Zenith 7695 amp circuit questions

[Daikini]

As a means to improve my understanding of tube amps, and to kill some time while I wait for my Stereomour kit to ship, I'm rebuilding this amp (see schematic attached) I pulled from a console record player.  I've done all the usual things to improve performance (remove tone controls, upgrade all caps, increase cap values for coupling and power supply caps, individual cathode resistors and bypass caps for each channel, etc etc), and it sounds nice, but there are a few things that I still don't understand:

1.)  Why does it seem to need some treble roll-off between the driver and power stages? Without R72/C55 and R80/C61 to roll off the treble, it's just piercing.

2.)  Are C57 and C62 (the caps across the output transformer primaries) there to provide negative feedback?  When I remove them entirely, there is audible distortion in the high frequencies.  As I increase the value, I get relatively more bass response and less treble (does not sound good), and when I decrease the value, the amp sounds more open and lively than with the stock value, but distortion creeps into the treble.

3.) Can anyone explain the output transformer primary tap/power supply arrangement to me?  I have gathered from internet research that this is some kind of hum-cancelling arrangement, but I don't understand it. 

4.)  The 7695's that were in it to start with are fairly quiet, but when I put any other 7695 it's noticeably noisier.  Is this just tube variability among manufacturers/tubes, or is there something about the circuit I could adjust to improve it?

Thanks for any insight you can give.

[Paul Birkeland]

Did you change C54/C59 from 0.01uf to 0.1uf?  (That would help)

The caps across the primary of the output transformer will essentially short some high frequency signal out (they provide a low impedance path for this high frequencies to AC ground).  These may be a function of making the amp perform well with those particular output transformers.

For the general performance, I would perhaps look into adding a global feedback loop if the output transformers are presenting you with problems.  If this is not appealing (it isn't particularly appealing to me frankly), then triode strapping the 7695 may really be the way to go, though I don't see much information available online.

The power supply/output transformer configuration taps the primary of the output transformers to provide a source of voltage for the driver stage and screens of the output tubes. 

-PB

[Paul Joppa]

Very scary circuit:

* No power transformer means the chassis is connected directly to the power line, relying on correct house wiring to keep it at the neutral instead of the hot voltage, and to keep the neutral at least close to earth ground. I would not even take it out of the console without an isolation transformer in place, and a grounded chassis.

* A single rectifier means it draws a net DC current from the power line, and thus will unbalance the power transformer of anything else connected to that half of the power panel. It also feeds rectifier reverse recovery spikes directly into the power line where they can be piked up by any other audio equipment nearby.

Now, questions:

>> 1.)  Why does it seem to need some treble roll-off between the driver and power stages? ...

Hard to say; there seem to be many equalizations in the circuit, which may compensate the source, transformer, speaker, etc. Not knowing the characteristics of the output transformer or speakers, I'd want to measure the frequency response, possibly even at intermediate points, to make a guess as to why each one is there.

>> 2.)  Are C57 and C62 (the caps across the output transformer primaries) there to provide negative feedback?  When I remove them entirely, there is audible distortion in the high frequencies.  As I increase the value, I get relatively more bass response and less treble (does not sound good), and when I decrease the value, the amp sounds more open and lively than with the stock value, but distortion creeps into the treble.

It is not feedback. It may be to prevent transformer leakage inductance and/or speaker impedance rise from rendering the plate circuit unstable. Beyond that, see above.

>> 3.) Can anyone explain the output transformer primary tap/power supply arrangement to me?  I have gathered from internet research that this is some kind of hum-cancelling arrangement, but I don't understand it.

Yes, a portion of the output transformer serves as a choke in the power supply, where the hum that it reduces in that application cancels some of the residual hum at the power supply output. It also cancels some of the DC in the primary winding, allowing a smaller and cheaper output transformer. Since you have changed the power supply capacitor values, the hum balance no longer works as intended.

>> 4.)  The 7695's that were in it to start with are fairly quiet, but when I put any other 7695 it's noticeably noisier.  Is this just tube variability among manufacturers/tubes, or is there something about the circuit I could adjust to improve it?

Since the only change is different tubes, it does see likely that it's due to tube differences. Without knowing what the differences are, or what the noise is like (hum/buzz/hiss/rushing/rumbling) it's difficult to guess what might be happening.

