Buffered cathodyne vs. LTP

[octal]

I was wondering if anyone has tried a cathodyne PI followed by a gain stage style phase inverter (as found on a few old Gibsons and Silvertones) and how it compares to the standard LTP. The disadvantages I can think of are tube count and that it might be difficult to apply NFB to the buffered cathodyne. The advantage is gaining isolation between the clamping action of the output tubes grids at vg=0 and the PI stage, which should stay balanced well into clipping. I'd be interested to hear if anyone has experimented with this.

[octal]

The attached Silvertone 1433 schematic is an example of the phase inverter I'm calling "buffered cathodyne." Somebody here has to have given it a try!

[tbryanh]

The Silverface Super Twin Reverb uses a buffered phase inverter. It does this with cathode followers, not with common cathodes stages as in the Silvertone:
http://www.ampwares.com/schematics/s...180w_schem.pdf

Aiken Amps also mentions a buffered phase inverter (cathode followers) for eliminating blocking distortion and clamping action:
Add a ... cathode follower between the phase inverter and the grid of the output tubes ... This effectively isolates the output tube grid circuit from the phase inverter and its associated AC coupling, and provides a very low impedance source for the output stage. This will prevent the output stage from going into grid clamp, and will eliminate the long time constant of the AC coupling. http://www.aikenamps.com/BlockingDistortion.html

Considering the high output impedances the common cathodes stages between the splitter and the PA in the Silvertone have, it does not appear they prevent blocking distortion and clamping action or keep the drive signal to the PA balanced well into clipping.

I would guess that the driver stage in the Silvertone (the cathodyne splitter and the two common cathode stages that follow it) operates similar to the long tailed pair. It may be that is was a transitional circuit in use between the time that the cathodyne circuit was phased out and the long tailed pair was phased in.

Negative feedback could probably be applied in a way similar to the way it is done in the Bogen CHB35A public address amp:
http://www.schematicheaven.com/hifia...gen_chb35a.pdf

[Merlinb]

I was wondering if anyone has tried a cathodyne PI followed by a gain stage style phase inverter (as found on a few old Gibsons and Silvertones) and how it compares to the standard LTP. The disadvantages I can think of are tube count and that it might be difficult to apply NFB to the buffered cathodyne. The advantage is gaining isolation between the clamping action of the output tubes grids at vg=0 and the PI stage, which should stay balanced well into clipping. I'd be interested to hear if anyone has experimented with this.

While it is true that the cathodyne doesn't behave very well when overdriving, buffering its outputs with more valves seems like overkill to me. Ok, you can get more swing from the buffers, and therefore more overdrive, but if you're using small-bottle output valves, you hardly need all that swing. And if you're using big bottles, then the LTP is so much easier. Also, it's very easy to isolate the cathodyne from the oputput stage by using big grid-stoppers on the power valves, like 100k or more, which can leave you with a very well behaved cathodyne and no blocking distortion at a fraction of the cost!

[defaced]

The Silverface Super Twin Reverb uses a buffered phase inverter. It does this with cathode followers, not with common cathodes stages as in the Silvertone:

VHT does the same thing with it's bigger (4 x KT-88) power amps. I've got my amp set up that way now but haven't been able to really give it a good workout because I'm running 2 x KT-88.. for now.

[octal]

Those are all well-taken points. I did think of one interesting point- the even order harmonic distortion in the "buffer" stage will be cancelled out (because the buffer stage is working like a push-pull amp) so I imagine the amp's distortion spectrum would be skewed more in the direction of odd harmonics. Since there are going to be large voltage swings in the buffer stage, the distortion reduction could be significant. Whether that's desirable or audible is open to debate

On a bit of a tangent- I understand how DC coupling the output tube's grids (i.e. using transformer coupling or a DC coupled CF) would prevent blocking distortion & grid clamping, but I'm completely confused about why it would be advantageous to have a lower output impedance AC coupled PI circuit vis-a-vis blocking/clamping/bias shift. It seems like a lower impedance source would just charge the coupling capacitors more fully when grid clamping occurs on the positive swing, thus driving the tubes further into cutoff on the negative swing.

Any thoughts?

[hasserl]

I think Ken Gilbert used cathode followers between the PI and the power tube grids on his BAGA, he reported great power tube distortion.

[Merlinb]

why it would be advantageous to have a lower output impedance AC coupled PI circuit vis-a-vis blocking/clamping/bias shift. It seems like a lower impedance source would just charge the coupling capacitors more fully when grid clamping occurs on the positive swing, thus driving the tubes further into cutoff on the negative swing.

