|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
63.246.183.22
In Reply to: RE: Question about transformer phase splitter as used in "Seth Amp". posted by Palustris on January 31, 2016 at 08:44:15
"The EXO-173 PPS choke was engineered to be used in that manner."
I agree. That's a high inductance nickel choke, designed for this specific use. It's unlikely you'll get good frequency response with a substitute if you build the circuit as-is. OTOH, you could use a transformer like the Edcor XSM15K/15K the way it was intended to be used. Drive the primary through the capacitor, and take output from the secondary with the CT grounded. Unlike the EXO-173, the Edcor is designed to drive a specific load. You'll need to load the output with the correct operating impedance (resistor) to prevent frequency response anomalies. Bear in mind, I'm not saying this will produce performance that's identical to the EXO-173, but it does provide an easy means for passive phase splitting.
--------------------------
Buy Chinese. Bury freedom.
Follow Ups:
I'm not following. I guess where I am confused is how the inductance comes into play in the circuit. It's not being used to load the plate it looks like it is merely being used to slit the signal. I'm not arguing I'm just hoping someone can make me "see" what they are talking about. I have contacted Mike about buying a pair of the EXO-173s. I've never had much luck understanding the AC side of things in a tube circuit and I don't expect you to bend over backwards to explain it to me but if you can clear things up a bit I would appreciate it.
OK, I see where you're coming from. Here's my attempt to describe it...The original circuit uses a center-tapped choke. The two halves of the winding are closely coupled, perhaps bifilar-wound to assure symmetry. Because of the coupling, any current flowing through one half of the winding creates an equal but opposite current in the other half.
The reason the choke must be specifically designed for this use is that one winding is operating with a virtual open circuit (the grid of the lower 2A3). Even if voltage balance weren't an issue, most transformer windings are designed to operate with a specific operating impedance. Failure to maintain the proper load causes frequency response anomalies, such as high- or low-end rolloff. Operating a standard transformer at too high an impedance can also create undesired resonances that degrade frequency response within the passband.
Regarding the voltage balance I mentioned above, even in this case it will gradually degrade as frequency rises. Inter-winding capacitance will affect the open-circuited winding well before the winding that's attached to the lower operating Z of the driver tube. However, the same design techniques that allow the transformer to exhibit good high-end response into an open circuit should move this characteristic well out of the range of audible frequencies.
Hope this is clear. I'll be glad to go into more detail if needed.
--------------------------
Buy Chinese. Bury freedom.
Edits: 01/31/16 01/31/16 01/31/16
Because of the coupling, any current flowing through one half of the winding creates an equal but opposite current in the other half.
Correct. It behaves as a 1:1 inverting transformer and since it needs to be inverting, bifilar is not an option. There was a large discussion here a number of years back about this and it is still my belief that using the terms "Autoformer" or "CT choke" stem from how the circuit is often drawn and not how it actually operates. My description of the circuit has the primary of a 1:1 inverting transformer acting as a grid choke for the in phase signal to one of the output tubes and the inverted secondary providing signal for the other output.
dave
--------------------------
Buy Chinese. Bury freedom.
Your drawings are correct and the "in phase" bifilar will have very wide bandwidth. A bifilar works because even though there is a large capacitance between windings the voltage gradient is such that there is very little AC potential between adjacent wires. When you invert on of the windings you have have that capacitance to ground which makes a nice path for your high frequency information.
Here is the simplified circuit in spice.
the 5n of capacitance is arbitrary and going to a slightly more elaborate model shows the same behavior.
Now to be fair I removed the obvious offending 1N cap to ground and the behavior is still there.
The moral of the story is do not invert a bifilar unless you want a nice lowpass filter. In practice the inverted bifilar typically used as an IT kills everything above 10Khz but the simple nature of this filter can actually be used to your advantage at the output of a Dac where a 40khz brickwall filter spells sonic death.
dave
Well, I guess the impedance in this application is a deal breaker for out-of-phase bifilar. I've designed and measured such transformers out to frequencies approaching 1 GHz, but that was always much lower impedance. I do see your point about about strays and opposing voltages.
--------------------------
Buy Chinese. Bury freedom.
So even though it's not directly loading the plate of the input tube the inductance does play a role. Rather than try to load the primary by coupling the secondary ct to ground with a resistor I think I will just buy the right part. I plan on using the same topology with a e182cc and a couple of 12b4as in PP. I'm hoping to get at least 4 watts out of the circuit. Thoughts?
If you're budget allows buying Magnequest chokes and transformers, the hard part is over! :)
--------------------------
Buy Chinese. Bury freedom.
"I guess where I am confused is how the inductance comes into play in the circuit. It's not being used to load the plate it looks like it is merely being used to slit the signal."
It's not between the power supply and the plate but it is still part of the plate load.
Everything hanging off the plate is part of the plate load and if there's not enough inductance in that winding there will be a loss of low frequencies.
I like TK's idea of just using the transformer as a transformer, the way it was intended to be used.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: