Looking at the Schematic, the output stage appears to be a classic White Cathode follower.
I looked it up in "Valve Amplifiers" Third Addition by Morgan Jones. Page 113,114.
Basically, the output stage has no gain. It also requires that the input signals are inverted and equal phase applied to the grids.
Like the M2, Voltage gain is from the transformer.
The real trick is the bias. Can't wait to see how Nelson did it.
Like the Futterman amp, they require (some)feedback to lower the output Z.
I love it.
Thanks Nelson.
I looked it up in "Valve Amplifiers" Third Addition by Morgan Jones. Page 113,114.
Basically, the output stage has no gain. It also requires that the input signals are inverted and equal phase applied to the grids.
Like the M2, Voltage gain is from the transformer.
The real trick is the bias. Can't wait to see how Nelson did it.
Like the Futterman amp, they require (some)feedback to lower the output Z.
I love it.
Thanks Nelson.
Thanks flg! It is a pleasure to follow the discussion, unfortunately beyond my horizon.....
Not necessarily. It is somewhat beyond our imagination also. We'll probably figure it out. Maybe after a few beers? I would think if the shematic has errors, N.P. would have realized it by now. That basically means, there is something non-semetrical here in order for it to be whatever he said, like a circlotron?
One of the charms about the transformer vs JLH/PLH phase splitter - go to
http://imageshack.us/photo/my-images/153/plhp.jpg/
is that it delivers pure symmetry in the same way a circlotron does - each like
polarity part is subjected to the same conditions, although in opposite
phase.
It is possible to do this with a phase splitter ala JLH/PLH, but if you look
carefully you will see that the bottom output device is run Common-Source
mode and the top device is Common-Drain. It is a rare circuit that will
deliver real symmetry under these conditions.
At first glance I would agree with that, but after read this papers from the follow URLs I am a bit confused and I'm not sure anymore.
The Tube CAD Journal: symmetrical solid-state output stages
Tube CAD Journal: No Gain Output Stages
Tube CAD Jounal: No Gain Output Stages
Tube CAD Jounal: OTL Gain Output Stages
Tube CAD Jounal: OTL Gain Output Stages
Last edited:
lalala
possible variations
all with different amount of Zen
some having Mu , some uM , some MuuM
edit : added original Pa's (with title ) sch , redrawn slightly , for easier comprehension
I understand how the first (leftmost) circuit would work, and was assuming that Nelson's F6 schematic (rightmost) had an error in the position of the phase dot in the bottom secondary. Did Nelson REALLY intend that the phases be the same into the output MOSFET gates? Not according to this post http://www.diyaudio.com/forums/pass-labs/216616-f6-amplifier-9.html#post3102389. I suspect an error in the original schematic.
Yes, I agree. A few posts later I responded with http://www.diyaudio.com/forums/pass-labs/216616-f6-amplifier-10.html#post3102675 but I don't believe I got a reply.
lalala
possible variations
all with different amount of Zen
some having Mu , some uM , some MuuM
edit : added original Pa's (with title ) sch , redrawn slightly , for easier comprehension
I understand how the first (leftmost) circuit would work, and was assuming that Nelson's F6 schematic (rightmost) had an error in the position of the phase dot in the bottom secondary. Did Nelson REALLY intend that the phases be the same into the output MOSFET gates? Not according to this post http://www.diyaudio.com/forums/pass-labs/216616-f6-amplifier-9.html#post3102389. I suspect an error in the original schematic.
Could it be that he is more clever than us?
No we are more clever. He's doin' all the hard work for us.
If F6 schematic is drawn correct than :
Upper bias source ( V ) stand there floating from the common ground point and is bootstrapped ( positive feedbacked ) from ``hot `` speaker output power line , but lower bias source (V) and associated power device is Not( bootstrapped ).
Result is that the both same sex output power devices work `` in phase `` , but with different inputs drive voltage amplitudes , so final Amp results will be .......?
At the first glance F6 circuit appear to be strange & unusual anyway !
Best Regards for all !!!
Upper bias source ( V ) stand there floating from the common ground point and is bootstrapped ( positive feedbacked ) from ``hot `` speaker output power line , but lower bias source (V) and associated power device is Not( bootstrapped ).
Result is that the both same sex output power devices work `` in phase `` , but with different inputs drive voltage amplitudes , so final Amp results will be .......?
At the first glance F6 circuit appear to be strange & unusual anyway !
Best Regards for all !!!
