1. Time for some upgrades in server hardware and software to enhance security and take AK to the next level. Please contribute what you can to sales@audiokarma.org at PayPal.com - Thanks from the AK Team
    Dismiss Notice

Rx For the Magnavox 8800 Series

Discussion in 'Tube Audio' started by dcgillespie, Apr 24, 2018.

  1. s-petersen

    s-petersen Scott Subscriber

    Messages:
    4,572
    Location:
    Middletown,New York
    I am rebuilding a 175 for another AKer, The schematic in the beginning is the one I am following, without controls. In this text it says a 10 meg resistor, on the schematic a 3.3m. Near the end of the thread it mentions the 3.3m again, so I am guessing it's 3.3m. Is this still used in the non control version?
     

     

    Please register to disable this ad.

  2. dcgillespie

    dcgillespie Fisher SA-100 Clone Subscriber

    Messages:
    9,709
    Location:
    Ball Ground, GA
    Yeah, that was written about as clear as mud, right? Sorry about that. The point I was poorly trying to make is that when the AC Balance control is included, it is executed just as the schematic shows, using a 3.3M resistor. When the control is omitted, the tap should be comprised of two matched 470K resistors, with the bottom resistor having a 10M resistor in parallel with it.

    Sorry for the confusion. I hope that helps!

    Dave
     
  3. s-petersen

    s-petersen Scott Subscriber

    Messages:
    4,572
    Location:
    Middletown,New York
    Thanks Dave!
     
  4. Kidmoe

    Kidmoe AK Subscriber Subscriber

    Messages:
    335
    Location:
    Southeast Mass.
    In the interest of keeping the stock look as much as possible, how would I wire up a simple single bias adjustment such as that in the modified 9300? I would like to use the stock speaker connectors for test points and install a single pot in the space where the hum pot used to be, again much like the basic modified 9300 has. I would be starting with known, well matched output tubes so the limited bias adjustment should still provide some benefit, though not as much as the full adjustments outlined by Dave.
     
  5. dcgillespie

    dcgillespie Fisher SA-100 Clone Subscriber

    Messages:
    9,709
    Location:
    Ball Ground, GA
    Kid -- May I assume that you're looking to build the modified version using cathode bias with the original OPT? If so, then the only way that can be accomplished is to revert (nearly) back to the cathode circuit of the original design. That is, on the modified schematic, the 470K grid return resistors for the output tubes would return to ground instead of to the balancing circuit. The four cathode terminals would then be connected back together again. From the four cathodes, there would be a single 100 uF/25 volt cathode bypass cap connected to ground. From the cathodes, there would also be one side of a 25Ω 2 watt pot, with the wiper and other side of the pot tied together, and connected to a 100Ω 5 watt resistor, and the other side of the resistor being grounded. Your test points would be ground, and the junction of the pot and resistor. Adjust the pot for 16.0 vdc across the test points (or 100Ω resistor). If you want to reverse the action of the pot, simply connect the wiper of the pot to the outside terminal connected to the cathodes, rather than the resistor.

    I hope this helps!

    Dave
     
  6. cozido512

    cozido512 New Member

    Messages:
    24
    Dave - could you please post the schematic with the floating paraphase inverter? I may have a chance to breadboard it on my bench next week, and see if I get the same result as you (with another OPT though, since I don't have a Z-565).
     

     

    Please register to disable this ad.

  7. Schmidlapper

    Schmidlapper Well-Known Member

    Messages:
    761
    Location:
    McConnellsburg, Pennsylvania
    Post #15, Details, the first attachment is the current schematic. It is easily overlooked because it is a plaine white thumbnail and blends in.

    Bill
     
  8. cozido512

    cozido512 New Member

    Messages:
    24
    Thanks, but that’s the modified schematic with the pentode output and the paraphase inverter, not the one Dave is currently working on, which has the ultralinear output and the floating paraphase inverter.
     
  9. dcgillespie

    dcgillespie Fisher SA-100 Clone Subscriber

    Messages:
    9,709
    Location:
    Ball Ground, GA
    co -- The pertinent changes I made to the driver for the fixed/cathode bias, pentode/UL tests were that the output tube grid return resistors were reduced to 220K (this because of operating the output stage with fixed bias), while the plate load resistors for the driver stage were reduced to 150K each to help accommodate the increased load that the reduced grid return resistors represented. The AC balance circuit remained unchanged, while the NFB resistor became 1 KΩ, with a 10 KΩ resistor paralleling the 1000 uF inverter stage cathode bypass cap -- this to accommodate both the required driver bias for its new operating conditions, and to produce approximately the same NFB level (~12 db) when using the Z-565 transformers (in pentode mode) with FB taken from the 8Ω tap. The rest of the changes were to the output stage DC Bias/Balance circuits to accommodate fixed bias operation, and to maintain proper stability and response when using the Z-565 transformer -- both of which have no significant impact on the effect noted.

