[ChrioN]

http://headwize.com/projects/showfil...szeke1_prj.htm

Anyone built it? Better than cmoy or grado?

 

[MisterX]

It's pretty nice but make sure you understand that it is a buffer more then an amplifier.
(it has almost unity voltage gain)

 

[robzy]

Quote:

Originally Posted by MisterX (it has almost unity voltage gain)

Just under to be exact...

Im also curious about the szekeres, im considering using it for my first p2p project...

Rob.

 

[ChrioN]

Quote:

Originally Posted by MisterX It's pretty nice but make sure you understand that it is a buffer more then an amplifier. (it has almost unity voltage gain)

Ah ok! How much current?

 

[rickcr42]

Typical Single ended Pure Class-A sound which is to the warm side of nuetral if you use typical Class-A/B Solid State as your reference "nuetral".I call this hard rather than nuetral but ...............

What you have with Class-A is zero crossing from the postive going signal because there is only one active element reproducing the entire signal.Class-a brings an "always on" condition so it responds to the input signal instantly rather than "turning on" in repsonse to that input signal and that is why they run hot.Sometime "on" means it only heats up in response to a signal,always "on" means it is running full out all the time.
what this gains you in sonic terms is the ultimate in low level detail retreival.The little bits that are lost when the A/B must go from "off" to "on" .

Quote:

Ah ok! How much current?

as much as you want within the limits of the power supply current reserves and set by adjusting the resitive element at the source of the mosfet (in this case R4 but can be any resistive element of the proper value).Use Ohms Law to set the current according to the power supply voltage.

Of all my amp I like my Szekeres amp best of them all for driving low impedance headphones with the current set to 250mA per channel.
Calling the Szekeres an amp is not exactly accurate though it does "amplify" or increase current drive.More accurate would be to call this a "High Current Voltage Follower" or "Unity Gain Buffer" or even an "Impedance Converter" there being no true amplification going on in the voltage domain.1 volt in is pretty much 1 volt out.
Having no voltage gain is meaningless if you use it after a CD player or on the output of your system preamp but if you want to use it with a low output source like a portable CD player you will need to add some kind of gain stage before the mosfet follower amp and that usually no more than X3

 

[dsavitsk]

Hey Rick, How would you compare (sound quality only, consider price to not be an object here) a well built Szekeres (meaning high quality parts, high quality PS, etc) after a preamp to a pair of good transformers (Sowter, Electra-Print, etc.) for driving low impedence phones?

Also, I seem to recall you expressing a preference for using a resistor rather than a CCS current sink in the Szekeres. Are the curves more linear, or is there some other reason?

-d

 

[rickcr42]

Quote:

Hey Rick, How would you compare (sound quality only, consider price to not be an object here) a well built Szekeres (meaning high quality parts, high quality PS, etc) after a preamp to a pair of good transformers (Sowter, Electra-Print, etc.) for driving low impedence phones?

Apples and oranges though more similiar than dissimiliar tonally.The thing with the Szekeres built straight and honest is there is no way around the big DC blocking cap so this has a definate sonic imprint on the sound.This needs to be on the order of 1000uF for a F3 of 5hz which gives you around 0.5 dB down in the bass @ 20 hz.I have had good luck using the elna Silmic here unbypassed.I do use and input transfomer (1:2 Step Up for +6dB gain)as the gain stage which in turn goes to a 10K pot,jfet gate buffer then szekeres amp so there are two active and three passive devices that can impact on the sound :

Passive :
1-input trafos
2-volume pot
3-output coupling cap

Active :
1-unity gain class-A JFET buffer
2-unity gain High Current MOSFET Follower

Another critical part of this design (and actully ALL single ended class-a amps) is the power supply.This stage has close to zero actual PSRR and the power source is more audible than any fancy designer parts.Mine use two stages of regualtion follwing a CLC DC supply so is :

1-AC to DC CLC coupled to
2-LM350K Pre-Regulator to
3-Individual L/R (dual mono) final regulation

Quote:

Also, I seem to recall you expressing a preference for using a resistor rather than a CCS current sink in the Szekeres. Are the curves more linear, or is there some other reason?

Purely sound.It specs better and has prettier traces with a CCS but sounds better with a passive current source.A prime example of why designing an audio device purely from spec sheets a road to true limitations instead of the perceived limitations of what atually sounds good.

