Hi all,
I'm working out a project to use a couple of Sunn transformers I salvaged to make my first 2xKT88 amp. I wanted to try a tetrode-mode sort of thing, which got me thinking about grid current screwing with the cathodyne balance and/or introducing blocking distortion. I'm following the Sunn 200S for the basis for the power amp section and I'll add a switch to disconnect the screen taps and re-route in a 300vDC screen supply. In case you're wondering, I planned to get the 300vDC by using an adapted VVR circuit to drop the 560vDC down into range so I don't kill the screens.
Well, I did the whole DC coupled negative rail thing in my last amp and I'm not sure I want go that far this time. I think if I were to do so I'd need to add a smaller transformer for the negative rail as the Sunn transformers seem ill-suited to this... so I'm wondering if some benefit is to be had by just buffering the two outputs of the cathodyne with mosfet source followers (DC coupled to the anode and cathode) and AC couple the sources to the grids with a normal bias arrangements. The idea being the ability to lower the grid resistors from 100k to maybe 25k to lower the time constant for re-charging the bias caps after grid current at clipping - without loading down the phase inverter. I figured this, combined with larger than normal gird stoppers on the KT88 (maybe up to 25k) perhaps the stability of the bias voltage might be improved. Here's a scribble:
That's the big question.... the little question is how could I select the source resistors for the followers? Since the both sit at very different potentials, should I use different source resistor values or would that affect their output impedance or some other performance factor? I was thinking maybe 3mA of bias each, so the lower follower would get a 33k resistor and upper would get a 110k. Or should I pick a value the works of for both but let them run at higher/lower currents?
Thanks for any help.
Brian
I'm working out a project to use a couple of Sunn transformers I salvaged to make my first 2xKT88 amp. I wanted to try a tetrode-mode sort of thing, which got me thinking about grid current screwing with the cathodyne balance and/or introducing blocking distortion. I'm following the Sunn 200S for the basis for the power amp section and I'll add a switch to disconnect the screen taps and re-route in a 300vDC screen supply. In case you're wondering, I planned to get the 300vDC by using an adapted VVR circuit to drop the 560vDC down into range so I don't kill the screens.
Well, I did the whole DC coupled negative rail thing in my last amp and I'm not sure I want go that far this time. I think if I were to do so I'd need to add a smaller transformer for the negative rail as the Sunn transformers seem ill-suited to this... so I'm wondering if some benefit is to be had by just buffering the two outputs of the cathodyne with mosfet source followers (DC coupled to the anode and cathode) and AC couple the sources to the grids with a normal bias arrangements. The idea being the ability to lower the grid resistors from 100k to maybe 25k to lower the time constant for re-charging the bias caps after grid current at clipping - without loading down the phase inverter. I figured this, combined with larger than normal gird stoppers on the KT88 (maybe up to 25k) perhaps the stability of the bias voltage might be improved. Here's a scribble:
An externally hosted image should be here but it was not working when we last tested it.
[/URL]
That's the big question.... the little question is how could I select the source resistors for the followers? Since the both sit at very different potentials, should I use different source resistor values or would that affect their output impedance or some other performance factor? I was thinking maybe 3mA of bias each, so the lower follower would get a 33k resistor and upper would get a 110k. Or should I pick a value the works of for both but let them run at higher/lower currents?
Thanks for any help.
Brian
Last edited:
Pick the value for the same current through both.
Or, you could use current sources in place of the resistors.
Or, you could use current sources in place of the resistors.
How is this power amp going to be implemented? By this I mean will it be used for the playback of recordings or for guitar amplification?
This will ultimately be for instrument amplification. Not a HIFI ultra low distortion application. I would like to avoid a huge increase in parts count but are there any good examples of current source implementations on high voltage fets? I've done current sources for solid state circuits with BJTs before, so I'm familiar with the principle... one thing I've always wondered is how one settles on a "proper" idle current. If all we're talking about is a BJT for each source, an emitter resistor for each BJT, and a voltage reference for the the two bases then I could go for that if it would make this more robust.