You're right- a low output impedance doesn't reduce blocking.

[GibsonLover]

The last four tubes in that are similar to what Genelex did on one of their amps. A friend at work is building one today - I hope to hear it tomorrow night. I'll try and report back, but he's building it to replace his Dynaco MKIII, so I may not hear a guitar through it. The circuit might better be used for Hi-Fi, as you are correct about the distortion. The buffer is a differential amplifier which excels in doing just that. Check the Genelex schematic for adding feedback - you don't have to go to the PI for local feedback. Returning to the input stage is global, and again, more of a Hi-Fi application.

[pdf64]

'On a bit of a tangent- I understand how DC coupling the output tube's grids (i.e. using transformer coupling or a DC coupled CF) would prevent blocking distortion & grid clamping, but I'm completely confused about why it would be advantageous to have a lower output impedance AC coupled PI circuit vis-a-vis blocking/clamping/bias shift. It seems like a lower impedance source would just charge the coupling capacitors more fully when grid clamping occurs on the positive swing, thus driving the tubes further into cutoff on the negative swing.'

See Aiken -
http://www.aikenamps.com/BlockingDistortion.html

I think the point is that grid clamping wouldn't have nearly as big an effect, because a low impedance source could stay more linear when the power tube grids start to conduct. But the coupling caps would have to sized to cope with the ~1k ohm conducting grid load, rather than for the normal 220k grid bias feed resistors. Otherwise the low end at high power would be reduced. Peter.

[octal]

'On a bit of a tangent- I understand how DC coupling the output tube's grids (i.e. using transformer coupling or a DC coupled CF) would prevent blocking distortion & grid clamping, but I'm completely confused about why it would be advantageous to have a lower output impedance AC coupled PI circuit vis-a-vis blocking/clamping/bias shift. It seems like a lower impedance source would just charge the coupling capacitors more fully when grid clamping occurs on the positive swing, thus driving the tubes further into cutoff on the negative swing.'

See Aiken -
http://www.aikenamps.com/BlockingDistortion.html

I think the point is that grid clamping wouldn't have nearly as big an effect, because a low impedance source could stay more linear when the power tube grids start to conduct. But the coupling caps would have to sized to cope with the ~1k ohm conducting grid load, rather than for the normal 220k grid bias feed resistors. Otherwise the low end at high power would be reduced. Peter.

I'm a bit skeptical about this... I did read the article and unless I'm missing something, I don't think that Aiken does suggest a low impedance drive source as a solution for the various problems caused by grid conduction. It's not possible to achieve AB2 operation with capacitive coupling to the output tubes of the amp.

[tbryanh]

#6 "... [Why would it] be advantageous to have a lower output impedance AC coupled PI circuit[?] It seems like a lower impedance source would just charge the coupling capacitors more fully when grid clamping occurs on the positive swing, thus driving the tubes further into cutoff on the negative swing. "

It does appear that a low impedance source would charge the coupling capacitors more fully.

When the coupling capacitors charge, because of the low impedance source of the driving signal, the grids do not cause clamping. In other words, the low impedance does not drop much voltage across itself from the high grid currents, so the signal peaks are not attenuated during grid conduction like they are with the high impedance source.

Because the capacitors charge more fully with the low impedance source, the signal to the power tubes rides on a "negative voltage" instead of riding on "zero volts". In single ended operation, this would tend to cause the negative half cycle to clip, but with push pull operation, it does not.

With push pull operation each tube conducts for less than a full cycle, so it is normal for the tubes to go into cuttoff. When one tube is cuttoff, the other tube is conducting, so no clipping occurs. The OT just sums everything together.

The issue here is crossover distortion. When the signal to the power tubes rides on a "negative voltage" instead of riding on "zero volts", there could be more crossover distortion.

[tbryanh]

It seems clear in the Silvertone amp that the common cathode stages between the splitter and the power tubes are used to create a greater drive signal to the power tubes, but why are the cathode followers used in the Silverface Super Twin Reverb amp?

[Merlinb]

but why are the cathode followers used in the Silverface Super Twin Reverb amp?

Perhaps to improve the linearity of the power stage? (they were designed to be clean, almost pseudo-hifi amps after all). The CFs would buffer the LTP from the heavy load of the power valves' grid-leak resistors, and so reduce THD, and also helping to maintain balance and reduce intermod' distortion if the output valves clip.