Your comment is great. Who's having the last laugh!No we are more clever. He's doin' all the hard work for us.
The circuit is just are Nelson said in http://www.diyaudio.com/forums/pass-labs/216616-f6-amplifier-9.html#post3102389. The output current of each MOSFET is determined from its gate-to-source voltage which is determined from it respective transformer winding plus bias voltage. Total symmetry, with opposite phases. The hard part is figuring out how the bias voltages are generated.
The feedback looks to be straight-forward, but I am not used to designing with transformers. Assume the feedback divider consists of resistors R1 and R2, when R1 is from the bottom of the transformer primary to ground, and R2 is to the Output. It is easy to see that the DC closed loop gain will be (R1+R2)/R1. I believe that the AC closed loop gain will be approximately the same, depending somewhat on the open loop gain, which is determined by the transformer winding ratios, and load resistance, and the MOSFET transconductances at the idle current.
DC idle state analysis determines the output voltage
The feedback looks to be straight-forward, but I am not used to designing with transformers. Assume the feedback divider consists of resistors R1 and R2, when R1 is from the bottom of the transformer primary to ground, and R2 is to the Output. It is easy to see that the DC closed loop gain will be (R1+R2)/R1. I believe that the AC closed loop gain will be approximately the same, depending somewhat on the open loop gain, which is determined by the transformer winding ratios, and load resistance, and the MOSFET transconductances at the idle current.
DC idle state analysis determines the output voltage
Remember that the device which performs the split operates in both
Common Drain and Common Source modes (Fet example).
As I see it, John's examples, while they work fine, are not "pure"
symmetry since the two driven devices do not see the same source
impedance value, their output impedances are different, the top one is a
follower (Common Drain) and the bottom one is Common Source (both
voltage and current gain).
Amusingly, the top Common Drain device is driven by the phase splitter's
Common Source operation and the bottom Common Source device is
driven by the splitter's Common Drain operation.
As I indicated, if you juggle the values around for the express purpose of
creating symmetry from a single-device phase splitter, you can get quite
close.
Batteries?
The DC closed loop gain is very low. There is no DC drive connection
between the input and output except for the feedback resistor, and sourcing
current to the load through that isn't going to create much excitement.
ahhhh... i feel happy i had a small insight .... the top half is not what i thought (ie source follower or CD) and the reason of course is that there is NO source degeneration going on.
Keeping in mind the mantra I inferred from reading pass articles 10x ... the FET knows nothing about the world apart from what it sees between the gate and the source and will attempt to react accordingly...
In a classic source follower, the load is connected between the source and ground. BUT crucially, the signal is connected between gate and GROUND. That makes the current flowing through the load push UP the source voltage and this reduce the VGS accordingly and this denegerates VGs (creates 100% NFB) and makes the voltage gain drop to 1x.
In the top half of NP circuit, increases of current also cause the source to float higher, BUT the gate floats higher also, so VGs is unaffected!!!!! The fet doesn't know anything else apart from VGS and so it blissfully just conducts away based on what it sees between gate and source.
So we get voltage gain in top half = voltage gain in bottom half !!!!!
.
.
.
So this is REAL symmetry ..... wow ... wow ... wow .... do you guys agree ?
now to go back to ZM comments re the phasing and my silly smileys to think some more ...
Keeping in mind the mantra I inferred from reading pass articles 10x ... the FET knows nothing about the world apart from what it sees between the gate and the source and will attempt to react accordingly...
In a classic source follower, the load is connected between the source and ground. BUT crucially, the signal is connected between gate and GROUND. That makes the current flowing through the load push UP the source voltage and this reduce the VGS accordingly and this denegerates VGs (creates 100% NFB) and makes the voltage gain drop to 1x.
In the top half of NP circuit, increases of current also cause the source to float higher, BUT the gate floats higher also, so VGs is unaffected!!!!! The fet doesn't know anything else apart from VGS and so it blissfully just conducts away based on what it sees between gate and source.
So we get voltage gain in top half = voltage gain in bottom half !!!!!
.
.
.
So this is REAL symmetry ..... wow ... wow ... wow .... do you guys agree ?
now to go back to ZM comments re the phasing and my silly smileys to think some more ...
Hang on. Just because there is not a resistor does not mean there is no resistance. There will still be some level of resistance on the pcb tracks, and the resistance of solder joint. That could possibly add up to 0.05 Ohms. As a guess. Anyway. I think I will just wait for Nelson to put me out of my misery (like a critically wounded dog)