    Dave
     
  10. cozido512

    cozido512 New Member

    Messages:
    24
    Cool, I will see if I can duplicate what you see, albeit with a different OPT...
     
  11. dcgillespie

    dcgillespie Fisher SA-100 Clone Subscriber

    Messages:
    9,709
    Location:
    Ball Ground, GA
    If I may, what transformer are you using and do you know what the screen tap is (% of winding)?

    Dave
     

     

    Please register to disable this ad.

  12. cozido512

    cozido512 New Member

    Messages:
    24
    It’s 4.8k, with 4 & 8 Ohms secondaries, and the screen tap at 20%. I will load the 8 Ohm tap with a 16 Ohm resistor, so Ra-a is 9.6k, not optimal, but should be close enough to see what’s going on.
     
  13. dcgillespie

    dcgillespie Fisher SA-100 Clone Subscriber

    Messages:
    9,709
    Location:
    Ball Ground, GA
    Agreed, although with less screen tap percentage, the effect may be somewhat reduced, but should still be enough to provide evidence of the effect over that of straight pentode operation.
     
  14. dcgillespie

    dcgillespie Fisher SA-100 Clone Subscriber

    Messages:
    9,709
    Location:
    Ball Ground, GA
    For those following along with the issue I've discovered wherein a Floating Paraphase Inverter produces unequal clipping when directly driving a UL output stage, I can now provide the following additional information on this subject:

    1. Use of a lower impedance inverter tube such as a 12AT7 in the floating paraphase position helps matters slightly, but does not entirely eliminate the effect.

    2. The application of global NFB acts to accentuate the problem, as might be expected. When the top output tube clips, the FB signal acts to correct the condition by applying more signal to that side of the push-pull circuit, in an effort to restore the missing crest of the sine wave. This of course causes the clip to be even more accentuated.

    3. With no global FB connected and close observation, the onset of clipping is actually seen to occur at the same time on both crests of a sine wave, but the characteristics of the clipping on each wave crest are seen to be very different, with the clip on the bottom of the wave very pronounced, while the clipping on the top of the wave is very soft and hardly noticeable at first.

    4. The drive impedance itself seems to be rather unimportant, as driving the output stage directly from a push-pull OPT (then acting as a push-pull driver transformer) results in identical clipping as that event is approached, achieved, and exceeded. However, if a conventional 12AT7 driver stage (for example) is inserted between the driver transformer and the output stage, clipping is again still seen to be identical on the tops and bottoms of the waveform. This shows that the impedance level itself is of limited importance (meaning that it can be either high or low), but that it's being equal is of significant importance.

    These results would tend to doom the Floating Paraphase inverter design when it is used to directly drive a UL output stage while seeking the ultimate in UL performance. Efforts to apply FB around the top inverter section to lower its drive impedance produced limited results, as the drive impedance is in fact lowered, but the move also increases the drive load on that section as well, producing (again) imbalance.

    At this point then, unless any new additional information comes along showing otherwise, a Floating Paraphase Inverter -- when used to directly drive a push-pull audio output stage -- should only be used to drive straight pentode or triode output stages, as the imbalance produced when driving a UL output stage is notable, and significant.

    I welcome any information that might be provided from others doing their own tests regarding this matter.

    Dave
     
    Last edited: Jun 25, 2018
  15. cozido512

    cozido512 New Member

    Messages:
    24
    Thanks for sharing your results, unfortunately, I won't be able to breadboard the circuit as planned - an unexpected business trip has just popped up. But it seems the difference in the plate resistances of the floating paraphase inverter is just too great to overcome when directly driving the output tubes in UL mode, by my estimation, the upper tube has ~80k, and the bottom tube has only ~1.7k - a whopping 47x difference! Oh well, learned something new today.
     
  16. gadget73

    gadget73 junk junkie Subscriber

    Messages:
    36,764
    Location:
    Southern NJ
    So in UL mode, is the input impedance of the 6V6 grid lower vs pentode mode and thats what causes the inverter to be unable to drive it properly? If so, does it get even lower if you triode strap a pentode or use an actual triode? Just asking for my own understanding here since its just not something I've ever seen discussed before.
     

     

    Please register to disable this ad.

  17. 6DZ7

    6DZ7 Super Member

    Messages:
    2,176
    What about some FB brought in from the UL tap of the bottom output ala the ST-70? I've always wondered why that was thought to improve anything that the GNFB was lacking. Laurent used that type of FB in a 7199/7189 amp but left the UL to the screen totally undone.
     
    Last edited: Jun 25, 2018
  18. dcgillespie

    dcgillespie Fisher SA-100 Clone Subscriber

    Messages:
    9,709
    Location:
    Ball Ground, GA
    Agreed. But old fashioned breadboarding is just so passé anymore: Any old slab of wood will do!

    SAM_2461.JPG

    As was shown earlier then, the standard Paraphase inverter (Magnavox's original design) actually provides a better -- as in equal -- drive impedance scenario when directly driving a UL output stage than the Floating Paraphase design does, and therefore makes for a "better" inverter in that instance. This is particularly true for the design of the 8800, since that unit does not include a cathode bypass cap for the upper AF Amplifier section of the inverter stage (as is used in the 175/185 chassis), which acts (among other things) to lower the output impedance of that inverter section, again creating an imbalance. As well, it was also shown that to achieve the low distortion capabilities that UL operation provides, the matter of accurate drive balance is also of some importance, which is not a strong suit of the standard Paraphase inverter design. As a result, more testing needs to be done to see if a UL output stage is any more sensitive to drive imbalance, versus that of it's plain-jane pentode mode cousin, relative to distortion generated. But resolving that issue is another question for another time, and more properly the topic of a separate thread dedicated to that issue.

    Until then, if maximum performance is the goal, then UL operation of any of the popular Magnavox push-pull designs -- even with a really high quality OPT -- may just in fact represent a step backwards rather than forwards, unless the appropriate equipment is available to adjust the unit for optimum performance. For now then, in the interest of trying to keep this project even remotely close to its original form, the amplifier will be left in fixed bias pentode mode, and as such, represents a very fine 20 watt 6V6 stereo amplifier, that few other designs using this tube could even approach matching. It's a very far cry from its original humble bargin basement beginnings, and as such with some finality now, allows a proper schematic drawing to be plugged into the to-do list of things to finish up for this project.

    So just when you think that every aspect of vacuum tube amplifier design is settled, something new appears -- or something that certainly seems to be new anyway since I have found no previously published information on the findings of this work -- either directly, or in passing as comments of a larger text -- discussing the principles, application, integration, and/or installation of Ultra-Linear output stages to maximum benefit in audio amplifier design work. Work will continue of course, but for now, comments are always welcomed.

    Dave
     
  19. dcgillespie

    dcgillespie Fisher SA-100 Clone Subscriber

    Messages:
    9,709
    Location:
    Ball Ground, GA
    Gadget -- It's not the input impedance of the 6V6 when operated in UL mode that is of concern, it is the unequal drive impedance of the Floating Paraphase inverter that is at issue. As the output stage reaches the onset of clipping, it attempts to start drawing grid current. During this time, the unequal drive impedances presented to the control grids -- in conjunction with the UL signal presented at the screen grids -- causes the tubes to clip unevenly. When the screen grids are grounded with respect to AC (pentode operation), then all else being equal, the clipping then otherwise becomes equal in every aspect from both tubes.

    6 -- In fact, I use that approach in the final (pentode) built of the unit, even though the UL taps are not used for their primary purpose. The effect of such feedback works to either eliminate, or greatly minimize any step networks needed, and usually installed at the plate of the AF Amplifier stage. In Hafler/Laurent's designs, they used a pentode AF Amplifier stage, whose output impedance at the plate is always very high. While step networks at that location can be very effective in achieving proper HF stability, they also tend to cause a greater than normal increase in HF THD in the finished design, due to the significant load placed on the tube. By applying some FB from the appropriate UL tap, that load is either eliminated or greatly reduced, allowing the same level of stability to be achieved, but with far less rise in 20 kHz THD in the process.

    Hafler remarked that the FB applied from the UL tap worked to balance out the unequal drive characteristics of the Cathodyne phase inverter, which of course is simply not possible. But due to the uniqueness of that FB path at the time, and discussion of the Cathodyne's unequal drive characteristics being a hot design topic in the press of the day, the statement made for a great marketing ploy that time has shown he clearly used to good advantage.

    Dave
     
  20. 6DZ7

    6DZ7 Super Member

    Messages:
    2,176
    YES! Thank you. That little tidbit keeps slipping my mind. I have that reference from Hafler's article. It was such a short mention. I remember discussing that in the forum somewhere ages ago, too. Brain and brain, what is brain?
     

Share This Page