Last edited:
If this if for musical instrument use then don't use current sources; they can lead to some really awful overdrive sounds (OK that's probably not going to happen in this particular instance because the MOSFETs can't be driven to cut-off, but they are best avoided nevertheless. Resistors are cheaper. In fact, in this application the argument for CCSs is weak even for a hifi amp) Just choose some resistors that give approximately the same drain current in each MOSFET; it's not critical.
3mA would result in close to 1W dissipation in the lower MOSFET, which is pretty hot for a TO-220 with no heatsink, so you might want to be a bit more conservative (or use heatsinks).
I assume you are already aware that you'll want to add gate stoppers and gate-source protection zeners, if they're not already built into the MOSFETs.
Last edited:
Since this is an MI amp thread I will move it to Instruments & Amps where I think you will get more feedback from the fellows into instrument amps.My only thought would be to suggest an LTP based phase splitter as they seem generally to be preferred for MI applications.
Have you though about using an interstage transformer?
Is this for guitar or bass?
I have had minimal play time through a class AB2 amplifier that I know of so I can't speak of it's tonal or playing characteristics that much. Instrument amps are a preference thing and everyone has different tastes. I personally like really stripped down raw amps that are for the most part engineered poorly.
BUT one of my rigs that I use rarely is a preamp I made running into a Mcintosh MC30 power amp through a 2x12 cabinet loaded with 103db efficient alnico drivers. I use this rig with a lot of effects and it has a super big clean sound to it, I get my overdrive from a pedal with this setup. The Mcintosh MC30 I believe has 12AX7 cathode followers driving the power tubes. Like I said I do not push the power amp into distortion so I am unsure how it sounds but for cleans it's really defined and articulate without being too sterile.
Is this for guitar or bass?
I have had minimal play time through a class AB2 amplifier that I know of so I can't speak of it's tonal or playing characteristics that much. Instrument amps are a preference thing and everyone has different tastes. I personally like really stripped down raw amps that are for the most part engineered poorly.
BUT one of my rigs that I use rarely is a preamp I made running into a Mcintosh MC30 power amp through a 2x12 cabinet loaded with 103db efficient alnico drivers. I use this rig with a lot of effects and it has a super big clean sound to it, I get my overdrive from a pedal with this setup. The Mcintosh MC30 I believe has 12AX7 cathode followers driving the power tubes. Like I said I do not push the power amp into distortion so I am unsure how it sounds but for cleans it's really defined and articulate without being too sterile.
Thanks, Merlin. I'll shoot for 2mA I guess? A 47k resistor on the lower follower and a 175K on the upper follower seem about right for that. I should have mentioned the potentials involved... B+ is 425, Anode is 340, cathode is 90. Maybe you dug up the schematic to know that already. I just ordered a bunch of 12v zeners for gate-source protection.
Yes but I dare not do it. Seems too complicated to pull off correctly. The cathodyne setup I'm working from seems pretty well suited to the application, I'm just worried about what happens when the output clips - hence the question about buffers.
For Musical Instrument use I do not see the point in buffering the cathodyne.
Yes, the balance is badly upset when the anode and cathode loads are not equal and that hppens as soon as the output stage moves out of pure Class A.
This can be handled:
First RECOMMENDED reading is what the Wiz has to say.
The Valve Wizard
Next is a discussion on output impedance at the anode and cathode which is were
most guys loose the plot and the arguement goes "west".
For equal loads on Anode and cathode on the Cathodyne (I've use symbol "u" = tube mu)
Zout = RL.ra/RL(u+2)+ra
The ra term on the bottom line is insignificant compared to RL(u+2) term so drop it. Then the RL terms top and bottom lines cancel leaving
Zout approx = ra/u+2
At typical values of u (>=20) u+2 approx = u, so simplify again
Zout approx = ra/u = 1/gm
So for equal loads at the anode and cathode the Zout at both anode and cathode is approximately 1/gm or about 650 Ohms for a 12AX7
If driving output stage directly The equal loads on Anode and Cathode will NOT be equal if:
1) Output stage strays out of Class A (When a tube cuts off it has no gain so Miller capictance will change, particularly with Triode Mode Output, less so with Ultralinear and less so again in Pentode Mode)
2) Output Stage strays into grid current (trying to overdrive the output tube). When this happend the load resistance presented by the output tube grid drops significantly.
Note also that the value of the anode and cathode load resistors does not appear in that calculation.
Big Amps tend to use 82K or 100K anode and cathode load resistors whereas smaller amp tend to use 56K.
This is purely for ease of swinging the required grid drive voltage. All of the EL34 and 6L6 Amps I've seen with cathodyne splitters use 100K or 82K (example ORANGE MATAMP)
Smaller Amps with EL84 or 6V6 which do not require the same voltage swing, typically use 47K or 56K. That is, they are swinging the same current in the cathodyne as the bigger amps but generating less voltage drive into the output tubes.
We have said that the problems come when we cut off an output tube (minor problem) or drive it into grid current (major problem) thus unbalancing the cathodynes anode and cathode loads.
Calculating Overdrive Effects
If MATHS gets on your wick then skip to the SUMMARY below:
If the Anode load drops significantly then:
Zout cathode = RL+ra/(u+2) x ra/RL The ra/RL term insignigicant so
Zout cathode approx = RL+ra/u+2 At usual values of u
Zout cathode approx = RL/u + ra/u = RL/u + 1/gm
***** That is Zout increases by RL/u *******
If the cathode load drops significantly then:
Zout anode = RLxRL(u+1)+RL.ra / RL(u+2)+ra
RL squared (u+1) is much larger than RL.ra and RL(u+2) is much larger than ra so
Zout anode approx = RLxRL(u+1)/RL(u+2)
and at reasonable values of u
******* Zout anode approx = RL ********
SKIP TO HERE:
SUMMARY:
As the loads on Anode and cathode become unbalanced (as when overdriving the output tubes) then
Zout anode increases from 1/gm toward RL (from 650 Ohms toward 100 KOhms in the MATAMP)
Zout cathode increase from 1/gm by maximum factor of RL/u (from 650 Ohms toward 650 + 100K/100 = 1650 Ohms in the MATAMP)
The imbalance from an over driven output will therefore be worse with anode and cathode loads of 100K vs 56K but the difference between the anode and cathode output impedance in overdrive is so dramatic it doesn't really matter if its between 650 Ohms and 100K or between 650 ohms and 56K, Also the effect of this will be asymetrical, affecting one side of the push pull and therefore introducing large gobs of 2nd harmonic distortion.
The trick here is to use large grid stops on the output tubes to reduce the loading effects on the cathodyne. Take those output tube grid stop resistors way up to say 47K. This is effectively tuning the amount of 2nd harmonic distortion you get in overdrive. Note this is what the Wiz says too.
Note also that The Wiz recommends a big grid stop on the cathodyne triode too to prevent grid current from disturbing the operating point when the cathodyne triode is itself overdriven.
I also had some email correspondence with Kevin O'Connor on this subject and he told me that most of his commercial amps run cathodyne splitters. He agreed that the way to address overdrive is to use large grid stops on the output tubes but said he had never found it necessary to use large grid stops on the cathodyne where he ran gridstops in the more usual 15K to 22K range.
Is this post useful or should I pick up my marbles (if I can find them) and go home.
Cheers,
Ian
Yes, the balance is badly upset when the anode and cathode loads are not equal and that hppens as soon as the output stage moves out of pure Class A.
This can be handled:
First RECOMMENDED reading is what the Wiz has to say.
The Valve Wizard
Next is a discussion on output impedance at the anode and cathode which is were
most guys loose the plot and the arguement goes "west".
For equal loads on Anode and cathode on the Cathodyne (I've use symbol "u" = tube mu)
Zout = RL.ra/RL(u+2)+ra
The ra term on the bottom line is insignificant compared to RL(u+2) term so drop it. Then the RL terms top and bottom lines cancel leaving
Zout approx = ra/u+2
At typical values of u (>=20) u+2 approx = u, so simplify again
Zout approx = ra/u = 1/gm
So for equal loads at the anode and cathode the Zout at both anode and cathode is approximately 1/gm or about 650 Ohms for a 12AX7
If driving output stage directly The equal loads on Anode and Cathode will NOT be equal if:
1) Output stage strays out of Class A (When a tube cuts off it has no gain so Miller capictance will change, particularly with Triode Mode Output, less so with Ultralinear and less so again in Pentode Mode)
2) Output Stage strays into grid current (trying to overdrive the output tube). When this happend the load resistance presented by the output tube grid drops significantly.
