Single Ended Parallel Lead - review my planned overhaul

[GlideOn]

Posted also on TDPRI forum, but it's slow as molasses over there getting feedback!

Seem to be lots more folks with Single Ended designs posting here!

I currently have an AX84 SEL with stock Hammond 270EX OT and 125ESE PT. My buddy and I completed it past January with a chassis and board I had ordered at least a decade ago from now-defunct Doberman Amps. Literally blew the dust off and got to work.

As you may know, it is a Cascaded preamp design ala Marshall JMP/JCM800.

Cathode Follower V2.

No Phase Invertor (not needed single ended).

Run into 6550.

It sounds great at bedroom volumes, but quickly falls apart and gets fizzy, indistinct at practice/drummer volumes.

It's pretty clear looking at the stock schematic it was never truly designed with big bottle tone in mind - lots of things look values copy/pasted from running EL84 power.



Now that I have lived with it for a few months, I've gathered up the courage to improve upon it.

Basically more power via 2x EL34 paralleled tubes (explanation below).



Firstly, the PT and OT changes:


AnTek torodial AS2T300 300v-0-300v/160ma PT.

AnTek torodial MS-30W25 single ended 30W OT.

AnTek torodial LK5H150 5H/150ma Choke.

They all came with respective metal covers and mounting screws. I'll have to drill new holes to mounts, but not a big deal when Torodials mount with single large-ish screws.


These should handle anything thrown at them for a 25w-30w design.

The EL34 are chose ln because they are the only high power big bottle tubes that will fit once these very wide transformers are installed, approx 1/4in of clearing the PT. More power than a single 6550 too.

I also opted for adding a third 12AX7 "Fake Phase Invertor" stage using one half of 12AX7. I hope it got it right, but referenced a schematic (attached #3 design below) that has been posted on here many times with other Single Ended builds. The notable change is that the Master Volume/Grid Leak is after the stage - not before - in order to extract that extra gain. I'm a big fan of Post-PI Master Volumes.

I've done many mods to amps over the years, including a 1968 Traynor YBA-1 that has JMP typology and LarMar MV.

Old Gibson GA-200 amp that has been "Trainwrecked"

A friend's 1979 Twin Reverb UL with a Lead Channel.

So I feel pretty confident in this build which has given me great pride so far.


So...



The other (extensive) changes from stock include:


- Added Grid stopper of 15k to V1 Input/Grid.

- Rework of the V1 preamp, eliminating redundant Gain 2 for fixed value.

- Gain 1 upgraded from 500k to 1M.

- Moving the Master volume to Gain 2 slot, changed to 250k value, moved the position of 5.6k Grid Stoppers.

- Added "Fake PI" stage to V3, uses half of 12AX7 as minor gain boost but to primarily add Presence Control to former MV slot at end of preamp controls.

- Presence Control 25K lin (in former MV slot).

- NFB of 56k, sent to 8ohm jack.

- Added new V5 power tube slot, parallel to V4 complete with own 1K/5w Screen Grids.

- Beefed up shared Screen Filter from 10uf to 22uf (500v).

- Added individual 330r/100uf cathodes to V4, V5.

- Changed 1.2k B+ Dropper to 4.7k.

- New 5H/150ma Torodial Choke in place of 100r.

- Beefed up Mains Filter to 100uf.

- Separated power amp grounding to new ground under PT.

- Added individual Grid Stoppers to V4, V5.

- Full Bridge Rectifier diodes (going to use terminal strips bolted to spare hole but spread out for easier viewing). Necessary as new PT lacking Center Tap.


Going into an old Mesa MKIV chassis, hanging upside down. A little small for the chassis, but the Black Shadow C90 sounds great and it is portable. Also picked up the whole thing for $70 which was a steal.

Let me know what you all think!

