Single ended EL34 amp thoughts

[nickfl]

... So, my current thoughts are:

Plate - 330v
Screen - 318v
Cathode - 21v (270ohm resistor)... If anyone sees a major flaw in my plan, please let me know!

Lemme know if you've already explained this but I'm curious about your plate and screen voltages and specifically the delta between the two.

I don't know but I've been told...that if you get too much of a voltage drop between the plate and screen voltage you going to generate some compression that you may not want. This is something I picked up off of a HiFi thread here recently so it is hearsay but has me wondering?? In that thread there was a lot of talk that an amp with a SE design using an EL34 output might not have but a couple volts difference between the plate and the screen. It was in fact possible to have them at the same voltage or even the screen a couple volts higher than the plate and the amp performance was still considered to be satisfactory.

Disclaimer - it might not be true but it has my attention. Anybody else heard of this?

I'm sort of operating on the impression that a smallish voltage drop between the plate and screen is a common best practice. In the valve wizard single ended EL34 tutorial, he ends up with a screen voltage 12v lower than the plate voltage, so I'm going with a similar figure. I would think that the concerns that Hi-fi guys have about unwanted compression etc, would not necessarily apply to guitar amp design, and might very well be exactly what we want in our application, just as they might be upset by levels of harmonic distortion that we would consider ideal.

[martin manning]

...but if you consider a new production tube to be capable of only 20W dissipation...

But this is just "some guy on the internet" saying that. Where's the data?

I'm sort of operating on the impression that a smallish voltage drop between the plate and screen is a common best practice.

That depends on where you want to place the Vg1=0 anode curve with respect to the load line. Generally, the higher the plate voltage is the lower the screen voltage wants to be for a given level of distortion.

[tictac]

I'm at a lost why I don't see any fixed bias SE amps

I've built a fixed bias SE amp with a Bluesmaster type preamp and a 6L6GC power tube and it works great

I had a leftover 50W Marshall PT which voltages are too high for a cathode biased amp but it works fine for a fixed bias version.....

TT

[martin manning]

No reason not to. I suppose since they are usually simple low-powered amps, the the extra complexity isn't warranted. I can see where there might be some interesting sounds, second harmonic distortion typical of SE, without the squish (technical term for bias shift) of cathode bias. Not sure what you mean by PT voltage is too high for a cathode biased amp??

[tictac]

The PT voltage too high for cathode bias thing is from something Dave Funk explained years ago. It had to do with the fact (according to him) that as you get into the higher B+ range (420-500V example) a cathode biased amp needs to have a higher value cathode resistor which for a reason I can't remember causes cross-over distortion. I'm sure I'm not explaining this correctly but the actual explanation was in his Tube amp book he did years ago....

Also he was talking about a PP cathode biased amp not an SE (which cannot have cross-over distortion anyway), So I think my brain did a correlation of High B+ = not good for cathode bias.

When you think about it you don't see many (if any) cathode biased amp that operate in the 430-500V B+ range like fixed bias amps do so maybe there's something to it. A more knowledgeable tube guy than I maybe able to explain this phenomena better and even why you don't see many fixed bias SE amps...

TT

[tubeswell]

FWIW, there is something of a discussion going on over at Ampage on modern EL34 reliability and possible workarounds, which makes for interesting griping.

http://music-electronics-forum.com/t446 ... /#poststop

[dorrisant]

I'm at a lost why I don't see any fixed bias SE amps

I've built a fixed bias SE amp with a Bluesmaster type preamp and a 6L6GC power tube and it works great

TT

Agreed. I tried both fixed bias and cathode bias in the same amps and prefer the fixed bias route... but don't let the cat out of the bag or everyone will be doing it...

https://www.youtube.com/watch?v=kFOQ2JMwUe8&t=16s

Not to derail, how does using fixed bias in a single ended amp change the way you should adjust the bias (% of dissipation)? What determines whether you bias to "Class A" or not? Would you adjust the bias to 90% and let it go? That is what I've done with the amp in the above video... Hope it doesn't let the smoke out..

Merlin had some good stuff to read about both types of bias... At least that's where I think I grabbed the idea from... Oddly the page was about push-pull.

"To push the stage into Class AB we slide the Class A load line up the graph while maintaining its gradient, in exactly the same way as for the Single Ended Output Stage and see what bias voltage is required, the grey lines show this process. In fact, this is what we are doing whenever we adjust the bias on a fixed-bias amp! If we bias hot enough (i.e., push the Class A load line up far enough) it will become pure ClassA.
Of course, the bias point must not be pushed above the maximum dissipation curve, and it is usual to bias well under it to avoid damage (70% of maximum dissipation is often recommended, although anything up to about 85% would be ok).
The load line indicates we need roughly 12V bias voltage. This can either be provided via a cathode bias resistor [below right], or by applying a fixed negative voltage to the grid [below left]. If using cathode bias, each valve can be given its own bias resistor and capacitor, or they can both share the same one (of half the resistance, twice the power rating), which tends to encourage extra second harmonic distortion [below right].
Cathode bias often lends a natural compression or 'squishiness' to the sound, due to the increase in bias voltage when one valve enters Class B conditions, though the larger the bypass capacitor, the less will be this effect. A small capacitor (less than 100uF say) also increases non-linear distortion, which may be significant in hifi. Using a very large capacitor (greater than 470uF say), or using no capacitor at all, reduces this effect.
Fixed bias on the other hand, remains the same at all times. This allows maximum output power to be developed, and the reduced compression gives faithful transient response, or a stiff or 'barking' overdriven sound.
Furthermore, there is no reason why we cannot use a little fixed bias and cathode bias simultaneously to achieve the desired mix of compression and 'bark'."

Here: http://www.valvewizard.co.uk/pp.html

[Firestorm]

Fixed bias in SE creates the potential to drive the tube into cutoff if the signal is strong enough. "Class B" SE does not sound good. Self bias/cathode bias ensures that this can't happen (with a reasonable value for Rk) since as more current flows in the tube it creates a larger voltage drop in the cathode resistor to keep things in check.

[roberto]

If I were you, I would search a custom set of transformers, with the capability to run at 300, 375 and 425 V with different Raa on the OT, and I would configure the power supply à la fender the twin (http://www.blueguitar.org/new/schem/fender/the_twin.gif) to have a half voltage 1/4th power switch option. Rebiasing can be done accordingly to the bias method you'll chose.

[martin manning]

Not to derail, how does using fixed bias in a single ended amp change the way you should adjust the bias (% of dissipation)? What determines whether you bias to "Class A" or not? Would you adjust the bias to 90% and let it go? That is what I've done with the amp in the above video... Hope it doesn't let the smoke out..

A single ended amp is a Class A amp by definition (current flows over the entire signal swing). You can drive it into cut-off regardless of whether it is fixed bias or cathode bias if the negative signal swing is large enough. If that happens, it is operating in Class AB (current flows for more than half but not all of the signal swing), and it is essentially like a push-pull amp with half the tubes missing. Note that if the bias point and load impedance have been been selected for a center-biased condition, the positive signal swing will be clipped as well.

A Class AB push-pull amp operates in Class A for small signal levels, and transitions into Class AB (negative and positive signal swings alternately drive the power tubes into cut off) long before it reaches full power (where Vg1 is driven to 0 and the signal is clipped). With hotter bias, a larger signal is required to transition to Class AB. If the bias is hot enough, it could reach full power without ever transitioning into Class AB. This is what is meant by "biasing into Class A", but it is not likely that you could actually do this without exceeding Pa max at the idle point.

I would bias a single ended fixed-bias amp at ~90% Pa max as if it were cathode biased, so that the operating point and load impedance results in a center-biased condition.

[didit]

[...]
So, my current thoughts are:

Plate - 330v
Screen - 318v
Cathode - 21v (270ohm resistor)

If I am doing things correctly, this is biased to about 90% and pretty close to right in the center of the load line. The interactive data sheet says this should be around 9.5 watts output, maybe a little less in practice.

Fairly conventional "big bottle" SE, works fine. Not exceptionally tall or clean. Screen a bit conservative perhaps, but cool nevertheless.

I'm going to work on implementing this in my schematic and then start thinking about how I might want to tweak the preamp (I'm probably going to move away from some of the matchless derived elements, since there are some odd values there and I don't really understand why they work...). If anyone sees a major flaw in my plan, please let me know!

One possible flaw. Encourage sticking with some variant off your selected Matchless preamp. It's interesting for being different. Ignore all objections from the crowd on low ohm plate load. Lowered first stage gain is more than made up for by pentode following, and output impedance adds drive into the tone stack. Only weakness is pentode's high impedance driving SE power stage versus a 12AX7 PI in that original design. Cathode follower potentially necessary. Just an educated intuition, and some maths are necessary to really confirm.

Best .. Ian

[nickfl]

[...]
So, my current thoughts are:

Plate - 330v
Screen - 318v
Cathode - 21v (270ohm resistor)

If I am doing things correctly, this is biased to about 90% and pretty close to right in the center of the load line. The interactive data sheet says this should be around 9.5 watts output, maybe a little less in practice.

Fairly conventional "big bottle" SE, works fine. Not exceptionally tall or clean. Screen a bit conservative perhaps, but cool nevertheless.

I'm going to work on implementing this in my schematic and then start thinking about how I might want to tweak the preamp (I'm probably going to move away from some of the matchless derived elements, since there are some odd values there and I don't really understand why they work...). If anyone sees a major flaw in my plan, please let me know!