Note also that the value of the anode and cathode load resistors does not appear in that calculation.
Big Amps tend to use 82K or 100K anode and cathode load resistors whereas smaller amp tend to use 56K.
This is purely for ease of swinging the required grid drive voltage. All of the EL34 and 6L6 Amps I've seen with cathodyne splitters use 100K or 82K (example ORANGE MATAMP)
Smaller Amps with EL84 or 6V6 which do not require the same voltage swing, typically use 47K or 56K. That is, they are swinging the same current in the cathodyne as the bigger amps but generating less voltage drive into the output tubes.
We have said that the problems come when we cut off an output tube (minor problem) or drive it into grid current (major problem) thus unbalancing the cathodynes anode and cathode loads.
Calculating Overdrive Effects
If MATHS gets on your wick then skip to the SUMMARY below:
If the Anode load drops significantly then:
Zout cathode = RL+ra/(u+2) x ra/RL The ra/RL term insignigicant so
Zout cathode approx = RL+ra/u+2 At usual values of u
Zout cathode approx = RL/u + ra/u = RL/u + 1/gm
***** That is Zout increases by RL/u *******
If the cathode load drops significantly then:
Zout anode = RLxRL(u+1)+RL.ra / RL(u+2)+ra
RL squared (u+1) is much larger than RL.ra and RL(u+2) is much larger than ra so
Zout anode approx = RLxRL(u+1)/RL(u+2)
and at reasonable values of u
******* Zout anode approx = RL ********
SKIP TO HERE:
SUMMARY:
As the loads on Anode and cathode become unbalanced (as when overdriving the output tubes) then
Zout anode increases from 1/gm toward RL (from 650 Ohms toward 100 KOhms in the MATAMP)
Zout cathode increase from 1/gm by maximum factor of RL/u (from 650 Ohms toward 650 + 100K/100 = 1650 Ohms in the MATAMP)
The imbalance from an over driven output will therefore be worse with anode and cathode loads of 100K vs 56K but the difference between the anode and cathode output impedance in overdrive is so dramatic it doesn't really matter if its between 650 Ohms and 100K or between 650 ohms and 56K, Also the effect of this will be asymetrical, affecting one side of the push pull and therefore introducing large gobs of 2nd harmonic distortion.
The trick here is to use large grid stops on the output tubes to reduce the loading effects on the cathodyne. Take those output tube grid stop resistors way up to say 47K. This is effectively tuning the amount of 2nd harmonic distortion you get in overdrive. Note this is what the Wiz says too.
Note also that The Wiz recommends a big grid stop on the cathodyne triode too to prevent grid current from disturbing the operating point when the cathodyne triode is itself overdriven.
I also had some email correspondence with Kevin O'Connor on this subject and he told me that most of his commercial amps run cathodyne splitters. He agreed that the way to address overdrive is to use large grid stops on the output tubes but said he had never found it necessary to use large grid stops on the cathodyne where he ran gridstops in the more usual 15K to 22K range.
Is this post useful or should I pick up my marbles (if I can find them) and go home.
Cheers,
Ian
Last edited:
Hi Gingertube,
Thats quite helpful, as usual thanks for taking the time.
I do feel like I should start closer to the beginning because maybe my goals/assumptions aren't obvious.
I have the transformers from a 2xKT88 Sunn amp, so I want build something close to the Sunn design. 6AN8 pentode/triode for a driver/cathodyne into KT88 UL. The 200S pre-drive, phase inverter, and output topology are perfect for what I want to do except for 2 things:
I'd like to try a tetrode mode of operation w/ the KT88 (Just 'cuz I've never done it) and I'd rather not use the 6AN8 pentode/triode - prefer some more generic 12A*7 tubes. This introduced some problems: Grid current at clipping (the Sunn design looks to never get to the point of overdriving the output at all when in UL, tetrode needs about half the signal to get full output) and trying the get similar open-loop gain from a 12AX7 pre driver and the same output swing from a 12AX7 cathodyne.
I borrowed the Orange Graphic MKII driver/splitter as that seemed like a drop in replacement. Probably not as much open-loop voltage gain as a pentode driver but I was thinking I'd boot strap the cathode side of the splitter back to the a split plate load on the driver (2 170K resistors) if I needed more.