Quote:

Avery Fisher "if it measure good but sounds bad you are probably measuring the wrong thing"

I tried every manner of CCS and some damn strange passive options and the simple high quality resistor the best option all things considered.Efficiency sucks but so ? This is DIY and what would not be acceptasble if a commercial amp is fine as a "one of" project.To say this puppy generates heat would be understating the obvious.

(the weird included photocells,thermistors,coils of heavy guage wire,inductors,triode heaters)

 

[Budgie]

I too love my version of the Skeres for low impedance phones. (Sony CD3000's)
I built mine with a CCS and bipolar power supply and was able to eliminate the output cap after carefully adjusting to minimize offset in the output. You still will need an input cap, unless you use a stage in front of the Skeres.
The sound is lively and dynamic, but without being cold and analytical. A good combination of tube like warmth and solid state impact. A good power supply is a must with a class A amp. I used a commercial bipolar regulated supply that I had in my parts box. "Power One"brand if I remember correctly.
It worked well for me.

I know Rick prefers the resistor bias. I would suggest trying both a resistor then a CCS and see which you like. While I trust Ricks opinions, tastes do very. I like the CCS better. Maybe if our amps were identical, I would prefer the resistor too. Thats part of the fun of this hobby!

 

[rickcr42]

Quote:

Maybe if our amps were identical, I would prefer the resistor too. Thats part of the fun of this hobby!

EXACTLY !

That is why we have these discussions and why I also like to toss out how important system matching is.
With a "loose" sounding system or one with flabby undefined bass then the apparant lower octave tightenting a CCS provides maybe would be a better choice but in my system and with my personal taste the resistor seems to add a more of a natural flow to the music rather than an analytical Hi-Fi sound that in my system a CCS adds to this design

 

[Budgie]

You insulting my phones, man?











(heehee)

 

[rickcr42]

Quote:

You insulting my phones, man?

nope just your taste

(j/k

)

Terms like "fat bass" or "lean bass" are really too strong since the actual sound differences are tiny at best.I just prefer the purely resistive over a long time of listening as having the most natural balance overall to my ears and even that could possibly be different if my R/C choices were different.Called tuning by ear then close 'er up (though the second part the hard one and why i made it modular

).
One thing that is amazing is just how good a single mosfet in a simple circuit can sound when matched up with the right cans.

 

[ECM]

I have built three versions of this amp. One with a dual power supply and resistor CCS. Another with a single power supply and active CCS. And yet another with a single power supply and resistor CCS.

My favorite of all three? Hands down it's the amp with a single power supply and resistor CCS. I use a 680uF output cap.

The active CCS sounds too sterile and "mechanical," if that makes sense.
The dual supply version gave me trouble with DC offset. It was drifting too much.

The single supply with a resistor CCS sounds just right. I agree with Rick, the music just flows and sounds more natural.

Hey Rick, I thought you were using an inductor for the CCS in your amp.

I was thinking of using four channels to build a balanced amp. I already have all the parts, but I'm weary of cutting the cables on my Grado 125's. Does this idea sound feasible?

 

[Teerawit]

Quote:

Originally Posted by ECM I was thinking of using four channels to build a balanced amp. I already have all the parts, but I'm weary of cutting the cables on my Grado 125's. Does this idea sound feasible?

Why don't you recable it instead of reterminating it?

 

[rickcr42]

Quote:

Hey Rick, I thought you were using an inductor for the CCS in your amp.

I did for a while in an attempt to get more efficiency at this point but went back as I always do to the simple "Big Mo" resistor.Just gets the sonics right.
I put my amp a while back in modular sections with nothing actually on bard unless it was a sub-module perf bord so I could over time experiment with this "simple" amp
The Mosfets are bolted drirectly to a rear panel mounted 4x5x1.5 heat sink with the baising resistors and RF bypass capacitor with everything else open for experimentation by using the barrier strips that connect to the G/D/S of the Mosfet

Things tried

1-single and dual power supplies fro 5VDC to 40VDC
2-AC and DC coupling the output
3-every manner of active and passive Current source
and set as high as 750mA per CH
4-every type of input from direct connection to the gate to a volume pot to triode stage to another mosfet stage to jfets.

My next (hopefully final) move is to lose both the volume control and jfet gate isolation and go to straight "Passive" 1:2 Stepup line transformer/"Active" Mosfet Unity Follower/Output "DC Blocking" cap then maybe button this sucker up permanantly and "retouch" the chipped paint on the front panel from all these damn experiments then do the Walnut Cabinet and sit back and listen