Ultra linear and phase inverter drive

[Chuck H]

then probably after a while you won't hear anything because big bottles are expensive so you'll be broke (and probably deaf from the extreme volume.)

Excellent.

[Tage]

Well I do have room for another dual triode. But now I need to clarify if I can simple ac couple my PI to a triode acting as a cathode follower or should I couple it to a dc coupled ganistage-cathode follower ala SVT, not that that is the kinda power I'm going for or anything... but it is about the only design I know of with the low impedance drivers.

[trobbins]

Why don't you overdrive the earlier stages in order to attain controllable distortion high-level output, without trying to achieve the overdrive character from the PI or output stage. You can set your levels up so as to keep the PI and output stage relatively 'clean' but still achieve the same sort of output power rating.

[loudthud]

Well I do have room for another dual triode. But now I need to clarify if I can simple ac couple my PI to a triode acting as a cathode follower or should I couple it to a dc coupled ganistage-cathode follower ala SVT, not that that is the kinda power I'm going for or anything... but it is about the only design I know of with the low impedance drivers.

It's better to DC couple the cathode followers to the power tubes so you can keep the power tube grid impedance low and avoid really huge coupling caps. The disadvantage is that you need a fairly stiff minus supply, maybe -100V or -150V at 20mA or so. Apply the adjustable bias to the follower grids and you won't need 2 watt bias pots.

[bob p]

... but it is about the only design I know of with the low impedance drivers.

Fender Studio Bass

Super Twin & Super Twin Reverb are similar.

[Gingertube]

I've intimated before that my ideas for guitar amps are "infested" with HiFi Amp design principles BUT this is one area where the HiFi ideas can really help.
Many commercial bass amps have reliability issues directly relating to grid 1 impedances being way too high.
The way to address this is mosfet source followers to buffer the PI outputs.
Direct couple the source followers to the output tube grids with just an appropriate grid stop in that feed. That takes firm control of the output tube grids, addresses bias purturbations and thermal run away issues caused by grid current. It also shunts grid current noise to ground.
Put the AC coupling caps between the PI and the source followers and apply the bias voltage to the gates of the mosfets.
This makes the PI perform much better too, not only is the PI more lightly loaded but that load is now a Class A load. This will definitely help for the "usual" schmitt (diff amp) splitter BUT it MASSIVELY improves performance of a Cathodyne (split load) PI where balance depends upon equal loads at the anode and cathode. The Class A loading of the PI also eliminates blocking distortion problems.
Yes you need a negative rail. A minimum of 3 times the bias voltage is what is required (I usually use 4 to 5 times the bias voltage).
Someone above stated that you are then at the mercy of the negative rail noise (that is, the negative rail noise will appear at the output tube grid).
That is true to a degree. Looking into the source follower output pin (the source pin) you will see a lowish impedance to 0V (signal ground). The amount of noise which appears at the top of the source follower load can be determined by simple voltage divider theory. The "bottom" arm of the divider is the AC impedance looking into the source pin. The top arm of the divider is the source follower load resistor back to that -ve rail.
Since you are already adding in a bit of SS (the mosfet source follower) then the answer is to add in another 2 bucks worth of SS and replace the source follower load resistors with a current sources. The VERY high AC impedance of the current source (compared with a resistor) will divide that -ve rail noise down to insignificance. For the current sources I generally use a "Ring of Two" transistor current source. A BC547C as the bottom transistor with an MJE340 on top. There are of course better and more sophisticated current sources but these simple ones are more than adequate for the task.

Can I give you a schematic - well not a bass amp but I used the same scheme in a HiFi Amp for which the schematic is here:
EL84 Amp - Baby Huey - Page 61 - diyAudio (See post #602)
It is "infested" with a few other HiFi'ish ideas (forced balance for example) but the current source loaded source follower buffers between the PI and the output tubes is there for all to see. My current HiFi Amp is this circuit but with 6V6 Outputs and 6SL7 splitter.

Hope this gives you some ideas, stimulates some discussion or at least a "Pick up your HiFi Marbles (if you can find them) and go home" comment.

Cheers,
Ian

[Gingertube]

Another +1 to add to Chuck H.
The lower frequency roll off is set by the output tube internal impedance rp and the primary inductance of the output tranny. Ultralinear connection reduces that rp by a factor of about 3 times shifting the low frequecy roll off 3 times lower. The problem then becomes core saturation at low frequencies and to address that BIG iron is what you need.
Cheers,
Ian (Cork Sniffer ??)

