Sziklai pair - 1fet+1bjt.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Dear Sirs,

I am very intrigued about the topology called SZIKLAI PAIR for using it in a future line preamp.
I have found schematics of pairs made out of two bjts.
I wonder if it is possible to build a Sziklai pair out of a fet and a bjt.
I understand that with a fet at the input I could do without an input coupling cap.
I would maintain the output coupling cap to be safer (my current power amp does not have an input DC blocking cap).
I am pretty sure this very simple topology has the possibility to sound extremely nice.

Suggestion would be very much valuable, welcome and aprreciated here.

Kind regards,

beppe
 
At the cost to turn out monotonous I am attaching an example of
sziklai pair.
I would like to replace the input bjt with a nice fet and so avoiding the input cap.

Thank you for your patience.
Kind regards,

beppe
 

Attachments

  • line_preamp.jpg
    line_preamp.jpg
    56.8 KB · Views: 1,437
you can replace input transistor with a JFET
but to remove input cap is not a good idea
not in this circuit

it is theoretically possible to use a JFET input with only a resistor connected to GND
because Jfet SOURCE (emitter) is at higher potential than GATE (base)

but to make this
you would need to build a different circuit
different resistor values and voltages

It is always good to have input cap ( high quality FILM Capacitor = MKT or MKP )
at a preamp input

Not all sources have Cap at output.
If you are sure you will only use source(s) with output Cap
you should remove input Cap.

But if you use a potentiometer before preamp, you have to use cap
for this single ended supply circuit.
Even if your CD-player have output Cap.
 
lineup said:
you can replace input transistor with a JFET
but to remove input cap is not a good idea
not in this circuit
it is theoretically possible to use a JFET input with only a resistor connected to GND because Jfet SOURCE (emitter) is at higher potential than GATE (base)
but to make this you would need to build a different circuit
different resistor values and voltages
It is always good to have input cap ( high quality FILM Capacitor = MKT or MKP ) at a preamp input
Not all sources have Cap at output.
If you are sure you will only use source(s) with output Cap
you should remove input Cap.
But if you use a potentiometer before preamp, you have to use cap for this single ended supply circuit.
Even if your CD-player have output Cap.

Dear Sir,

you have answered kindly and completely to my question.
I abandon the idea.
Thank you so much.

Kind regards,

beppe
 
Hi beppe,

However you can read this:
http://www.diyaudio.com/forums/showthread.php?postid=423652#post423652

I worked on it (CFP) a few years ago, not yet finished, I got sometimes very, very good results, I say sometimes because this circuit is very prone to oscillations and I think that from time to time it oscillates, and then the sound becomes harsh.
Today, I do not really know if the circuit is worthwhile vs fet only, taking in account instabilities problems, maybe the circuit needs to be "best tuned", I have not much time for it but one of this days... I can send you the latest design, more detailed, in a few days if you are interested.
 
Justcallmedad said:
Hi beppe,
However you can read this:
http://www.diyaudio.com/forums/showthread.php?postid=423652#post423652
I worked on it (CFP) a few years ago, not yet finished, I got sometimes very, very good results,
I say sometimes because this circuit is very prone to oscillations and I think that from time to time it oscillates, and then the sound becomes harsh.
Today, I do not really know if the circuit is worthwhile vs fet only, taking in account instabilities problems, maybe the circuit needs to be "best tuned",
I have not much time for it but one of this days...
I can send you the latest design, more detailed, in a few days if you are interested.

Dear Sir,
thank you so much for your extremely kind and valuable reply.
As I said in a very old thread the attached schematic is part of an old line preamp I had the opportunity to listen and appreciate very much indeed.
I would like to stick with a 2 active devices solution to build a line preamp and make all the efforts to set-up a really nice power supply.
The original voltage rails are 0-60V.
I did build a prototype with point-to-point wiring and all in all I found it very promising (really).
Much better than the expensive commercial preamp I normally use.
I intend to use the original BC107 and a TO126 pnp for practical reason of heatsink, soldering the device directly on their pins.
The schematic you are referring at are very much more complex.
I would like to keep it extremely simple and basic.

