Simple Symetrical Amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi Bigun ;)

More power dissipation - why is that bad ?

It is not bad, just different. Good for someone, not so good for the others. :rolleyes:


I don't understand why Cbc is going to have much influence since unlike the VAS there isn't much of a voltage swing on the collector of the input devices ?

More explanation here. ;)


Did you try listening to both ?

For a long time I didn't built an amp without cascoded input, see it as an advantage but I fully aware that the sound signature is simply different. :)

P.S. How is doing your SSA built? I am very enthusiastic about the quality of your sch especially with compound output drivers. :xfingers:
 
SSA Full Balanced Input

SSA full balanced input current feedback front-end. Maybe I'll try it someday. :rolleyes:
 

Attachments

  • SSA Balanced Input.jpg
    SSA Balanced Input.jpg
    190.2 KB · Views: 1,160
VERY inspiring design LC
I might try SSA bipolar version, scaled down for headphones , 32-55 ohm @+/- 24V reg.PSU with MJE150xx output.
In simulator works like promised!
I have also simulated (don't laugh!) SSA quasi. Not so symmetrical, I admit, but I have many power NPNs left from past century :)
 
Disabled Account
Joined 2006
VERY inspiring design LC
I might try SSA bipolar version, scaled down for headphones , 32-55 ohm @+/- 24V reg.PSU with MJE150xx output.
In simulator works like promised!
I have also simulated (don't laugh!) SSA quasi. Not so symmetrical, I admit, but I have many power NPNs left from past century :)

Try the circuit shown at the bottom of the page, the original JLH circuit, sounds great, use cfp for the outputs and highish beta input transistors. With HD600 phones only 2nd and 3rd harmonics with dominant 2nd and no high order harmonics. THD is exceptionally low. Youll be hard pressed to find a better sounding headphone amp.

http://www.tcaas.btinternet.co.uk/jlhphones.htm
 
SSA full balanced input current feedback front-end. Maybe I'll try it someday. :rolleyes:
May-be need of a trimer in the input to tune the common mode rejection. Once again, i do not understand the cap between the two emitters lines. Their trans are feed by a constant current, not ?
Beautiful, as habit, the way you mix the two first stages. Can be implanted on all the boards, as you can simply remove the two transistors and 4 resistances of the - stage if you do not need-it.

By the way, there is just one version missing to explore all possible combinations: with fets both in input and output.

I had spend hours with LTspice, trying in vain to make working a design witch work in the real life. What a piece of sh... and a lossy interface !....
 
Last edited:
Once again, i do not understand the cap between the two emitters lines. Their trans are feed by a constant current, not ?

In theory the potential between middle points of a splited feedback loop-bridge is DC. All AC variatios should be in phase so zero difference, in theory. So the capacitor is there to bring the bridge to near the ideal theoretical conditions. ;)


Beautiful, as habit, the way you mix the two first stages. Can be implanted on all the boards, as you can simply remove the two transistors and 4 resistances of the - stage if you do not need-it.

Yes universal unbalanced/balanced SSA PCB fit the needs for RCA or XLR inputs as well. :cool:
 
Last edited:
Thanks for the quick reply!

More explanation here. ;)

yeah - I got the bit about reduced miller capacitance, but.... with my limited experience I am relying on what I read about Cascodes. I read that they are helpful, even compulsory, for JFET inputs because these little fellas can have high capacitances from gate to drain which hurts freq. response. I read that cascodes work wonders on the VAS, in fact I may yet use one, because this is where the vast majority of the voltage gain is being made, and hence it's where the miller-multiplication rears its ugly head. But for a BJT input device I don't see a lot of miller multiplication and I don't see that theres too much h.f. roll-off. In simulations I find it hard to see any benefit from the cascode of the input device. The OLG plots for with and without appear the same.

For a long time I didn't built an amp without cascoded input, see it as an advantage but I fully aware that the sound signature is simply different. :)

This is where I was heading with my question. There has to be something more to it than a marginal 'engineering benefit'. But 'different' doesn't imply 'better' and so I am left with the thought that the cascode is not 'earning it's place' in the design yet ? - not that I doubt your experience but I like to question things in order to learn.

P.S. How is doing your SSA built? I am very enthusiastic about the quality of your sch especially with compound output drivers. :xfingers:

Progress is a bit slow due to my day-job. This week they sent me off to a 1-week training course (which was excellent by the way).

