[Niteshade]

You do not see too many transformer based splitters out there- most are tube. Which kind do you like?

[nullspace]

I like autoformers:


Sometime soon, I'll be trying a 12AT7 long-tailed pair driving EL84 output.

John

[mgalusha]

I'm using Lundahl 1544A's for phase splitting in front of my McIntosh 275 and the improvement over running it single ended with the internal phase splitting is quite substantial IMO.

BTW, John that is a cool looking 2A3 PP, would love to hear it. Your design?

[poseidonsvoice]

I'm using Lundahl 1544A's for phase splitting in front of my McIntosh 275 and the improvement over running it single ended with the internal phase splitting is quite substantial IMO.

BTW, John that is a cool looking 2A3 PP, would love to hear it. Your design?

I'm with Mike here. Transformer and transformer only. With the advent of excellent interstage transformers from Electraprint, Magnequest,  Lundahl, Tribute and O-Netics (and I'm sure several others), the only aspect preventing a designer from using a transformer for phase splitting is cost.

Anand.

[nullspace]

Thanks Mike. If ever you're in Phila., let me know -- you (and any other AC'er for that matter) are more than welcome to stop by.

It's my design, more or less. The direct-coupled 6EW7 is basically Gary Kaufman's 6EM7 spud amp (see here) mingled with a Free Lunch/ Monkey CCS-version Bottlehead circuit ('Afterglow in SP #15), and both Magnequest and Electra-Print have autoformer-based phase splitter DIY schematics posted on their websites.

I think the interesting part is the parafeed PP output, which you don't see very often. I wired it so I could easily switch between series-feed and parafeed, and I like parafeed significantly more. I have a couple things I'd still like to try -- diode-bias for the large half of the 6EW7 (an 0D3) and fixed bias for the 2A3 stage.

John

[mgalusha]

I noticed the parafeed on the outputs and the plate chokes on the 2A3's. Nice to know it works well. 

Oh, and the thanks for the invite, it goes both ways. I've had a few AC'ers drop in if they are passing through town, all have been good folks.

[rklein]

Blair, although you are in the middle of building(or have finished) my NS-60 SE amp, would this option be available to me down the road and would it be beneficial to do in my case?  I have had the opportunity to hear Mike's Mac 275 with the Lundahls and it was pretty impressive.

Thanks,

Randy

[jkeny]

Has anybody ever tried John Swenson's Beam Deflection Tube - it's a splitter & volume control all in one tube! http://www.audioasylum.com/cgi/t.mpl?f=tubediy&m=141747
http://johnswenson1.home.comcast.net/~johnswenson1/stereo/BDTpreamp.gif[/img]]

[Niteshade]

I will have to give it a try! This sounds like a worth-while project.

Randy: your amp is using 6SN7's as phase splitters and that can not be changed. It works incredibly well, so you're not missing out. I can't change it over to a transformer coupled circuit.


Mike, what improvements have you noticed over the tube phase splitters?

[mgalusha]

Mike, what improvements have you noticed over the tube phase splitters?

Blair,

The Mc 275 uses a somewhat unconventional arrangement. The amp is fully balanced all the way through with balanced input and drivers.

They use 1/2 of a triode to invert the signal before the input stage and use that for the negative phase feeding into the input stage. The positive phase is fed directly into the input stage. Since the inverted signal goes through an RC network there is almost certainly some time delay, hard to avoid with RC.

Running it in balanced mode it removes the inverter circuit but of course requires a true differential signal, not just impedance balanced like many balanced designs. Using the transformer to generate the differential signal it seems to sound more coherent, perhaps because the internal method may introduce some slight offset in one phase of the signal. I have not tried to measure this but it rather transforms the sound of the amp.

mike

[GBB]

Mike,
The MC275 schematic shown here:
http://www.pmillett.com/file_downloads/McIntosh/MC275_sch.pdf
sounds a bit different from what you describe. I don't see the direct input for the positive phase that you mention. Is your MC275 different? 
For the schematic that I linked, I don't see a good way of using an input transformer as a phase splitter without messing with the gain structure and feedback loop. 
I'd be interested in hearing more details.
Thanks in advance.
---Gary

[mgalusha]

Gary,

Mine is one of the newer reissues and it's quite a bit different than the classic version shown in the schematic on Pete's webiste. I had to buy the service manual to get the schematic so I don't figure it's right to scan and post it though I don't see any sort of copyright. It has dual feedback loops, one for each phase. The output transformers have a dedicated feedback winding, so it's not tied to the speaker outputs, which are single ended as the negative is tied to ground.

The block diagram in the manual shows the unbalanced inputs going through a level control and then into the "unity inverting amp" which is just a grounded cathode stage to invert the phase.

After that it's input amp > driver amp > cathode follower > output amp, 2 of each stage per channel. The feedback loops connect back to the input amps.

Using balanced mode it connects the negative input phase to the input stage directly. I use the transformers externally. I have them in a little box with really short cables and XLR connectors. No internal mods to the amp. BTW, if one tries to use balanced mode with an impedance balanced vs differential source it sounds like crap since one phase is never driven at all. That little experiment is what prompted the purchase of the service manual as I wanted to know why. As soon as I looked at the schematic it was of course obvious and I built the transformer box within a day or two as I already has the Lundahl's.

[GBB]

It has dual feedback loops, one for each phase. The output transformers have a dedicated feedback winding, so it's not tied to the speaker outputs, which are single ended as the negative is tied to ground.

The block diagram in the manual shows the unbalanced inputs going through a level control and then into the "unity inverting amp" which is just a grounded cathode stage to invert the phase.

After that it's input amp > driver amp > cathode follower > output amp, 2 of each stage per channel. The feedback loops connect back to the input amps.

OK, I think I've got it.  The original MC275, according to the schematic above, has the following configuration:

Input gain stage > phase splitter > driver amp > cathode follower > output amp
So compared to that in the "new" amp, the 1st gain stage is changed to a simple inverter, which acts as a phase splitter if needed.  And the phase splitter stage then becomes the input gain stage.  After that it's close to the same, with similar driver amp > cathode follower > output amp.  This is consistent with what's published on the web about changing the 12au7 in the original amp to a 12ax7, and changing the 12bh7 to a 12az7 to get the right gain structure.  Since the feedback is gotten from a separate winding, one can take + and - phases and you're there.

Thanks for sharing the concepts.
---Gary

[Ericus Rex]

Isn't the tube phase splitter more commonly called the driver?  Or is the driver something else entirely?

[Steve]

Isn't the tube phase splitter more commonly called the driver?  Or is the driver something else entirely?

Not necessarily, but as you stated it is commonly used to drive the output tubes. If the phase splitter active circuitry is designed with gain, the phase splitter can eliminate stages for better fidelity. Saving a stage is almost always better.

Although difficult to design perfection, electronic phase splitters do offer the possibility of perfection with proper/flat tonally balancing, possible wider bandwidth, virtually no resonance problems.

Cheers.