Thank you very much again.
Kind regards,

beppe
 
beppe61 said:

I would like to stick with a 2 active devices solution to build a line preamp and make all the efforts to set-up a really nice power supply...

I did build a prototype with point-to-point wiring and all in all I found it very promising (really)...

I would like to keep it extremely simple and basic...


Yes my schematic is for a power amp.

Why don’t you base your preamp on a fet source follower / White follower / complementary follower ? 2 fets, 2 resistors and a good power supply (even batteries). Very easy, no caps, no offset correction, and very good!
 
Justcallmedad said:

Yes my schematic is for a power amp.
Why don’t you base your preamp on a fet source follower / White follower / complementary follower ?
2 fets, 2 resistors and a good power supply (even batteries). Very easy, no caps, no offset correction, and very good!

Dear Mr. Perez,

thank you again.
Please could you explain me a little your proposal?
It sounds extremely interesting but I am afraid I have not understood very well.
I have to declare my almost complete ignorance in electronics.
The only schematic I have seen with only 2 fets are those nice buffers designed by E.Borberly.
Could you kindly provided me any link to an actual schematic?

Thank you immensely.
Kind regards,
beppe
 
beppe61 said:
I understand that with a fet at the input I could do without an input coupling cap.
I would maintain the output coupling cap to be safer (my current power amp does not have an input DC blocking cap).
I am pretty sure this very simple topology has the possibility to sound extremely nice.

Beppe,

here is a possible schematics and results of FFT simulation for 1V RMS output - the gain here is about 7. I did use similar gain stages in the past and if you have time to select JFETs, it could be a very good circuit.

Input here is directly coupled, however in real life you may decide to add a simple RC filter to remove RF, say 10K and 100pF.

Cheers

Alex

P.S. - I've edited the schematics to add the filter.
 

Attachments

  • jfet_bjt_s1.gif
    jfet_bjt_s1.gif
    4.8 KB · Views: 921
Re: Re: Sziklai pair - 1fet+1bjt.

x-pro said:

Beppe,
here is a possible schematics and results of FFT simulation for 1V RMS output - the gain here is about 7.
I did use similar gain stages in the past and if you have time to select JFETs, it could be a very good circuit.
Cheers
Alex

Dear Alex,
thank you so much for your extremely kind and much valuable reply.
That is indeed what I was thinking about !
I still have two question:
1) as I intend to build a line stage gain= 7 is a little too much.
Could you think it could be possible to arrange the resistors value to get let's say a voltage gain of 3 ?
2) would you think that increasing the voltage supply could give some nice improvements in dynamics?
Please excuse my silly questions but I have been too many times said that the higher the voltage supply the better that I am become almoct obsessed with high voltage value.
Anyway this is what I was looking for on principle.
Have you actually built it ?
Very very interesting indeed.

Thank you again.
Kind regards,

beppe
 
Re: Re: Re: Sziklai pair - 1fet+1bjt.

beppe61 said:
I still have two question:
1) as I intend to build a line stage gain= 7 is a little too much.
Could you think it could be possible to arrange the resistors value to get let's say a voltage gain of 3 ?
2) would you think that increasing the voltage supply could give some nice improvements in dynamics?
Please excuse my silly questions but I have been too many times said that the higher the voltage supply the better that I am become almoct obsessed with high voltage value.
Anyway this is what I was looking for on principle.

You can do gain of 3, however as it is indeed a very simple circuit you have to understand the limitations and requirements of it to get a good result. I will try to give here a brief explanation. The R1 and R2 chain gives you the gain value, however to get high open loop gain R2 needs to be relatevely small as it provides a local NFB for J1. On the other hand the DC condition of this circuit would be set by the Vgs of the JFET for a given drain current. This current is defined by R3 and Vbe of Q1, and in the illustration equals about 0.7-0.8 mA . Vgs defines the current through R2, which is in turn the sum of currents through both transistors Id+Ic. Hence the DC voltage on the collector is more or less defined by the Vgs and gain. In the example Vgs is about 1.7V but it would obviously depend on the cut-off voltage of the JFET. If you'll choose a JFET with higher Vgs for the same current the gain will be lower. That why I've said about selecting JFET's - as you'll need two matched FETs with required parameters (for two stereo channels) to get the gain you want. Higher voltage is usually a good thing, but in this circuit you need to get about a half supply on the collector of Q1 and as this is Vgs and gain dependent you'll need to find a good compromise... .

