|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
69.108.92.40
In Reply to: what are good sounding circuits for Electraprint choke-splitter? posted by freddyi on March 14, 2007 at 19:22:11:
I'm driving PP 300b's with a paralleled 6sn7 at 16ma. 250vdc plate with a 6.5uf coupling cap.You might want to look at what Dr. Jean-François Lessard has done with 2a3's. He's using the MagneQuest EXO-173 (same function). I had a small hand in the design on this one.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Follow Ups:
From the Seth article:"It acts as the phase splitter in a simple but so elegant way. Putting the choke center tap to ground will create a reflected signal inverted 180o on the bottom leg from the one that is fed on the top leg (where C2 is connected)."
Has anyone confirmed this with measurements? In my playing around with high impedance centre-tapped audio transformer primaries in this application (no DC) the unterminated end rings like the Gong Show at high frequencies. The driven end of course depends on source impedance. The upshot is p-p balance goes way out of whack at the top end. I don't yet know if this is due to iron quality (vintage Hammond and a Freed so far) or topology but strongly suspect the latter is the main contributor. I notice the Seth doesn't load the undriven end to compensate.
A few things to take into consideration.The circuit is essentially a grid choke on the grid of one phase that is also the primary of a 1:1 inverting transformer feeding the other phase. (i know i harp on this)
the grid choke half will give you a darn near perfect square wave and as many have measured the inverting side can give you some interesting behaviors. When you add the perfect signal to the imperfect signal you do not get any sum/difference oddities and the imperfect side is 'temperd' by the perfect side. I think this has a lot to do with the sound of this approach.
In comparison, a SE:pp transformer is going to have two distinctly imperfect square waves, and when you add those signals together the sum/difference waveform becomes much more of a crapshoot.
Moving back to the grid choke / 1:1 inverting situation, it is also interesting to note that if an attempt is made to tame any odd behavior by loading, the gain of each phase will decrease as the load is increased but they will not track each other exactly. This means that as you load the secondary to fix a mismatch at HF, you will create an overall gain mismatch between the phases. Of course once measured in circuit, the gain mismatch (if judged excessive) could be corrected for by spec-ing an inverting transformer with the appropriate step-up ratio for the load required to neaten up the high frequency behavior.
all of the above will net you the circuit that measures the best which by no means suggests it will be the circuit that sounds the best. I personally give a lot of respect to the people who have tried, listened to, tweaked and liked this circuit in spite of the 'measured flaws'.
The transformer used in the test was a vintage Hammond 322 (80,000CT:20,000SE) interstage, nasty iron at best. The hope was it at least made the grade as an inductor. The 20k side was left open and the 80k used as the splitter - one end driven by a 50 ohm function generator, the other floating. The driven side was perfect, the ringing on the float side was over +- 100% of the square wave! While it's true a following differential stage will temper this, at best it halves the error and there's nothing to be done if the ringing overdrives one side.
In a somewhat related scenario (using the centre-tapped floating 70 volt windings of a P-P jukebox amp for output stage cathode feedback) I found exactly as you describe. Loading is a tricky business. In that case decent HF balance (level and phase ) was achieved by loading one side only with a series RC of 50 ohms and 0.47uF. Haven't tried similar on the 322 yet. So I think it can be done, the question of course being, as mqracing says, whether the result sounds any good in a phase splitter. (In the cathode feedback example it was mighty handy in avoiding a 55kHz oscillator.)
when you take an 80K tapped winding and hook it up as you did, you essentially made a 20K:20K inverting transformer to drive the inverted grid. you may get better results using the 20K primary as your grid choke/primary and then the approriate end to CT of the secondary as your source for the inverted signal.
Hi Dave:you wrote;
:::the grid choke half will give you a darn near perfect square wave and as many have measured the inverting side can give you some interesting behaviors.:::
As tre' pointed out in this thread... there is a difference btwn a simple CT choke used as a phase splitter and some of the more evolved designs that several magnetics designers have brought to fruition. Our design (PPS version), for example, has a lot of interleaving btwn the driven and undriven half windings. And hence knocks down considerably the leakage L of the simple archtype ct choke. That and careful control of the c's yields a device with good phase and freq response well, well above 20Khz. And from one vantage point (the co-efficient of coupling) behaves moreso as a traditional two winding transformer.
And, interestingly, some of the moreso conventional two winding phase splitting interstages have themselves been noted in these forums to have some "interesting" square wave behaviours (to borrow your language). I don't think that the square wave issue is directly attributable to whether the device under consideration is a one winding or two winding device. Both types can be optimized for square wave behaviours if that is the goal.
MSL
Hey Mike, how about letting me post on your forum again?Thanks, Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Hey Mike, how about letting me post on your forum again?
Why not try contacting Mike privately rather than in a thread which has absolutely nothing to do with your question?
se
Steve, that's a great idea. Do you have an email address for Mike?Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Steve, that's a great idea. Do you have an email address for Mike?
I do, but since his AA profile is set to not accept unsolicited EMail and he doesn't have an EMail contact listed on his website, I would guess that he doesn't want it made public.
However his phone number is listed on the MangneQuest website .
se
Hi Steve:Very polite of you to not give out my email address if you were not sure that I have made it public. But in fact, I frequently publish my email address on AA forums when appropiate and anyone is welcome to email me at acrosound at aol dot com
And as you suggest anyone is welcome to call me as well. My phone number is listed on our company website. Generally evenings are best to get a hold of me by phone.
Actually I should take the few minutes and update my AA profile and have it listed there as well.
again thanks,
Mike
The Electra Print PSA-2D has 100K across the whole winding to control that. There are also winding considerations. I tried other windings that weren't made for it and they didn't sound very good. I don't think either the MQ or the EP are JUST center tapped grid chokes. Mike and Jack have "played" with the windings to make them balance just right.Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Muchas gratias! I'm much more comfortable giving it a spin now.
Thermionically addicted.
I used an E180CC (paralleled) driving UL 7027A'a with an unbypassed
Rk. I ran it off a large 450VDC supply, used OPT's a friend found on the roadside (literally) they where off Eico HF-22's? The whole project was
kind of a joke, it went together fast, I figured it might sound barely ok.
Well, for a geek who normally goes to way more trouble in every area
of amp building and design, I was floored the second I heard it. How can this sound this good? There's no DHT's, UL (oh my), degenerative FB, 1.0uf coupling cap, etc...I only had it around for a few days before I loaned it to friend, who won't give it back now. Powerful, open, extremely detailed, even low level info, good.
This topology is probably even better with 2A3's, but I haven't tried that yet, as I'm focused more on SE experiments at the moment, but this is a great "no circuit" circuit...!
Using a center tapped grid choke as a phase inverter does keep it simple. Simple is good? I like my amps.Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
This post is made possible by the generous support of people like you and our sponsors: