F6 Amplifier

Hey thanks buzz, I was actually going to email you first to avoid exposing myself lol. no the pertinent point in my post is, without the buffer, can the Xformer be fed directly with a balanced input? there seems to have been a few posts concerning how to connect a balanced input but nothing showing it, so the confusion has been over how to implement a balanced buffer stage in front of it? it seemed pretty straight forward to me to just run the one without the buffer, but the confusion amongst posters had me thinking I was missing something obvious and maybe I still am...

see I woulnt really need the buffer most likely, but I would prefer balanced input and the transformer in this case is most certainly the most elegant way to do that conversion, since its already there.

has anyone got a decent line transformer model for iCircuit? yes I know its a toy, but it would help me to understand the circuit to model it out and i'm on mac

yes running directly would certainly be something I would try, perhaps with an AK4399 DAC i'm kludging together, I liked the ESS with TX 'IV', it seemed to be very natural, however already a bit less live than the NTD1 and here it may take that further, which I wouldnt like, as its why I stopped with TX on the ESS to begin with.
 
Hello there,

I just read the article some more and stumbled across one thing:

If I understand correctly the circuit should be converted to a current source amp by removing the loop feedback. So my question is, would this be easy to do by simply skipping some parts or would it be more complicated and require redesigning the circuit?
 
Official Court Jester
Joined 2003
Paid Member
I obviously need to let this steep into my brain more, or what I sometimes like to call a brain anyway. but I cant really see why the buffer cant be deleted and a balanced signal fed directly to the transformer? doesnt the transformer not care? what does the following amp stage think of a negative phase? does it care?. I would have normally thought that the input 'ground' leg of the TX is just a fake negative phase; or rather it doesnt really have phase like that, it only really has difference between the positive and ground, same as it would see the difference between positive and negative. but i'm probably missing something really basic.......

quite possible , doable and done already

but , if you arrange xformer to work like that , you can't use F6 feedback approach

maybe if you wire input xformer as center tapped , connect center tap to same feedback node as in original , so balanced input became symmetric

gain will change also - you'll need 6db more for full throttle

but , checking that demands morning coffee ........ and I'm just going to organize that
 
Official Court Jester
Joined 2003
Paid Member
Hello there,

I just read the article some more and stumbled across one thing:

If I understand correctly the circuit should be converted to a current source amp by removing the loop feedback. So my question is, would this be easy to do by simply skipping some parts or would it be more complicated and require redesigning the circuit?

not so simple

mainly because of this
 

Attachments

  • open closed loop vs freq.jpg
    open closed loop vs freq.jpg
    78 KB · Views: 486
quite possible , doable and done already

but , if you arrange xformer to work like that , you can't use F6 feedback approach

maybe if you wire input xformer as center tapped , connect center tap to same feedback node as in original , so balanced input became symmetric

gain will change also - you'll need 6db more for full throttle

but , checking that demands morning coffee ........ and I'm just going to organize that

Having big speaks coming on line, i might very well try this. Dual channel ess dac with one output feeding xformer, directly into f6 output stage and other feeding xformer followed by opa, feeding hybrid f4. Iam excted now. Wghere do i ut j2 and m2:D
 
Having big speaks coming on line, i might very well try this. Dual channel ess dac with one output feeding xformer, directly into f6 output stage and other feeding xformer followed by opa, feeding hybrid f4. Iam excted now. Wghere do i ut j2 and m2:D

definitely worth a try and I probably will too, but despite having pretty nice sound directly feeding an ESS dac into a transformer is not exactly high on the performance list, I will probably end up feeding the D1 into it.

thanks for the answer ZM, looks like i'll have to do the traditional learning technique of Pass Article and learn by osmosis.

luckily its a simple enough circuit parts wise to build up and play while learning, better buy a couple spare R125 :redhot: the BudP transformers can be wired for 1:2 as well, so I could make up 3dB that way, but it does seem the trick in this amp is in the feedback, so best put my thinking cap on
 
Official Court Jester
Joined 2003
Paid Member
definitely worth a try and I probably will too, but despite having pretty nice sound directly feeding an ESS dac into a transformer is not exactly high on the performance list,

.....

question is - is that combo optimized

......

thanks for the answer ZM, looks like i'll have to do the traditional learning technique of Pass Article and learn by osmosis.

..

best explanation of right approach , I had opportunity to see here , in Papaland ;)

especially for Mighty Moi , not interested to learn just tech things from Paparticles :rofl:

....... but it does seem the trick in this amp is in the feedback, so best put my thinking cap on

simple as is , each tiny piece in this bloody amp is having important part in doing the trick
 
Official Court Jester
Joined 2003
Paid Member
Having big speaks coming on line, i might very well try this. Dual channel ess dac with one output feeding xformer, directly into f6 output stage and other feeding xformer followed by opa, feeding hybrid f4. Iam excted now. Wghere do i ut j2 and m2:D

two paths :
- ESS voltage out , feeding repeater ; output impedance of ESS is important

- ESS current out , feeding mic. xformer class xformer ; then SUSY chip ; will you have enough swing on output ?
 
yeah well I consider that folklore, high output impedance and current limiting make driving LCD2 and other headphones directly with es9023 an unfriendly concept.

it is possible to take minimalism too far.

power to you if you are getting good sound, really. I just think this gets posted around a bit too much and any cases it works properly will be rare, also lacking volume control

going balanced will make them work harder
 
Last edited:
@ ZM : optimized? hmm maybe not, just used passive IV into TX, but i'm pretty addicted to the NTD1, nothing ive tried since has really done it for me, I had planned on using something else for the tweeters, but i've just ended up buying parts for another D1, despite the fact I will end up with a 4 channel dac that dissipates about 100W in the IV stage and is as big as a poweramp....

i'm just not sure its possible to match, nevermind better the opc/pass NTD1 with transformers. Matching as it does, the THD and detail a high quality opamp IV, but with no global feedback and less than a handful of parts in the signal. it gave me that ease of sound but without paying the penalty in detail.

