In some readings, I found that a center tapped transformercan be used as a splitter for a push pull output stage. That being said, of course it would have to be a high quality transformer.
I have not seen any designs utilizing this technique. Any comments on this application?
Roger
I have not seen any designs utilizing this technique. Any comments on this application?
Roger
I use it regularly. I just finished a pair of 45 P-P amps with a Lundahl LL1660 as a phase splitter. ISO makes a nice transformer for that also, the NC-126.
Was a common technique before feedback was incorporated into tube amps. Then, the additional transformer just added too much phase at both low and high frequencies to make feedback practical. Then there is the added cost.
Not having global negative feedback, and often not having any negative feedback, makes it possible to design . . .
What I call: "Forward Facing" circuits.
Most often successful by using Triodes, Triode Wired Pentodes, and Triode Wired Beam Power tubes.
That is what my amplifier designs are like today.
But . . .
Most all types of circuit topologies can be successful, but only if the design takes all the problems and tradeoffs into account, and deals with them properly.
I am glad there are fans, designers, and builders of SE, PP, OTL, global negative feedback, Ultra Linear, Schade, ParaFeed, etc.
If you have never heard a good sounding example of all of these designs, I am sorry that you have not had that opportunity . . . Yet.
If I said I did not learn from and get/steal ideas from all of these topologies, my nose would be getting longer.
Signed: "Pinocchio"
What I call: "Forward Facing" circuits.
Most often successful by using Triodes, Triode Wired Pentodes, and Triode Wired Beam Power tubes.
That is what my amplifier designs are like today.
But . . .
Most all types of circuit topologies can be successful, but only if the design takes all the problems and tradeoffs into account, and deals with them properly.
I am glad there are fans, designers, and builders of SE, PP, OTL, global negative feedback, Ultra Linear, Schade, ParaFeed, etc.
If you have never heard a good sounding example of all of these designs, I am sorry that you have not had that opportunity . . . Yet.
If I said I did not learn from and get/steal ideas from all of these topologies, my nose would be getting longer.
Signed: "Pinocchio"
Last edited:
I have two amps with transformer phase splitters, my recent 6S4S PP amp with LL1660S ITs and an older junkbox build with triode wired PL81s as output tubes and Edcor XSM10k interstage transformers. The XSMs are actually used as autoformers as (as in the EL84 schematic above) after a mishap where the primary winding on one of them went open circuit.
rfbrw,
Thanks for posting that!
Except for the Ultra Linear Negative Feedback, that is as near to "Forward Facing" as you can get.
I designed and built a simple push pull amplifier with the ElectraPrint PSA-2N.
I am sure, like me, many people miss Jack Elliano's products, experience, and wit.
Sooner or later, we all have to retire.
Jack, Thanks for hosting me at your home, 2 times.
Thanks for posting that!
Except for the Ultra Linear Negative Feedback, that is as near to "Forward Facing" as you can get.
I designed and built a simple push pull amplifier with the ElectraPrint PSA-2N.
I am sure, like me, many people miss Jack Elliano's products, experience, and wit.
Sooner or later, we all have to retire.
Jack, Thanks for hosting me at your home, 2 times.
How about splitting phase before the amp? Use a phase splitting transformer on the output of your preamp. Or just use a DAC with differential output. Build the amp accepting differential signals and then not worry about it on the amp. Sure, you’ll need twice the drivers but twin triodes make that easy, just one per side.
I find it to be an elegant solution and skipping the phase splitter in the amp makes it a bit simpler circuit. I imagine it also simplifies feedback options but I don’t have enough first hand knowledge about that.
I find it to be an elegant solution and skipping the phase splitter in the amp makes it a bit simpler circuit. I imagine it also simplifies feedback options but I don’t have enough first hand knowledge about that.
I find a few mono amps from the 1960s using a phase splitter transformer. Cannot remember models.
isaacc7,
You almost "read my mind".
But there is no need for a phase splitting input transformer.
Let the CD player do the phase splitting.
I already posted my design of a low power balanced amplifier, in a thread in Tubes/Valves.
A CD player with XLR outputs feeds the 2 mono-blocks amplifiers with XLR inputs.
A dual triode ECC82 feeds a pair of 5881 Triode Wired, per mono-block.
The XLR input feeds 7 resistors:
2 for grid returns, 2 for grid stoppers, 2 series Rs, and a Rheostat across the series Rs to control the volume.
Link to my Balanced Mono-Blocks:
low-power-balanced-mono-block-tube-amplifier.404771
page-2
I hope the link works.
You almost "read my mind".
But there is no need for a phase splitting input transformer.
Let the CD player do the phase splitting.
I already posted my design of a low power balanced amplifier, in a thread in Tubes/Valves.
A CD player with XLR outputs feeds the 2 mono-blocks amplifiers with XLR inputs.
A dual triode ECC82 feeds a pair of 5881 Triode Wired, per mono-block.
The XLR input feeds 7 resistors:
2 for grid returns, 2 for grid stoppers, 2 series Rs, and a Rheostat across the series Rs to control the volume.
Link to my Balanced Mono-Blocks:
low-power-balanced-mono-block-tube-amplifier.404771
page-2
I hope the link works.
Last edited:
Here you can see a Tubelab SSE board, which is a two channel SE amp board for 6L6GC, EL34, 6550 and other tubes with a similar pinout. I used an Edcor UMM 8 transformer to split the phase on the input of the amp and simply wired a push pull OPT across the two plates with the CT tied to B+. Output tubes were "RUBY" 6L6GC's. This works quite well when driven from a low impedance source. When fed from a 600 ohm audio oscillator I got very low THD numbers like 0.724% at 1 watt, 1.035% at 10 watts and 1.99% at 35 watts where visible clipping was present.
My crummy old LG DVD player did not like the transformer load. I did a similar experiment with a small Parts Express "LEPAI LP-2020A+" solid state amp driving the transformer. I also used an old Hammond 1608 OPT instead of the guitar amp unit in the other experiments. I fed this from my PC through a Focusrite Solo and ran some tests. Clipping and OPT saturation were evident at low frequencies due to the 10 Watt OPT and power was lower due to the 8000 ohm load, Same Ruby 6L6GC tubes. Connecting up a speaker provided some really good sound, with minimal low frequency artifacts. My Yamaha NS-10M Studio monitors have very little output below 70 Hz despite large cone movements and they are in the 20+ ohm range below 100 Hz. I can't seem to find any of the old data though.
More recently I did the same thing with an UNSET board with similar results due to a request from the past. See post #563 here:
https://www.diyaudio.com/community/threads/unset-beta-board-build.376124/page-29
I tried the push pull hookup to see how an UNSET board did in push pull. It is likely that my ongoing guitar amp development project will use UNSET output topology due to its ability to transition smoothly from triode to pentode mode at the turn of a knob.
Attachments
An article on phase inversion:
https://dalmura.com.au/static/Phase Inversion Circuits Radio Craft 1948.pdf
I had used LL 1660S as phase splitter which doesn't have the inductance for bass response for my taste.
Inversion with center tapped coil (autoformer aka) with constant current source in the driver sounds
better than using LL 1660S.
A "cheap" center tapped coil is LL1670, you can find it here:
https://www.erhard-audio.com/Chokes.html
but make sure the choke is center balanced.
https://dalmura.com.au/static/Phase Inversion Circuits Radio Craft 1948.pdf
I had used LL 1660S as phase splitter which doesn't have the inductance for bass response for my taste.
Inversion with center tapped coil (autoformer aka) with constant current source in the driver sounds
better than using LL 1660S.
A "cheap" center tapped coil is LL1670, you can find it here:
https://www.erhard-audio.com/Chokes.html
but make sure the choke is center balanced.
Here is the thread - https://www.diyaudio.com/community/threads/my-new-el84-pp.115095/
and if you find the above a bit tame - https://www.diyaudio.com/community/threads/marantz-t1-well-sorta.380828/
In my head I see differential amps like mine as a 4 channel amp sharing A single power supply. The output transformers sum each side’s signals. Are there any potential advantages in thinking in terms like this when designing an amp or am I way off base?
Because from such an amplifier, you can expect a frequency response from about 100Hz to 10KHz and maybe 1% THD. Transformers were used that way maybe 80 years ago when such performance was considered hifi. Big transformers work much better than small transformers, so the output transformer is less of a problem. The self-resonance of a 10nF cap may be about 100MHz, well above audio frequencies but the self-resonance of a transformer will be close, if not within audio frequencies.
Don't forget the most obvious reason, cost. Transformers are often the most expensive parts in an amp or other electronic device. That's why they were designed out of radios, TV sets, and other high volume consumer electronics equipment.
Concertina splitters and differential pairs can make excellent splitters without the distortions that transformers have. A wide bandwidth transformer is expensive and transformers in general can cause headaches when attempting to enclose them in feedback loops.
See article "the Firal Eye" in Sound Practices. I use a Tango NC14, driven by a 300B (Elrog at low current) splitting into 2 x 813 (triode strapped). Works like a charm...