[cdemike]

Are you committed to single-ended? If you went push-pull with EL84s through an efficient speaker, I’d think you should be able to keep up with most of the drummers I’ve played with. If you have a John Bonham type of drummer, you may need more, but I’m not certain parallel EL34s will get you much more than a typical push-pull pair, especially since you’re not committed to class A with push-pull. With that faux phase inverter it seems you’re a lot of the way there sparing an OT replacement, which you’ll likely need anyway if you want to change from EL84 to EL34.

If you want to stay with single-ended, I think your plan is totally reasonable and will likely give you more power and headroom than your current setup. I do wonder if you might have more success with a single KT88 or 6550 at very high voltage and a screens supply that is about 1/2 B+ rather than the typical 3-5V lower than the plates. Probably more importantly I think you’ll want a beefier PT. 300-0-300 @160mA may be working pretty hard with 2x EL34s in class A. If I’m understanding you right, it seems like the big goal would be to get more headroom with is directly related to your power supply. What’s your B+ measuring currently?

[Stevem]

Keep in mind that if you want to keep a class A output stage along with more power you can just place more tubes in parallel.

If one is not loud enough then for your ears to experience double the volume you need to go to 4 tubes.

[GlideOn]

Are you committed to single-ended? If you went push-pull with EL84s through an efficient speaker, I’d think you should be able to keep up with most of the drummers I’ve played with. If you have a John Bonham type of drummer, you may need more, but I’m not certain parallel EL34s will get you much more than a typical push-pull pair, especially since you’re not committed to class A with push-pull. With that faux phase inverter it seems you’re a lot of the way there sparing an OT replacement, which you’ll likely need anyway if you want to change from EL84 to EL34.

If you want to stay with single-ended, I think your plan is totally reasonable and will likely give you more power and headroom than your current setup. I do wonder if you might have more success with a single KT88 or 6550 at very high voltage and a screens supply that is about 1/2 B+ rather than the typical 3-5V lower than the plates. Probably more importantly I think you’ll want a beefier PT. 300-0-300 @160mA may be working pretty hard with 2x EL34s in class A. If I’m understanding you right, it seems like the big goal would be to get more headroom with is directly related to your power supply. What’s your B+ measuring currently?

Brother, thank you for the reply however I already have the more powerful PT and OT covered via the Torodials seen the pic. Currently a smidge above 350v on the plate and slightly lower on screen. The Torodial 300-0-300 should give an additional 50-60v healthy volts AND required current for parallel EL34s. Any larger and it wouldn't fit the chassis - it's literally at the edges and 1/2 away from power tube socket(s).

I'm not really a fan of small bottled EL84 tone. I rather like the 6550 tone, but there is no single-ended Output Transformer in existence that will run two of them in parallel at the desired output, which defeats the point of using them. EL34 is the next power tube that will run at the desired voltage and current/bias.

And quite aware push pull makes more power pound per pound but that is not the intention. I full intend for a 20W to 25W heavy bastard, badass Class A combo.

Keep in mind that if you want to keep a class A output stage along with more power you can just place more tubes in parallel.

If one is not loud enough then for your ears to experience double the volume you need to go to 4 tubes.

Again, that's precisely the intention as illustrated in the (somewhat long) post with a pair of EL34s in parallel. Beyond that, I am limited by chassis space and output transformers lower than 2.5k ohm and single ended up to 30w, which is about as big and powerful as they get.

[wpaulvogel]

A pair of EL34’s at 420 volts and an output transformer with 1750 will get you about 22 watts clean RMS.

[cdemike]

You don't need two to get the power you're describing. You should be able to get comparable power from a pair of EL34s in class A as compared with a single 6550. See: https://shop.ehx.com/catalog/addimages/ ... ng-sol.pdf

That's especially true if you push the B+ by changing your rectifier setup. A full wave bridge rectifier with a resistive input load should provide ~530 at the plates. You could really punish that 6550 if you went with a capacitive input load and a full wave bridge rectifier, but I'm not certain how long it would survive that. Probably a situation better suited for KT120. I will say that you could just machine the chassis to mount the output tubes somewhere else. It seems to me that you could build the amp you wanted in the first place that way and run as many parallel 6550s as you wanted.

there is no single-ended Output Transformer in existence that will run two of them in parallel at the desired output

Edcor makes at least two OTs that would work under the conditions I described, including:
https://edcorusa.com/products/cxse25-2- ... 8dc7&_ss=c

In any event, I get the sense that you're not interested in other approaches. I think this amp should work as-designed.

