So you want SERIOUS do you?
But Joel, since they have a common cathode resistor, the circuit will only work if the tubes are quite close.
Imagine 2 curves differing slightly:
To make them fit each other, there are 3 steps:
1) Match the bottom - done with bias
2) Match the top - done with gain
3) Match the curvature - this is the tube's inherent distortion, and should be match on 2 tubes of the same type and make, providing boints 1 and 2 are met.
Your circuit matches only point out of 3.
If you want something that "works", then it's probably OK.
I have a serious distrust of common bias. Even if the devices match at first, they might not later on.
If, on the otherhand you want to strive for excllence, you must consider the consequences of each design action.
But Joel, since they have a common cathode resistor, the circuit will only work if the tubes are quite close.
Imagine 2 curves differing slightly:
To make them fit each other, there are 3 steps:
1) Match the bottom - done with bias
2) Match the top - done with gain
3) Match the curvature - this is the tube's inherent distortion, and should be match on 2 tubes of the same type and make, providing boints 1 and 2 are met.
Your circuit matches only point out of 3.
If you want something that "works", then it's probably OK.
I have a serious distrust of common bias. Even if the devices match at first, they might not later on.
If, on the otherhand you want to strive for excllence, you must consider the consequences of each design action.
Well, yes and no, Joel. In a conventional paraphase splitter that you've shown, the balance will change as the tubes age, since it's dependent on the mus of each section. You've got an adjustment pot in there, which certainly will help when the tubes aren't well matched for mu and gm, but remember that those parameters change over a tube's useful life. This can be minimized by the use of a balancing resistor (the so-called "self-balancing paraphase"), but then you have a splitter the balance of which can never be better than 1/mu. Or you can keep adjusting the pot throughout the lifetime of the tube, but not everyone wants to do that.
Also, one must keep in mind the other disadvantages of paraphase splitters, among which are the asymmetry of HF rolloffs and distortion between the two phases.
The paraphase is a useful and venerable circuit, but as with all engineering compromises, one must realize that it still has its disadvantages. It is not a perfect splitter.
Also, one must keep in mind the other disadvantages of paraphase splitters, among which are the asymmetry of HF rolloffs and distortion between the two phases.
The paraphase is a useful and venerable circuit, but as with all engineering compromises, one must realize that it still has its disadvantages. It is not a perfect splitter.
Looking for perfection...or something darn close
SO gentlemen, in your opinion, what is the perfect phase splitter?
Off topic: Also, would it be possible to put the power supply and regulator components in their own chassis and run the clean power to another power amp chassis? Also, If I wanted to build a 5 channel power amp, could I have one massive mains transformer and then run the HT line through each channel?
David
SO gentlemen, in your opinion, what is the perfect phase splitter?
Off topic: Also, would it be possible to put the power supply and regulator components in their own chassis and run the clean power to another power amp chassis? Also, If I wanted to build a 5 channel power amp, could I have one massive mains transformer and then run the HT line through each channel?
David
HAIR SPLITTING.
Hi,
Nothing is ever perfect in this world...
Much will depend on the design targets you set.
What preceeds the spliiter,what follows and so on.
If you want a general overview of splitters the Bonavolta site has some interesting overview,a resume from a GA issue if I recall it correctly:
Phase splitters.
Yes...but you should provide for an individual HT reg per channel ideally.
Cheers,
Hi,
Nothing is ever perfect in this world...
Much will depend on the design targets you set.
What preceeds the spliiter,what follows and so on.
If you want a general overview of splitters the Bonavolta site has some interesting overview,a resume from a GA issue if I recall it correctly:
Phase splitters.
Yes...but you should provide for an individual HT reg per channel ideally.
Cheers,
Ditto what Frank said- if there were a "perfect" phase splitter, there would only be one phase splitter.
If I had a way of posting schematics, I'd give you one that I think would work for you in this application (it's one I've used in similar apps, using FETs), but lacking that, the GA reference and a look at some of the old stuff by Norman Crowhurst will give you a pretty fair overview. Consider diff amps, diff cascodes (the Hedge circuit), paraphase, and cross-coupled circuits, at a minimum.
Anyone who wants to give me a free schematic drawing package is welcome to do so!
If I had a way of posting schematics, I'd give you one that I think would work for you in this application (it's one I've used in similar apps, using FETs), but lacking that, the GA reference and a look at some of the old stuff by Norman Crowhurst will give you a pretty fair overview. Consider diff amps, diff cascodes (the Hedge circuit), paraphase, and cross-coupled circuits, at a minimum.
Anyone who wants to give me a free schematic drawing package is welcome to do so!
you mean something like these?
The Student edition is free
http://www.microcode.com/index.html
SwitcherCAD
http://www.linear-tech.com/software/
Cheers
Craig Ryder
The size of the coupling caps is irrelevant to the HF rolloffs, unless there's something going on there that I don't understand.
By making the coupling caps larger, you can move the LF rolloffs to a lower frequency, but you can't get away from there being one LF pole in one phase and two LF poles in the other. And making the caps larger carries other costs, too, by passing more infrasonic junk (thermal modulations, DC drift, 1/f noise...) and making decoupling much more critical.
By making the coupling caps larger, you can move the LF rolloffs to a lower frequency, but you can't get away from there being one LF pole in one phase and two LF poles in the other. And making the caps larger carries other costs, too, by passing more infrasonic junk (thermal modulations, DC drift, 1/f noise...) and making decoupling much more critical.
SPLITTER
Hi,
Well,once you start out with well matched valve halves and both halves work under the same conditions they do tend to age in the same way.
I would like to think that's an advantage.
As for xformers...I suppose a lot will depend on the quality of that as well and as usual:
Horses for courses,right?
Cheers,
Hi,
Well,once you start out with well matched valve halves and both halves work under the same conditions they do tend to age in the same way.
I would like to think that's an advantage.
As for xformers...I suppose a lot will depend on the quality of that as well and as usual:
Horses for courses,right?
Cheers,
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.