One possible flaw. Encourage sticking with some variant off your selected Matchless preamp. It's interesting for being different. Ignore all objections from the crowd on low ohm plate load. Lowered first stage gain is more than made up for by pentode following, and output impedance adds drive into the tone stack. Only weakness is pentode's high impedance driving SE power stage versus a 12AX7 PI in that original design. Cathode follower potentially necessary. Just an educated intuition, and some maths are necessary to really confirm.

Best .. Ian

Ok, you inspired me to go ahead and follow through on the clubman preamp. Here is a schematic updated for my power amp design with the clubman pre, I've added a gain/volume control between the parallel 12ax7 and the ef86 and slightly altered the component values around the ef86 to compensate for a lower B+ value at its supply node than the actual clubman has, but it should still be in the ballpark. I have the jacks wired for single/parallel input instead of the low/high input the clubman uses, that shouldn't make much difference to the rest of the amp, but I find it to be a more useful option. Lastly I added a raw switch to lift the ground connections on the tone stack.

Here's what I have at the moment:

Single Ended EL34 Clubman Pre.png

RE: adding a cathode follower. I was looking at another design today for a single ended amp which had an EF86 as an input stage and then a DC coupled cathode follower after it leading into a TMB tone stack and then straight into the power section. That and your comment made me wonder if I should consider adding another 12ax7 as a dc coupled cathode follower after the EF86? I wonder if this would be a good way to make up for the loss of the gain the LTP PI would add in a push pull amp, or would it be too much? Or would it make more sense to put the cathode follower before the tone controls?

[nickfl]

[...]
So, my current thoughts are:

Plate - 330v
Screen - 318v
Cathode - 21v (270ohm resistor)

If I am doing things correctly, this is biased to about 90% and pretty close to right in the center of the load line. The interactive data sheet says this should be around 9.5 watts output, maybe a little less in practice.

Fairly conventional "big bottle" SE, works fine. Not exceptionally tall or clean. Screen a bit conservative perhaps, but cool nevertheless.

I'm going to work on implementing this in my schematic and then start thinking about how I might want to tweak the preamp (I'm probably going to move away from some of the matchless derived elements, since there are some odd values there and I don't really understand why they work...). If anyone sees a major flaw in my plan, please let me know!

One possible flaw. Encourage sticking with some variant off your selected Matchless preamp. It's interesting for being different. Ignore all objections from the crowd on low ohm plate load. Lowered first stage gain is more than made up for by pentode following, and output impedance adds drive into the tone stack. Only weakness is pentode's high impedance driving SE power stage versus a 12AX7 PI in that original design. Cathode follower potentially necessary. Just an educated intuition, and some maths are necessary to really confirm.

Best .. Ian

Ok, you inspired me to go ahead and follow through on the clubman preamp. Here is a schematic updated for my power amp design with the clubman pre, I've added a gain/volume control between the parallel 12ax7 and the ef86 and slightly altered the component values around the ef86 to compensate for a lower B+ value at its supply node than the actual clubman has, but it should still be in the ballpark. I have the jacks wired for single/parallel input instead of the low/high input the clubman uses, that shouldn't make much difference to the rest of the amp, but I find it to be a more useful option. Lastly I added a raw switch to lift the ground connections on the tone stack.

Here's what I have at the moment:

Single Ended EL34 Clubman Pre.png

RE: adding a cathode follower. I was looking at another design today for a single ended amp which had an EF86 as an input stage and then a DC coupled cathode follower after it leading into a TMB tone stack and then straight into the power section. That and your comment made me wonder if I should consider adding another 12ax7 as a dc coupled cathode follower after the EF86? I wonder if this would be a good way to make up for the loss of the gain the LTP PI would add in a push pull amp, or would it be too much? Or would it make more sense to put the cathode follower before the tone controls?

Could I perhaps add a third preamp tube, as a gain stage and dc coupled cathode follower, with a switch to bypass the signal around the gain stage half, giving me the option of switching it to an ac coupled cathode follower inline after the pentode?

[didit]

10-15% unclear on the proposal. An updated schematic with the added triode pair please?

Best .. Ian

[nickfl]

10-15% unclear on the proposal. An updated schematic with the added triode pair please?

Best .. Ian

Ok... this was a lot harder to put into schematic form than I expected, but here it is. With no values for the components calculated yet, just an illustration of what I was thinking:

SE EL34 Clubman pre with CF switch.png

It seems like it should work as a way to switch between ac coupled cathode follower used to buffer between the EF 86 and the EL34s and adding a gain stage + DC coupled cathode follower for a presumably significantly higher gain in the preamp. I'm wondering if it would actually need to be a 3pdt switch to change anything else on the stage, like removing the connection from the grid leak resistor to the CF triode grid when it is set to DC coupling? I assume that if this is a workable idea, someone else has already done it, but I don't know where to look for an example.