In terms of grid current, the thing I was really trying to avoid was bias shift/blocking distortion. I'm not a fan and I had trouble with this in some JCM 800 style amps I built about 10 years ago. My last build was a 2xEL84 (http://www.diyaudio.com/forums/tubes-valves/251288-direct-coupled-output-stage-dc-bias-balance.html) amp with a fully DC coupled output and the overdrive tone is excellent… no splatty swirly anything. I don't think I have a way to set up a real negative rail on this build so I wondered if the following would't help maintain a stabile bias and prevent shift:
Large grid stoppers
Small bias feed resistors
A low DC resistance to ground on the phase splitter side of the coupling caps
With a max grid resistance of 100k for the KT88 I figured I'd need to reduce the bias feeds anyway as the grid stoppers go up. The lower they go the harder they are to drive and my pause splitter now has 100K resistors (thought the math above reveals that this isn't as big a deal as it seemed before) so some kind of buffer made sense. I chose Mosfets because I want to favor linearity in this case.
Sorry for going back to the beginning and starting over. That's just part of my process if you know what I mean. What do you think about where my head is at? Too much?
Thanks again,
Brian
Thats quite helpful, as usual thanks for taking the time.
I do feel like I should start closer to the beginning because maybe my goals/assumptions aren't obvious.
I have the transformers from a 2xKT88 Sunn amp, so I want build something close to the Sunn design. 6AN8 pentode/triode for a driver/cathodyne into KT88 UL. The 200S pre-drive, phase inverter, and output topology are perfect for what I want to do except for 2 things:
I'd like to try a tetrode mode of operation w/ the KT88 (Just 'cuz I've never done it) and I'd rather not use the 6AN8 pentode/triode - prefer some more generic 12A*7 tubes. This introduced some problems: Grid current at clipping (the Sunn design looks to never get to the point of overdriving the output at all when in UL, tetrode needs about half the signal to get full output) and trying the get similar open-loop gain from a 12AX7 pre driver and the same output swing from a 12AX7 cathodyne.
I borrowed the Orange Graphic MKII driver/splitter as that seemed like a drop in replacement. Probably not as much open-loop voltage gain as a pentode driver but I was thinking I'd boot strap the cathode side of the splitter back to the a split plate load on the driver (2 170K resistors) if I needed more.
In terms of grid current, the thing I was really trying to avoid was bias shift/blocking distortion. I'm not a fan and I had trouble with this in some JCM 800 style amps I built about 10 years ago. My last build was a 2xEL84 (http://www.diyaudio.com/forums/tubes-valves/251288-direct-coupled-output-stage-dc-bias-balance.html) amp with a fully DC coupled output and the overdrive tone is excellent… no splatty swirly anything. I don't think I have a way to set up a real negative rail on this build so I wondered if the following would't help maintain a stabile bias and prevent shift:
Large grid stoppers
Small bias feed resistors
A low DC resistance to ground on the phase splitter side of the coupling caps
With a max grid resistance of 100k for the KT88 I figured I'd need to reduce the bias feeds anyway as the grid stoppers go up. The lower they go the harder they are to drive and my pause splitter now has 100K resistors (thought the math above reveals that this isn't as big a deal as it seemed before) so some kind of buffer made sense. I chose Mosfets because I want to favor linearity in this case.
Sorry for going back to the beginning and starting over. That's just part of my process if you know what I mean. What do you think about where my head is at? Too much?
Thanks again,
Brian
Quibble: this is not true. The balance is still essentially perfect even for AB or B operation. It is only upset if the tube draws grid current, but that's true of any phase splitter.
Quibble accepted - it is really only at high frequency extremes that balance is affected (minor) due to Miller Capacitance disappearing as a tube cuts off (no gain => no Miller Capacitance).
This will be at so high a frequency as to not be relevant for a MI Amp (probably not for HiFi either).
So, as you say, it is just when driving the tube into grid current we need to worry about.
Cheers,
Ian
This will be at so high a frequency as to not be relevant for a MI Amp (probably not for HiFi either).
So, as you say, it is just when driving the tube into grid current we need to worry about.
Cheers,
Ian
- Status
- Not open for further replies.