Have you actually built it ?
Very very interesting indeed.

Thank you again.
Kind regards,

beppe

Yes, I did build circuits of this topology in the past many times with very good results. However the resistor values in the picture are for a simulation only :) . You'll need to change them to adjust for a particular FET, gain and supply voltage. There are not that many resistors, thought :) .

Alex
 
x-pro said:
I couldn't attach two files for whatever reason, so here is the FFT graph:

Dear Alex,

I am everything but an expert, but I would call this performance quite good , don't you?
Nevertheless I would not dismiss it because "too simple".
On the contrary I understand that even very famous audio designers have abandoned complex topologies in favour of simpler ones.
I well understend that the quality of the execution matters, but I have the strong feeling that the simplest topologies have the biggest chances to sound phenomenal.
Then I have to be honest with you.
I am very very ignorant in electronics.
I tried to learn the basics in audio design (I mean how to calculate a simple one bjt buffer for instance) but failed.
I must tell you this because your explanation are targeted to an educated expert.
I cannot explain how sorry I am not to understand them well.
Because if I were smarter I would have tried to become an audio designer, in the hope to marry passion and work.
Unfortunately this has not happened.
So any judgement I make on a circuit is from an aesthetic point of view (no, I am not an architect), and your is wonderful in its simplicity.
It is not clear to me which type of fet do you propose in the schematic.
I could not find any U309 fet.
My final solution would be, as I said, your topology with 48V voltage supply and a volatge gain of 3.
If you have any idea on how implement such a circuit I would be very interested in it. That would be my future line stage.
I am sure that this kind of circuit could be a real nice surprise for sound.

Thank you sincerely for your extremely kind and valuable advice.
I appreciate it very much indeed.

Best wishes,

beppe
 
x-pro said:
I couldn't attach two files for whatever reason, so here is the FFT graph:

Dear Mr. Alex,

may I ask you wich software have you used to simulate the circuit ?;)
I would like to ask you another thing.
Do you think that "playing" with the value of R1, R2 and R3 and/or the V2 value it would be possible to get an even better distortion spectrum ?:angel:
I would like to thank you very much again for this very interesting circuit.
I like it very much indeed.

Kind regards,:D

beppe
 
beppe61 said:
Dear Alex,

I would call this performance quite good,
don't you?

Yes, it is very good for such a simple circuit.
2nd harmonic is dominant: -80dB
3rd harmonic distortion is low: -100dB

These are very nice numbers!
I would call it Hi-fi, out of 2 transistors.


One thing, as circuit is now,
we have not added 1 more output buffer transistor.
So best is to keep a light load.
I would suggest 22kOhm or higher load impedance.

With a suitable added NPN follower output stage, it will drive any load.
Like a 5 kOhm potentiometer.
With almost same performance.
 
lineup said:

Yes, it is very good for such a simple circuit.
2nd harmonic is dominant: -80dB
3rd harmonic distortion is low: -100dB
These are very nice numbers!
I would call it Hi-fi, out of 2 transistors.
...


Thank you very much Sir.
I would like to get your opinion on the spectrum distortion graph.
Do you think it is the most dependable way to assess by instruments the quality of an amplification device?
This question about the ultimate measurement to assess the sound quality of an audio equipment will accompany me till my old age.
I mean, someone say these graphs tell nothing.
Others propose them as valuable.
I do not know who to trust.
What makes me crazy are units that measure worse and sound better !! how can this be ? but it seems it is.


Thank you very much.
Kind regards,

beppe

P.S. I could not enter your web site. Is it for Members only?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.