Tony,
I might try this as the ALL 6SN7 configuration had me glued to my seat.
So, I am very keen to see if I can achieve a well balanced sound with ZERO FEEDBACK, if possible.
What is a “tad more plate current” in this scenario - and how best to achieve it?
Thanks Tony.
So, reducing cathode resistance increases anode current.
Ohms law at work - of course - I was just unsure how much adjustment to make and wasn’t sure about the impact on the unorthodox “phase-splitter” functions shared between the first stage 6SN7 and the 2nd stage 6SN7.
Worth noting too, that the second 6SN7 has always been a 6SN7 - factory configuration. It is the first position 6SN7 that was formerly a 6SL7, that I recently changed - to perform better with ZERO GNFB.
That said, can I assume reducing R6 from 15K to 12K would provide benefit to BOTH factory configuration (6SL7, 6SN7, 6550 150K GNFB) and ALSO this revised configuration (6SN7, 6SN7, 6550 ZERO GNFB)?
Or would the reduced cathode resistor value ONLY provide benefit to the new configuration with ZERO GNFB?
So, reducing cathode resistance increases anode current.
Ohms law at work - of course - I was just unsure how much adjustment to make and wasn’t sure about the impact on the unorthodox “phase-splitter” functions shared between the first stage 6SN7 and the 2nd stage 6SN7.
Worth noting too, that the second 6SN7 has always been a 6SN7 - factory configuration. It is the first position 6SN7 that was formerly a 6SL7, that I recently changed - to perform better with ZERO GNFB.
That said, can I assume reducing R6 from 15K to 12K would provide benefit to BOTH factory configuration (6SL7, 6SN7, 6550 150K GNFB) and ALSO this revised configuration (6SN7, 6SN7, 6550 ZERO GNFB)?
Or would the reduced cathode resistor value ONLY provide benefit to the new configuration with ZERO GNFB?
Tony, as above, I'm just asking if this new value ONLY applies to the new configuration (6SN7 x 4, ZERO GNFB) or will it be equally beneficial for the original configuration? (6SL7/6SN7, 150K GNFB)
Last edited:
I've been re-reading an older thread about the contentious BEWITCH 6550 circuit and how its Phase Splitter operates.
I confess that I still do not understand much of it.
That being the case, what I want to be sure about, is that removing GNFB and changing preamp tube cathode resistor values does not mess with the complex Phase Splitter / Inverter stages on this amp.
With this in mind, I have collated several of the most insightful comments from that 2017 diyAudio thread.
Worth noting that the late (and sorely missed) DF96 contributed to this discussion, adding further clarification, for those who operate at his level.
Generalizations and Calculations on triodes & triode wired pentodes and triode wired beam power tubes
What appears to be simple is not always that simple.
Changing the plate load resistor, RL changes the gain of a triode.
6SN7 plate impedance rp = 7,000 Ohms; u = 17
6SL7 plate impedance rp = 44,000 Ohms; u = 70
If the cathode has a bypass capacitor (for this rule), the gain is: u x (RL/(RL + rp))
17 x (RL/(RL + rp))
Or
70 x (RL/(RL + rp))
Plug in RL = 15k, then Plug in RL = 12k
For some of my circuits, I start with RL = 3 x rp. Then the gain is u x 3/4 (u x 0.75)
If RL = rp, than gain is u x 1/2 (0.5)
12AX7 u = 100 rp = 62,500 Ohms. If RL also equals 62,500, the stage gain is 50. But the next stages grid resistor is in parallel with RL, so the plate sees a lower total load, and gain will be less than 50.
Gain x Bias voltage = Maximum plate swing (dynamic range)
Low bias voltage and low gain (combined) will limit the ability to drive the next stage.
As RL becomes smaller versus rp, the distortion increases.
Increasing plate current reduces rp (less Ohms at higher current).
One way to increase plate current without reducing RL Ohms, is to increase the B+ at the top of RL.
There are lots of interactions here.
(I probably forgot some other factors)
Please, if any readers notice an error here, do post a correction.
What appears to be simple is not always that simple.
Changing the plate load resistor, RL changes the gain of a triode.
6SN7 plate impedance rp = 7,000 Ohms; u = 17
6SL7 plate impedance rp = 44,000 Ohms; u = 70
If the cathode has a bypass capacitor (for this rule), the gain is: u x (RL/(RL + rp))
17 x (RL/(RL + rp))
Or
70 x (RL/(RL + rp))
Plug in RL = 15k, then Plug in RL = 12k
For some of my circuits, I start with RL = 3 x rp. Then the gain is u x 3/4 (u x 0.75)
If RL = rp, than gain is u x 1/2 (0.5)
12AX7 u = 100 rp = 62,500 Ohms. If RL also equals 62,500, the stage gain is 50. But the next stages grid resistor is in parallel with RL, so the plate sees a lower total load, and gain will be less than 50.
Gain x Bias voltage = Maximum plate swing (dynamic range)
Low bias voltage and low gain (combined) will limit the ability to drive the next stage.
As RL becomes smaller versus rp, the distortion increases.
Increasing plate current reduces rp (less Ohms at higher current).
One way to increase plate current without reducing RL Ohms, is to increase the B+ at the top of RL.
There are lots of interactions here.
(I probably forgot some other factors)
Please, if any readers notice an error here, do post a correction.
Hyperman75,
Wrapping Negative Feedback around circuitry that includes a phase splitter does not exactly cancel the amplitude mismatch . . .
There is always a small residual of that amplitude mismatch.
Why do you think I use a really long tail (LTP) with a very good CCS (Constant Current Sink) in the cathode circuit?
Because it has less amplitude mismatch, and less distortion to begin with, before negative feedback is applied.
Oh, did I mention I do not use forms of negative feedback, except for Local or Intrinsic feedback (LTP stage, Triode wired Beam Power tubes, Push Pull, and balanced stages with cathodes tied together to a single self bias resistor that is Not bypassed with a capacitor).
Uh Oh, I just admitted that I no longer use individual self bias resistors with bypass caps.
That breaks my own rule (well not exactly, because I use Extremely Well Matched tubes).
Generalization:
If a distortion is 2%, and you apply 6dB negative feedback, the residual distortion is now 1%.
I am not aware of anyone applying 120dB of negative feedback to get rid of [almost all] of the residual distortion.
Wrapping Negative Feedback around circuitry that includes a phase splitter does not exactly cancel the amplitude mismatch . . .
There is always a small residual of that amplitude mismatch.
Why do you think I use a really long tail (LTP) with a very good CCS (Constant Current Sink) in the cathode circuit?
Because it has less amplitude mismatch, and less distortion to begin with, before negative feedback is applied.
Oh, did I mention I do not use forms of negative feedback, except for Local or Intrinsic feedback (LTP stage, Triode wired Beam Power tubes, Push Pull, and balanced stages with cathodes tied together to a single self bias resistor that is Not bypassed with a capacitor).
Uh Oh, I just admitted that I no longer use individual self bias resistors with bypass caps.
That breaks my own rule (well not exactly, because I use Extremely Well Matched tubes).
Generalization:
If a distortion is 2%, and you apply 6dB negative feedback, the residual distortion is now 1%.
I am not aware of anyone applying 120dB of negative feedback to get rid of [almost all] of the residual distortion.
Last edited:
Thanks @TonyTecson
So, would 10K for R6 work okay?
(I just happen to have a nice 10K resistor on-hand...)
My calculations suggest that reducing R6 to 10K, the 6SN7 is still NOT working very hard - but not sure how this would impact the first 6SN7 tube.
So, would 10K for R6 work okay?
(I just happen to have a nice 10K resistor on-hand...)
My calculations suggest that reducing R6 to 10K, the 6SN7 is still NOT working very hard - but not sure how this would impact the first 6SN7 tube.
Well, no grown-ups protested, so I reduced R6 to 10K using a nice 4W non-magnetic wirewound resistor - again bypassed with a .01uF polypropylene cap.
I’ve powered up the amp and there were no red-plates and no smoke.
Initial measurements confirm that the 6SN7 plate voltage relating to R6 has been significantly reduced - and voltage across R6 itself has reduced by 5V.
I’m not sure what impact this change will have on the sound.
I will report.
I’ve powered up the amp and there were no red-plates and no smoke.
Initial measurements confirm that the 6SN7 plate voltage relating to R6 has been significantly reduced - and voltage across R6 itself has reduced by 5V.
I’m not sure what impact this change will have on the sound.
I will report.
Last edited:
Listening to the new value for 6SN7 cathode resistor R6 at 10K.
6SN7 plate voltage has reduced from ~240V to ~180V. Voltage across R6 has reduced from 60V to 56V.
I'm not sure if those changes are dramatic or negligible.
The bass seems to have improved but is now a bit boomy.
Otherwise, vocals are sounding a bit squashed and muffled.
Gone is the sublime "open" sound produced by the previous configuration.
So, a backward step so far.
I guess part of the problem here, is due to my lack of understanding about this circuit, I am unsure about what I am making the changes for. That is to say, I am not sure what the changes are hoping to achieve.
Obviously, improved sound quality is the overall goal - but reducing the second stage 6SN7 cathode resistor value (alone) has not helped meet this objective.
Once again, time has made a complete liar of me.
A few hours more play using the new configuration of the (2nd position) 6SN7 R6 shared cathode resistor - now reduced from 15K to 10K - and things are starting to sound very good.
I've made no other changes since my last post; just a few more hours of play and burn-in of the new resistors. The amp remains dead-quiet at idle - even at full volume - which is very pleasing for a 100% Chinese-made mid-power tube amp, with ZERO GNFB.
Without an O-Scope this is all highly subjective of course, but my ears are now telling me that the "open-ness" is returning and that the top-end is a bit more extended and sparkly.
Also very noticeable, is the increased dynamics.
The bass is no longer boomy and is pretty much on par with the previous ZERO GNFB configuration using all 6SN7 at factory circuit values.
One last comment on the sound is that it seems a little more honest - if not quite as relaxed-sounding - as the previous configuration. This means a little more detail and a little less honey-ed tones.
I would guess that this means that the (2nd position) 6SN7 is now at a more linear operating point - and perhaps I am hearing less overall distortion.
Now that R6 at 10K is sounding much better, I will have a think about reducing the resistor values of R9 and R10 - which are currently at 30K - reducing to ~22K - as (also) previously suggested by @TonyTecson (Thank you, Tony...
)
A few hours more play using the new configuration of the (2nd position) 6SN7 R6 shared cathode resistor - now reduced from 15K to 10K - and things are starting to sound very good.
I've made no other changes since my last post; just a few more hours of play and burn-in of the new resistors. The amp remains dead-quiet at idle - even at full volume - which is very pleasing for a 100% Chinese-made mid-power tube amp, with ZERO GNFB.
Without an O-Scope this is all highly subjective of course, but my ears are now telling me that the "open-ness" is returning and that the top-end is a bit more extended and sparkly.
Also very noticeable, is the increased dynamics.
The bass is no longer boomy and is pretty much on par with the previous ZERO GNFB configuration using all 6SN7 at factory circuit values.
One last comment on the sound is that it seems a little more honest - if not quite as relaxed-sounding - as the previous configuration. This means a little more detail and a little less honey-ed tones.
I would guess that this means that the (2nd position) 6SN7 is now at a more linear operating point - and perhaps I am hearing less overall distortion.
Now that R6 at 10K is sounding much better, I will have a think about reducing the resistor values of R9 and R10 - which are currently at 30K - reducing to ~22K - as (also) previously suggested by @TonyTecson (Thank you, Tony...
)
After days of listening to the new configuration of the (2nd position) 6SN7 R6 shared cathode resistor - now reduced from 15K to 10K - I have come to the conclusion that there is just too much "hardness", "dryness" and "stridency" in the new sound.
Despite the good levels of detail, this has made listening much less enjoyable.
In an effort to optimize the new 10K value - and perhaps (re)sweeten the tops a bit - I have since removed the .01uF polypropylene cap that previously bypassed R6. This did not help.
I have poured over 6SN7 and 6550/KT88 datasheets and am comfortable that no area of the configuration is stressing the tubes and conclude that I simply have not yet reached the (new) circuit sweet-spot - following the R6 value reduction.
On this basis, I am ordering new 22K resistors for R9 and R10, which currently remain at 30K.
I'm going to give the highly regarded AMTRANS AMRG 2W 22K resistors a try, for this spot.
Below is the current configuration that I am attempting to optimize.
All comments and feedback welcome.
Despite the good levels of detail, this has made listening much less enjoyable.
In an effort to optimize the new 10K value - and perhaps (re)sweeten the tops a bit - I have since removed the .01uF polypropylene cap that previously bypassed R6. This did not help.
I have poured over 6SN7 and 6550/KT88 datasheets and am comfortable that no area of the configuration is stressing the tubes and conclude that I simply have not yet reached the (new) circuit sweet-spot - following the R6 value reduction.
On this basis, I am ordering new 22K resistors for R9 and R10, which currently remain at 30K.
I'm going to give the highly regarded AMTRANS AMRG 2W 22K resistors a try, for this spot.
Below is the current configuration that I am attempting to optimize.
All comments and feedback welcome.
Thanks @TonyTecson
Are you recommending Schade Feedback - feeding back from the plate of 6550 to the G2 (PIN1) of the second 6SN7 - to lower the overall output impedance, in order to help towards better bass response?
Are you recommending Schade Feedback - feeding back from the plate of 6550 to the G2 (PIN1) of the second 6SN7 - to lower the overall output impedance, in order to help towards better bass response?
@TonyTecson
Sorry. I edited after your THUMBS-UP.
Should the SCHADE FB feedback to the second 6SN7 G2 or the 6SN7 Plate? Does it matter?
Sorry. I edited after your THUMBS-UP.
Should the SCHADE FB feedback to the second 6SN7 G2 or the 6SN7 Plate? Does it matter?
Sorry Tony.
As per my question above, the (second) 6SN7 point of connection is still unclear.