  - all marked  in red.
 
As a final check, with the machine still shut down, I connected a 100 Ohm resistor from Q406 Base  to GND -  the +B dropped to 0.5V,  and the -B dropped to  5.5V, and  stays there.

I replaced Q406 - no change.

All suggestions welcome.

Thanks
M Yachad

[FullRangeMan]

Welcome Yachad

[Phil A]

Welcome!

[Davey]

I'm trying to think back forty years, but I think you should probably investigate the power supply as well.  This is a dual-mono design so each channel has its own power transformer, bridge, reservoir capacitors, etc.
But, I think you're probably on the right track investigating the amplifier board.  C406 and C412 would probably be good candidates to check since the electrolytics would be the most prone to long-term failure.

FYI, I still have a Harmon Kardon 730 in use.  Other than controls getting scratchy, it's never missed a beat. 

Dave.

[ArthurDent]

Greetings & Welcome to AC M Yachad   

[jea48]

I have just completed a recap, new potis, and complete RF and Bias calibration.
The machine is working just fine, except for one unusual issue.

On shutdown, the Right channel +/- B rails drain from +/- 37VDC (when powered on) to a few fractions of mV.
That's the normal behavior of a DC-Coupled output amp on shutdown.

However, the Left channel +/- B rails drop to +/- 10VDC on shutdown, and never go to 0V.

See the attached picture.

After shutdown, I measured all the points with about +/- 10VDC  - all marked  in red.
 
As a final check, with the machine still shut down, I connected a 100 Ohm resistor from Q406 Base  to GND -  the +B dropped to 0.5V,  and the -B dropped to  5.5V, and  stays there.

I replaced Q406 - no change.

All suggestions welcome.

Thanks
M Yachad

Someone suggested below the Amp has a dual power supplies. Separate power transformers and all. Can you furnish a schematic diagram of both power supplies.

There may be resistors that connects the B- rail of each amp power supply to common chassis ground. The resistor for the left channel may be open.   

[yachadm]

Thanks for the kind words!

The PSU schematic is attached.

I don't see any bleed resistors, neither on the actual machine!

As I mentioned, I replaced all the caps already -Panasonic FC, and in the signal paths -Nichicon KA. So that base is covered already.

[yachadm]

Here's the Tone Pre-Amp - any resistor over 100K is probably a likely candidate?

However, I don't see any resistor connected directly between GND and either the + or - B rails.

[yachadm]

OK, I put up a nice clear Hi-Res schematic, but why does it show up as a low-res blurred picture?

[jea48]

Thanks for the kind words!

The PSU schematic is attached.

I don't see any bleed resistors, neither on the actual machine!

As I mentioned, I replaced all the caps already -Panasonic FC, and in the signal paths -Nichicon KA. So that base is covered already.

I can't see anything in the power supply design that would bleed off, drain, the DC charge in the electrolytic storage caps in the two power supplies. There in a bleeder resistor or designed load, or connected load.

Without such a design, imo, the natural storage ability of the electrolytic caps to hold the charge in the left channel power supply would be normal. As to what is causing the right channel channel power supply's DC power to bleed off is unkown, but jmo,  it is not due to the circuit design. 

[yachadm]

In 30 years, I haven't yet seen a bleeder in a power supply.

They are more normally in the Pre- and/or Power-amp stages, with a transistor switching to conduct the power to ground, as soon as power is cycled off. The B+ and B- rails go right through the amp system from the Power-Amp, all the way back to the start of the Pre-amp.

Therefore, I highly doubt that the right channel (which bleeds down in normal fashion) is defective (I may be wrong, if HK went against the grain, with this model)

This HK is a first for me, so I'm just wondering if someone has seen this before, or a similar problem on another amp.

[Davey]

I can't see anything in the power supply design that would bleed off, drain, the DC charge in the electrolytic storage caps in the two power supplies. There in a bleeder resistor or designed load, or connected load.