[soundmasterg]

Another +1 to add to Chuck H.
The lower frequency roll off is set by the output tube internal impedance rp and the primary inductance of the output tranny. Ultralinear connection reduces that rp by a factor of about 3 times shifting the low frequecy roll off 3 times lower. The problem then becomes core saturation at low frequencies and to address that BIG iron is what you need.
Cheers,
Ian (Cork Sniffer ??)

Big iron is definitly king for tube bass amps. The early Sunn 2000S (UL with a cathodyne inverter I might add) with the Dynaco transformers (before they switched to Schumacher) has a huge OT and will amplify cleanly down to 10 Hz! But that kind of OT costs dearly these days.

Greg

[Steve Conner]

I think we all agree that the bigger the OT, the better!

However, I saw a Wireless World article somewhere arguing that ultralinear allows a smaller OT for the same bass response. The lower Rp allows the transformer to have lower inductance without losing bass. You can then stack the laminations in bigger chunks to get more distributed air gap, and that helps with the saturation.

If applying the Paul Ruby mod to a fixed bias amp, the relevant voltage is twice the bias voltage. If you have your power tubes biased at -35V, the signal flaps them between 0 and -70, so your zeners have to be 70V or greater.

Fender's 300PS and 400PS were AB2 all the way. The power tube grids were driven by a transformer the size of a Deluxe OT, which in turn was driven by a triode connected 6V6 or 6L6. That's what I call an output stage! The same topology was used on a miniature scale in the Musicmaster Bass practice amp.

[jrfrond]

I'd like to point out that the old Marshall Major design accomplished all of this nicely with a UL OT and a pretty routine LTP 12AU7 PI (chosen for it's lower plate impedance) driving 4- KT88's. You don't REALLY need cathode followers unless you are worrying about blocking distortion, which is not happening unless the amps' is operation at constant overload. Lookee here: http://www.webphix.com/schematic hea..._bass_200w.pdf

It's also worthwhile to note that while the coupling caps to the power tube grids are at a fairly-high .47uF, the grid load resistors have been reduced to 68K to reduce the time constant.

[Tage]

Yeah the Major has been a grood referance for me, but I was puzzled by it's PI design, as the second grid seems to be connected back to a gain stage before the one that drives the other half of the 12AU7, is that a drawing error?

I also have been using the amp book calculators and was wonder what Kinda time constant I should be shooting for, as far as what is an acceptable bias excursion recovvery time, or is bias excurssion a different thing? I was gonna use 57K grid load resistors and .1uf caps as they should allow enough low end, maybe a .22uf

[R.G.]

The way to address this is mosfet source followers to buffer the PI outputs.
Direct couple the source followers to the output tube grids with just an appropriate grid stop in that feed.

I like the idea, Ian. In fact, it's pretty much what I've proposed here MOSFET Follies in 2000, and flogged on this forum a number of times.

You're right about worrying about hifi heresies - the current source is the obvious choice, but I could never bring myself to go ahead and say it. I caught enough flak here and other places over actually saying to introduce silicon into an tube amp.

People who have tried it have reported back that it works well.

[jrfrond]

Yeah the Major has been a grood referance for me, but I was puzzled by it's PI design, as the second grid seems to be connected back to a gain stage before the one that drives the other half of the 12AU7, is that a drawing error?

In fact, it is. The junction of the 1M grid-load resistor after the volume control and .047 coupling cap should be joined to the junction of the 2.7K and 100K cathode resistors. The stage just after the volume control is actually configured as a cathodyne PI, which then drives a long-tail pair PI. Cool and simplistic design.

[Chuck H]

I like the idea, Ian. In fact, it's pretty much what I've proposed here MOSFET Follies in 2000, and flogged on this forum a number of times.

You're right about worrying about hifi heresies - the current source is the obvious choice, but I could never bring myself to go ahead and say it. I caught enough flak here and other places over actually saying to introduce silicon into an tube amp.

People who have tried it have reported back that it works well.

I remember it being proposed here... By you. I thought it was well recieved and valid. The only rub is that it doesn't fit into the ideal of an "all tube" circuit (for marketing) and it has become passe in light of new digital sampling amps (main signal OR effects) and the uber gain preamp into a 100W power amp modern designs that all the wankers use. Better power amp distortion is sort of a tube purist crank kind of thing that MOSFETs don't really fit into. And the metal kids don't crank their amps or they couldn't get that V notch, mid scoop wanker tone. For builders like us it's great. But I don't know if it has a place in any non technically inclined demographic.

[exclamationmark]

That's probably true. Blackstar use a lot of solid state components in their amps, and have copped a lot of flack from prospective buyers, because their not 'all tube'. Though for the more technically inclined home builder, I think on some level we just like to bring our haphazard creations on stage and gloat. Sometimes I tell people my amp is solid state just for fun (it's not far from the truth either!)