[Daikini]

Thanks for the input.  That's a lot to think about. 

Yes, I went to a 0.22 uf coupling cap.  Lots more bass that way.  I think these things were designed so as not to reproduce enough bass to cause the record to skip.  Many are the limitations of having speakers and record player in the same cabinet.

I hadn't considered what this circuit would do to the mains power and other devices.  I did notice that when running it on an isolation transformer, the transformer made an audible hum (from the transformer, not the speakers) which was quite loud.  Is this a result of the net dc draw?

So if I want to get this thing to perform well and safely as a stand-alone amp, I need to come up with a power transformer and full-wave rectifier that will deliver the specified voltage at the first filter cap, right? 

If I want to play with global negative feedback, I connect one side of the secondary of the output transformer to the input of the circuit via a resistor, correct?  Do I need to ground the other side of the secondary?  How do I determine the ballpark resistor value?  What wattage rating resistor will I need?  I have seen designs that connect the NFB to the grid of the driver tube, and that connect to the cathode.  How do I decide which is appropriate in this case?

Thanks!

[Paul Joppa]

Yes, the buzz of the isolation transformer is a result of the DC draw. (And congrats for your care to use it!)

As usual, every solution brings its own problems. You can replace the single diode with a fullwave bridge (four diodes) to eliminate the DC. But that will leave the filament power fluctuating, which may induce hum. In the original circuit the sensitive 12AX7 has its filament at the neutral, or sort-of-grounded end of the 120v power.

All circuits are loops, so you must connect the output at two points to get feedback. Ground is usually used as one of them, common to the source and the load.

Negative feedback can be calculated, but it is complicated to do. At this point, the best bet would be to find a similar circuit that has feedback and copy it, then experiment with changing component values and phaseing. If the phase is wrong, it will oscillate. Even if the phase is right, if the component values are wrong there's a good chance it will oscillate. That's why you will probably need to experiment.  :^)

[Daikini]

I put a full wave bridge in, and it worked nicely.  I mounted the isolation transformer on the chassis (a hammond mfg box I've put this thing in), oriented 90 degrees from the OPTs, and got a hum that starts immediately when I turn the power switch on,  and is still there with the tubes taken out.  So it must be magnetic coupling between the transformers.  It's always something!  I guess isolation transformers are not designed with this application in mind.  Maybe a torroid iso transformer would be better?

I also started playing with negative feedback, and I see what you mean about experimenting.  I found a schematic online for what is essentially the same circuit that used negative feedback applied to the cathode.  If the cathode resistor is well bypassed, wouldn't the bypass caps just send the negative feedback signal straight to ground?  I found some other designs that split the cathode resistor into a large bypassed resistor and a smaller one, and apply the feedback at the junction of the two.  I didn't really want to mess with the cathode resistor/ bypass cap arrangement I have now, because it works well, so I decided to apply feedback to the grid. 

I have a 250K stereo pot on the input, and I figured applying feedback to the input side of the pot would form a voltage divider to control how much feedback goes in.  In order to sort out the phasing, I tried connecting an input pot terminal directly (no resistor) to one of the OPT secondaries and grounding the other, and I found that one side of the OPT resulted in loud squealing feedback/oscillation, and the other resulted in... nothing.  Total silence, turned all the way up, source plugged in.  I had assumed that I would get oscillation from both sides, as even the negative side would surely be too much signal with no resistor, so I was confused by this.  Could it be that the circuit just happens to put out enough signal to reduce the gain to zero?  I kept this polarity and added resistors.  Low value (2.2 K) brought the sound back but with a great reduction in gain.  At about 15K I couldn't tell the difference between open loop and closed loop.  I have 10Ks in there now, and I think the frequency response is flatter than without them, and the noise seems somewhat lessened, although I haven't measured it. I was hoping that a NFB loop would decrease some of my magnetic coupling hum-- is that possible?

Am I on the right track, or am I just shunting some of my input signal to ground through the transformer?

[Paul Joppa]

Well, you're experimenting and learning stuff, so you're on the right track in that sense. At this point you should look for a textbook (IMHO) to make sense of what you are learning. I usually recommend a text from a community college technician class, rather than a university engineering class - it's more useful in the real world. Sometimes there are "electronics for physicists" classes with good real-world texts as well.