Without such a design, imo, the natural storage ability of the electrolytic caps to hold the charge in the left channel power supply would be normal. As to what is causing the right channel channel power supply's DC power to bleed off is unkown, but jmo,  it is not due to the circuit design.

The amplifier circuit itself is the connected load and bleeds the capacitor charge at power off.  There is no need for a bleeder resistor in this sort of scheme.
Note also, the two power supplies are identical, so you would expect equal discharge in both sections....assuming the L/R amplifier circuitry is the same.  Obviously, in this case, it's not and that's what yachadm is investigating.

It still seems likely the issue is a problem in the left channel amplifier module, but I wouldn't eliminate a PS issue just yet.  The large electrolytics can act funny with age.....but I guess you've replaced those as well.  I haven't looked inside my unit in decades, so I don't remember if any easy left/right switch is possible.

Dave.

[yachadm]

Dave, thank you. You've understood the issue.

So, we are looking in the amplifier circuit, and not in the PSU.

The Power Amp is on a detachable PCB with spike connectors, and it is virtually impossible to swap the Left channel circuits onto the right channel rails.

And yes, there are 4 brand-new Nichicon GU reservoir caps in there.

[Davey]

This is probably some sort of biasing issue and you have a transistor (or two, or three) getting into cutoff prematurely and "disconnecting" the amp module from the capacitors and thus stopping the discharge.  It might be a bit tedious, but you should be able to zero-in the culprit with comparison testing to the other module.

This is an excellent, nostalgic, piece of equipment to restore/use.  It's worth the effort.

Dave.

[yachadm]

In the Left Output amp, I have already replaced ALL the semiconductors, except for Q402 and 404 2SC1400 - simply because I don't have stock (or any suitable subs).

I did test them, and they both show hFe of 700, no leakage, which is expected, so, they may be testing OK static, but failing under load.

So, I'm standing by on those to exchange when I get new 2SC2632 parts.

[yachadm]

OK, I think I have it.

The surges started happening again - still working via the DBT, so no damage. I saw the Bias voltage/current go up to 126mV!!! Then it dropped down to 35mV stable again.

After power off, I started touching the output amp components, to feel if any were hot, and to see if any affected the residual voltage.

I touched one of the 2 0R47 Bias resistors, and it made a hum on the speaker, and I saw the B voltage drop to zero.

I powered up again, and shutdown - zero B voltage .

So it looks like an intermittent bad resistor. That would also explain the surges also, and the unstable high Bias voltage/current.

It's too late tonight - I'll exchange the 4 Bias resistors tomorrow - I have Solen Audio Bias resistors.

[yachadm]

Problem of the +/- B rails not draining, is solved.

This was a very interesting investigation and understanding - it will not appear on any other amplifiers, because HK used 2 separate transformers for the Left and Right channels.

In order to investigate, I completely disconnected the L & R Output Amplifiers from the machine.

I then checked the L & R  +/- B rails during power on (+/- 37VDC), and shutdown (L +/- 10VDC, R zero).

So, I now understood that the Output Amplifiers have NO influence on the draining issue.

I started to trace the B rail wiring from the PreAmplifiers, to their source at the Bridges.

It turns out that the RIGHT transformer supplies B power to BOTH the Left AND Right PreAmplifiers, but ONLY the Right Output Amplifier. Therefore, the LEFT Transformer supplies ONLY the Left Output Amplifier.
This is NOT what is shown in the schematics!

Therefore, because the PreAmplifiers (and NOT the Output Amplifiers) cause the power drain on shutdown, and because the PreAmplifiers are connected to only the Right B rails, then ONLY the Right Channel B rails are drained.

To confirm this, I transferred the B wiring feeding the L & R PreAmplifiers FROM the Right Transformer to the Left Transformer.
Now, the Left B rails drain on shutdown, while the Right B rails remain charged.

This appears to be a cost-cutting step, instead of having separate B wiring for the L and R PreAmplifiers, as shown in the schematics.

So, after a normal shutdown, the Left Channel will remain charged, but the Right channel will drain to Zero.

Any comments?