Besides i'm pretty keen on drawing a line under the dac section of my rig, being 4 x balanced mono with another 2 channels on the way, to replace it would not be a cheap or efficient exercise, so I shouldnt really even consider it; but it wont stop me trying it since i'll have the parts on hand. there are other things that need attention, i've spend too much focus and budget on the dacs as it is.

feed it direct and you have issues of loading the dac and added parasitics, non-linearity, feed it through passive IV first and you lose in performance. a jfet or mosfet follower/buffer/current conveyer could be placed in front of the TX to drive it and do IV in between, but then thats basically the D1 + F6

i'm also not prepared for urushi painted permalloy cores built by ninja artisans $$$$

I do like what you and AR2 did with his Ackodac and jfet buffered TX, bet it sounds great, but best I continue with my best foot forward in the direction I have started

but enough on this topic probably. I wish we had those spoiler 'click to show' objects some other forums have. It sounds like i'm sorted for TX, I have 18vac/300VA antek tx, I have the jfets, I jave most resistors; are there any other really critical parts you guys are discovering have a profound effect? everyone is using LEDs or zeners for references?
 
Last edited:
Official Court Jester
Joined 2003
Paid Member
...

it is possible to take minimalism too far

pretty straightforward - cartridge's needle right into the vein

or , if you want it permanent - you can always (black humor mode on) - just take Syd Barret's approach - " one day almost genius , another day without any needs"


regarding DACs - for my modest and limited experience - nothing beats differential I output DAC(s) directly feeding xformer primary , with appropriate load strictly on secondary , followed by either SUSY or Supersymmetric ( or mix of two) gain stage

using xformer in I/V just as amplification device , not I/V conversion itself , is .......... say it mildly ........ stubborn and not elegant

and I saw that more than once , unfortunately not just hyped out of proportion , tnx to power of web
 
Last edited:
not so simple

mainly because of this
Zen Mod: The curve at 39 dB is a bit surprising. It implies that the transformer is performing poorly despite its excellent specifications by Jensen. Its only purpose in the current source amp configuration is to faithfully transmit the input signal, and be unperturbed by its surroundings. You know transformers very well. What's going on?
 
Official Court Jester
Joined 2003
Paid Member
Zen Mod: .....What's going on?

I can't give you exact recipe , not having experience with exact Jensen type ;

whatever - speaking strictly of xformer - loading each secondary with nominal impedance (150R) will result in better bandwidth

be aware that Jensen guys made it exactly that way

don't think even for a sec that Pa doesn't know that - he just took different approach - to correct entire amplifier , killin' seven flies with one blow

1938 - Mickey Mouse, Minnie Mouse - The Brave Little Tailor - YouTube
 
Hello there,

I just read the article some more and stumbled across one thing:

If I understand correctly the circuit should be converted to a current source amp by removing the loop feedback. So my question is, would this be easy to do by simply skipping some parts or would it be more complicated and require redesigning the circuit?


OllBoll: Your questions and Zen Mod's reply post are relevant, and valuable. Please follow this discussion:
  • Withou loop feedback, F6 is a current source instead of a voltage source amp.
  • The open loop gain of F6 is 39 dB [shown in the reply post by Zen Mod]
  • The closed loop gain of F6 is 14 dB.
  • You and DIYers have up to 24dB of available gain to burn [a term used by Mr. Pass].
  • One can methodically apply overall negative feedback [burn gain] like in F6 and end up with a closed loop gain of 35 dB, 30 dB, 23 db etc..
  • As available gain is consumed, OLG F6 moves methodically from a current source to a voltage source amp. Thus, the parent output impedance decreases, its audio bandwidth increases, its harmonic distortion [Zo] decreases etc.
  • One can burn the right amount gain to stop at a Zo =8 Ohms which is the output impedance of the SIT amp by Mr. Rothacher. Or stop at a Zo = 6.5 Ohms to drive the highly acclaimed Arpeggio loudspeaker. Or stop at Zo = 4 Ohms which is like that of the SIT amps by Mr. Pass.
In order to fully appreciate the following paragraph, please go to the thread in the Pass Labs Forum which is entitled "O. H. Schade (1938) meets MOSFETS". Go to post # 29 and watch the video link provided by DIYer avincenty. There is a high probability that the intrinsic pentode characteristics of the output stage in OLG F6 get reorganized to triode-like characteristics as the closed loop gain decreases; just like in the video. This burnt gain in F6 [IMO] is a special Pass Feedback used in his FET circuits like that of Schade Feedback in vacuum tube circuits. Overall loop feedback to a summing juction emanating from the ouput of a current source amp must be electrically different from that emanating from the output of a voltagte source amp and terminating at the same summing junction.