Self Split vs PP - There is a "fixed" power rating between the two options ?

[kagliostro]

I would like to know if there is a "fixed" rating between Self Split and PP options

I mean, using same B+ and Cathode Resistor is possible to know which will be the approximate output power in % respect the PP option ?

Thanks

Franco

[jjasilli]

See:  https://robrobinette.com/Deluxe_Micro.htm#Push-Pull

[kagliostro]

Thanks JJasilli

seems that the only useful (about this argument) info there is:

If you prefer a push-pull power amp this puts out a little more power than the single-ended parallel ....

But this is something of known as a real PP puts out more power than a self-split that uses the same tubes ....

Franco

[jjasilli]

I don't know this topic. The little I've read says real PP can use a bypass cap, but split PP cannot.  This contradicts Robinette.  But maybe he's wrong because it seems the cap would bleed the 2nd tube's signal to ground.


Anyway, the existence/absence of a bypass cap would account for a significant difference in gain.


Also, true  PP might be capable of different bias for more gain than split PP.

[frankenxtein]

Not that this is useful but , I built that one , not much power there PD 7.44 watts  with a self split 12BH7

http://el34world.com/Forum/index.php?topic=21486.msg228694#msg228694
 
Maybe some of the comments in that thread would give more information.

[sluckey]

The little I've read says real PP can use a bypass cap, but split PP cannot.

That's correct. At least that's the case when the grid on one tube is grounded. The cathode becomes the input for the grounded grid tube so you cannot use a cathode bypass cap.

[frankenxtein]

Just trying to learn something here this is all interesting. With a Phase Inverter? It's Push Pull? http://monster.partyhat.co/article/lunchbox-amp/

[sluckey]

The self split is also a push/pull amp. It just develops the push pull drive signals without using a dedicated phase inverter.

[PRR]

Figure "about 3/4".

AND you can use a P-P OT instead of a SE OT. (Smaller, cheaper.)

If you need a better answer than that, you probably should not be fooling with self-split.

[drgonzonm]

Franco,

As usual your inquiries offer challenges to the forum   
Awhile back, this forum was discussing the AX84 firefly project.  As you are likely aware, the firefly uses self-split power tubes.  I came upon an article, written
In Wireless Engineering, dated February 1946, that explains the operation of the self-split PP circuit  It was written by Amos and two patents were cited  492,407 (GE Ltd) and 508,644 (BBC).  I believe this are British Patent numbers.  Title: Push Pull Circuit Analysis. 

This article provides plenty of formulas, and graphs.   

I found within my notes,
1.  pentodes can be used.
2.  The self-split operates in Class A,
3.  the firefly output is about 1w.

With conventional PP, you might be able to operate the power tubes in Class AB, which provides more output.   

Now you have a source for better understanding the self-split. 

[kagliostro]

Many Thanks PRR

That approximate 3/4 is what I was looking for and obviously the interest is due to the reason you itemize

Many Thanks also to Drgonzonm

I'll try to find that article, for sure an interesting one

EDIT: I find the article as a pdf shared by Merlin on valvewizard

For those who can be interested, here the link:


http://www.valvewizard.co.uk/SelfSplit.pdf

Franco

[drgonzonm]

From what I could determine, there is no fixed ratio, but PRR's number is an excellent starting point.  And yes the cathode resistor is an important design consideration. With tweeking, the ratio can get closer to 1. 

The power ratio number you seek is:
P = 1/2(1 + i₂²/i₁²)
where the self-split power is proportional to i₁² + i₂²
and the conventional PP is proportional to 2i₁²


I enjoy the questions and issues you raise in your posts and threads.   I quoted an article found in Wireless Engineer by Amos dated February 1946, With the Title "Push-Pull Analysis, Subtitled "Cathode-Coupled Output Stage". 


From this article, I make the following recommendations to you.
1.  If you choose to use self split output;
    a.  Use pentode tubes, you will have a greater output.
    b. Familarize your self with the design aspects, Like any tube design, there are unlimited choices of components, and the inter-relationship between component choices.  My guess, would be to put together a spread sheet and model the impact of the choices.
    c.  Put together a library of resources.  This includes the article mentioned above, and copies of the two patents mentioned in the article. (#492,407 and #508,697)   I would recommend another ariticle: Compact Hi-FI power amplifier by melvin liebowitz/ Delaware Electronics Supply Co.  subtitled "Constructing a 6-watt amplifier in which the output tubes serve as their own phase inverter.  power tubes are EL84's.  This article is no where as detailed as the Amos article. 

Determination of the i's is circuit dependent, particularly the value of the cathode resistors.   
I left a lot out, and this post is long enough. 

[kagliostro]

Many Thanks Drgonzonm

Interesting info and ... I'll try to find also this documentation

Franco

[drgonzonm]

Many Thanks Drgonzonm

Interesting info and ... I'll try to find also this documentation

Franco

Send me an e-mail address via PM, and I will forward the two articles to you. 

Drgonzonm

[kagliostro]

OK, Many Thanks

Franco


p.s.: BTW all those who want to send me an email can use the email icon under my avatar present on each post

[Backwoods Joe]

What is the "low down" on the self-split style amp? Is this a way to get only a few watts from a pair of KT88"s. How about sonic qualities as compared to a LTP style amp? thx joe

[drgonzonm]

What is the "low down" on the self-split style amp? Is this a way to get only a few watts from a pair of KT88"s. How about sonic qualities as compared to a LTP style amp? thx joe

I suspect you can use KT88's.  you could even use them as triodes to cut the power more. 

I am not sure about the sonic qualities. After all, the tubes will be used in Class A. 

Start a new thread, so we can discuss this independent of this thread.  I will need some time to create the Connect circuit in JSchem, and also to create the Amos circuit in JSChem and comments.  Once posted, I am sure others on this forum will provide the guidance and modifications to create a guitar amp from a hi-fi amp.  The Compact article referenced wanting to create a Marshall like amp. 

[Ed_Chambley]

What is the "low down" on the self-split style amp? Is this a way to get only a few watts from a pair of KT88"s. How about sonic qualities as compared to a LTP style amp? thx joe

Backwoods Joe,
Magnetone made one of these.  Schematic attached and I have built a similar version.  The amp has a lot more harmonic content that a Push Pull.  I saw a you tube video of a guy analyzing the frequencies and the fundamental being the highest, of course, but surprisingly the first, second and third all were present in descending pattern.  Later I purchased a Philco radio from the 30's that had a self split output stage.


It is a fun amp and plenty loud.  You can also change the upper 470K resistor to a 2 watt pot and it will thicken the tone as your reduce.  Sounds great with low output single coils.  It sounds a lot like a really nice clean of a Marshall.  What I like to call chewy.  It is not DIRT pedal friendly without another input stage, but very nice with light modulation.


Sorry guys, couldn't help but post on this thread.

[PRR]

> Magnetone made one of these.

That is a different "self-split". Note that 6V6 cathode resistor is bypassed. One 6V6 screen has an extra resistor and cap coupled to the other 6V6. If the 6V6 works like most pentodes, this will derive an out of phase signal. I wonder how it works on "all 6V6 ever made", because 6V6 has low screen current and the residual varies wildly with different factories.

> yes the cathode resistor is an important design consideration. With tweeking, the ratio can get closer to 1.

If RK is infinity, it is perfect push-pull.

But if RK is even "very large", it wastes a large fraction of total B+ power. Since this scheme is about reducing cost (avoid a phase splitter), increasing power supply costs is "stupid". ("Stupid" may not apply to experimental guitar amps.)

You can make dynamic "RP" very large with an additional active device as a current source. Of course if you can afford another device, why not make it the little phase splitter instead of a high-current cathode part?

[robrob]

This contradicts Robinette.  But maybe he's wrong because it seems the cap would bleed the 2nd tube's signal to ground.

Yea, it looks like I was wrong on this one. I was comparing the output of the single-ended parallel power amp of the standard Deluxe Micro with the self-split version. I didn't realize how severe the performance hit was with the self-split compared to a phase inverter fed push-pull but I did factor in parallel power tubes' performance hit (two parallel power tubes put out less than 2x the power).

I edited the quoted line in the Deluxe Micro webpage to:

"If you prefer a push-pull power amp this puts out slightly less power than the single-ended parallel 12AU7. . ."

[jjasilli]

Thanks for clarifying that.


BTW I think your website is a great resource!

[MakerDP]

We just had a very good discussion on self-split over at AX84.com which culminated in me building one out of a repurposed Fender Super Champ XD chassis and iron set just a couple of weeks ago. The results were very good, IMO. I really like how it turned out.

but, in a nutshell, here is what we concluded:

They always operate in Class-A mode.
Self-split is a great way to get a more powerful "Single-Ended tone" without the expense of a huge SE transformer.
The in-general estimate for power output is about twice what a SE amp would do with just the one output tube with the same voltages.
There is more than one way to do a self-split configuration. There are screen-fed and cathode-fed designs. I have not built both yet to compare tonality. Cathode-fed designs, like the above mentioned Firefly project and the one I built, are mind-numbingly simple to wire up and require very few components.

What I learned from the project:
Dead Fender Super Champ XD's make AWESOME self-split platforms for 6V6 tubes if the reason the amp is dead is that the digital processing card went bad. If that is the case, the amp is un-repairable because Fender does not have any more of those cards. The resulting power transformer and output transformer are perfectly suited for a one preamp tube, two 6V6 power tube self-split design.
Self-split designs are super easy to wire point-to-point with terminal strips.
I discovered a new preamp tridoe/pentode tube that is really fun to play with (not really self-split related but it was a great discovery for me.)

Here is a link to the discussion if you are interested... http://ax84.com/bbs/dm.php?thread=499119

[jjasilli]

 

[PRR]

> They always operate in Class-A mode.

Class A is no cut-off.

If the input side ever cut-off, the other side would have no signal.

[drgonzonm]

What is the "low down" on the self-split style amp? Is this a way to get only a few watts from a pair of KT88"s. How about sonic qualities as compared to a LTP style amp? thx joe

You probably will get close to a LTP style amp. 
I am posting the Feb 1946 article about push-pull analysis.  I believe that PRR pointed out when the forum was discussing the firefly that the schematic was very close in design to the LTP.  Compare the schematic to the LTP on a 5f6a.  To me, its almost like bypassing the power tubes and feeding the LTP balanced signals in to an OT. 

Taking another guess, I would consider building a amp with an isolation transformer and powering the amp at 120 to 150v.   

Good luck