BTW, with respect to:

AnTek torodial AS2T300 300v-0-300v/160ma PT.

300-0-300 implies a center tapped PT. That PT is technically not center tapped as you noted but it can be wired as center-tapped if you run the secondaries in series with each other and ground the connection between them. So you could run full wave rectification if you wanted.

[GlideOn]

All, please review the V2 layout in the PDF attachment at end of post #1.

All of these wonderful points such as the full wave bridge rectifier, parallel EL34s have been illustrated there.


I already have the Torodials. They are in the picture of post #1. Larger transformers will not fit the 16x8 chassis nor clear speaker, not to mention they are absurdly expensive. Torodials were chosen not only for their footprints, but superior pound-per- pound in power output, magnetic flux and for the OT, much superior frequency range verging on HiFi.

No EIs I looked at came close to their performances and price points.

6550s will not fit with these, nor have the needed current/voltage demands supplied by even the upgraded PT. EL34s were chosen primarily due to being able to fit with less than 1/2in room next to PT. They are also cheaper to replace in pairs than even a single 6550 these days.

From AnTek's own site:

PT:
https://www.antekinc.com/as-2t300-200va ... ansformer/

OT:
https://www.antekinc.com/ms-30w25-30w-s ... ansformer/

Choke:
https://www.antekinc.com/lk-5h150-toroidal-power-choke/

A pair of EL34’s at 420 volts and an output transformer with 1750 will get you about 22 watts clean RMS.

Yep! Nearly dead-on to my calculations. I could post the load lines, but to summarize it was 400-410v at the plates, less so at the screens which still have a dedicated 5k/5w dropper. I also ballpark updated the main B+ dropper from 1.2k to 4.7k to keep the preamp in-spec with the desired voltage.

330r/100uf cathodes for each tube.

I saw a lot of 330r in many single ended EL34, so I just copied it and it ended up being the Goldilocks value - just a hair on the conservative side giving me around 82%-86% plate dissipation, just shy of max 90% for Class A. 300r or 270r would likely put it over the 90%.

[jabguit]

You don't need two to get the power you're describing. You should be able to get comparable power from a pair of EL34s in class A as compared with a single 6550.

Ummm....what am I missing here?

[GlideOn]

You don't need two to get the power you're describing. You should be able to get comparable power from a pair of EL34s in class A as compared with a single 6550.

Ummm....what am I missing here?

At first I thought he meant *to* instead of "two," but then the statement makes even less sense within the context of my post.

I have no idea either

[cdemike]

You don't need two to get the power you're describing. You should be able to get comparable power from a pair of EL34s in class A as compared with a single 6550.

Ummm....what am I missing here?

The post was saying that a single 6550 would be as powerful as a pair of EL34s under the right circumstances. Specifically if the B+ was derived from series secondaries (600VAC, non-CT), and if rectifier was correspondingly reconfigured as a full wave bridge with a resistive load. That should be clear if you read the Tung Sol data sheet linked above.

[GlideOn]

You don't need two to get the power you're describing. You should be able to get comparable power from a pair of EL34s in class A as compared with a single 6550.

Ummm....what am I missing here?

The post was saying that a single 6550 would be as powerful as a pair of EL34s under the right circumstances. Specifically if the B+ was derived from series secondaries (600VAC, non-CT), and if rectifier was correspondingly reconfigured as a full wave bridge with a resistive load. That should be clear if you read the Tung Sol data sheet linked above.

Thanks for clarifying, though irrelevant as single 6550 simply will not fit with the larger transformers I have gotten, that should be clear if you read the post(s) above

[martin manning]

Look here for calc's to make some quick estimates: http://www.valvewizard.co.uk/se.html

With 420V Va and 2x EL34 (50W Pa max) Za for center bias = 420^2/50 = 3528 ohms
Assuming 0V Va min, power output will be Irms*Vrms, Irms = 420/3528 * 0.707 = 0.0842A, Vrms = 420 * 0.707 = 297V, Pout = 25W
Screen voltage should be about half of anode voltage, which is nice if you run a FWB and a CT PT. Using a stacked reservoir you can run the CT to the middle of the stack and get Vg2 from there.

With 420V Va and 1x 6550 (42W Pa max) Za for center bias = 420^2/42 = 4200 ohms
Assuming 0V Va min, power output will be Irms*Vrms, Irms = 420/4200 * 0.707 = 0.0707A, Vrms = 420 * 0.707 = 297V, Pout = 21W
Again screen voltage should be about half of anode voltage.

[cdemike]

Ummm....what am I missing here?

The post was saying that a single 6550 would be as powerful as a pair of EL34s under the right circumstances. Specifically if the B+ was derived from series secondaries (600VAC, non-CT), and if rectifier was correspondingly reconfigured as a full wave bridge with a resistive load. That should be clear if you read the Tung Sol data sheet linked above.

Thanks for clarifying, though irrelevant as single 6550 simply will not fit with the larger transformers I have gotten, that should be clear if you read the post(s) above

This was meant to clarify the post in response to the candidly rude commentary about the initial post's clarity above; it's abundantly clear the point is moot since you are unwilling to machine the chassis to fit the larger tube. I'm unsure what's motivating the hostile tone in the recent posts. If you are trying to go in a different direction, saying so would suffice.

[GlideOn]

The post was saying that a single 6550 would be as powerful as a pair of EL34s under the right circumstances. Specifically if the B+ was derived from series secondaries (600VAC, non-CT), and if rectifier was correspondingly reconfigured as a full wave bridge with a resistive load. That should be clear if you read the Tung Sol data sheet linked above.

Thanks for clarifying, though irrelevant as single 6550 simply will not fit with the larger transformers I have gotten, that should be clear if you read the post(s) above

This was meant to clarify the post in response to the candidly rude commentary about the initial post's clarity above; it's abundantly clear the point is moot since you are unwilling to machine the chassis to fit the larger tube. I'm unsure what's motivating the hostile tone in the recent posts. If you are trying to go in a different direction, saying so would suffice.

With respect, there was never any moment of intended malice, much less hostility as you put it, as if it were possible to convey via mere words and pixels to a complete stranger I've never met before.

I would hope that were are all mature adults here with at the very least light hearted senses of humor, seeking to learn first and foremost and not seeking to making make a scene or detract from the subject nor derail the thread to air personal grievances.

I apologize If you were made uncomfortable from anything I posted.

Let's move on.

[GlideOn]

Look here for calc's to make some quick estimates: http://www.valvewizard.co.uk/se.html

With 420V Va and 2x EL34 (50W Pa max) Za for center bias = 420^2/50 = 3528 ohms
Assuming 0V Va min, power output will be Irms*Vrms, Irms = 420/3528 * 0.707 = 0.0842A, Vrms = 420 * 0.707 = 297V, Pout = 25W
Screen voltage should be about half of anode voltage, which is nice if you run a FWB and a CT PT. Using a stacked reservoir you can run the CT to the middle of the stack and get Vg2 from there.

With 420V Va and 1x 6550 (42W Pa max) Za for center bias = 420^2/42 = 4200 ohms
Assuming 0V Va min, power output will be Irms*Vrms, Irms = 420/4200 * 0.707 = 0.0707A, Vrms = 420 * 0.707 = 297V, Pout = 21W
Again screen voltage should be about half of anode voltage.

The edge in power of EL34s make sense, but I have no idea how you arrived at 0.707 figure.

Is that current or a fraction of voltage/power?