I am absolutely confident in the design of the amplifier, with or without cascodes. I don't have any doubt it will serve my application well. So I'm looking at a pcb layout design rather than prototyping the amplifier first. This means learning some new things, not only how to use Eagle software but also different way of layout and mounting of power devices. I am considering incorporating a speaker protection circuit on the pcb so I'm playing with some simulations of this part of the design today.
 
Beauty in Simplicity

Lazy Cat,

While reading this thread I'm so amazed with Yours pure genie simplicity, and Yours target . . the beauty of
purest sound playback possible.... Congratulation to You on such effective simple circuit with such great performance.
While You mention the SSA variants possible, there is the top one to be realised in future, an X Balanced version, similarly like F5. Perhaps unmatched performance especially if implemented with Lateral Fets in a BIBGT stile. A Holy Grail of Amp. :)
Thank You for all Your great share to the Audio community.

Best regards,

Andrew
 
IYes universal unbalanced/balanced SSA PCB fit the needs for RCA or XLR inputs as well. :cool:
I had made a symetrical amplifier for my recording studio with two bridged ones on separate power supply for each half bridge. So floating that i was obliged to add two high values resistances to the common ground on each.
I love bridged amps, they reduce distortion in a nice way, if they are designed for the low charge. I don't know why this is not more used by industry, as it can reduce prices by the mass effect.
 
Lazycat, the sims look very good to me, maybe it's my turn now.

Meanwhile I am enjoying the opamp version very much, with zero bias on the FETs. ;)

I'll start buiding my SSA tommorrow. No veroboard in stock. :(

I'm taking your word LC, that I don't need a servo. SSA better behave, or else... :bomb:




:D
 

Attachments

  • SSA.png
    SSA.png
    9.6 KB · Views: 783
  • SSA_1.PNG
    SSA_1.PNG
    47.9 KB · Views: 749
  • SSA_-graph.png
    SSA_-graph.png
    8.8 KB · Views: 737
  • SSA_bw.PNG
    SSA_bw.PNG
    53.1 KB · Views: 723
What is the PSRR figure of this design and how to improve it?
I believe PSRR is never a problem: A good amp need a good power supply. And, on my point of view, it is even as important than the amp design itself. Will try a good SMPS for my future class D+my modified SSA amp: i have a friend witch has changed the original PSU of a Quad amp for a SMPS and report a much better sound. May-be the little 100kz ripple you cant get rid off ? I had noticed several time that, adding a little amount of HF in an amplifier (as long it is fast enough to deal with it) improve the transparency and separation between instruments. i dont know why. Mechanical effect on the tweeters ? Or what in the amp ?
 
Lazy Cat,

While reading this thread I'm so amazed with Yours pure genie simplicity, and Yours target . . the beauty of
purest sound playback possible.... Congratulation to You on such effective simple circuit with such great performance.
While You mention the SSA variants possible, there is the top one to be realised in future, an X Balanced version, similarly like F5. Perhaps unmatched performance especially if implemented with Lateral Fets in a BIBGT stile. A Holy Grail of Amp. :)
Thank You for all Your great share to the Audio community.

Best regards,

Andrew

Thank You Andrew ;)

Yes quite some SSA variants were presented in this thread, most of them just as an idea, others already driving the speakers. SSA splited feedback loop is actually a classic resistor bridge that allows DC operating conditions + AC feedback to the input stage. That opens a lot more possibilities than usual two resistors divider network and enables the designer to implement SSA principle to many types of standard input configurations, like some examples shown in this thread. :)

I agree that the last presented SSA balanced input plus BIGBT with laterals drivers would be a great amp ... only to find the time to make it is for me at the moment mission impossible. :spin:

Regards, Andrej
 
Lazycat, the sims look very good to me, maybe it's my turn now.

Meanwhile I am enjoying the opamp version very much, with zero bias on the FETs. ;)

I'll start buiding my SSA tommorrow. No veroboard in stock. :(

I'm taking your word LC, that I don't need a servo. SSA better behave, or else... :bomb:




:D

The reality check is also OK. :D

So OPAMP is lowering distortions in B-class bias setup. What, you supply the amp from a battery? :cold:

DC servo is realy not needed here at SSA. If you make all the steps correctly: hFE match, input BJT thermall joint and proper connections according to schematic, than you will get +/-10mV max offset variations below 1Hz, which is negligible. ;)
 
SSA Balanced Input L-Mosfet Output

I already made it in post #823 but that was more like transforming classic differential input to full balanced SSA version. :rolleyes:

This one is complete transformation of the classic SSA to full balanced input with lateral mosfet output. As always, some will find it useful, others will stick to unbalanced input as well. Enjoy. ;)
 

Attachments

  • SSA Balanced Input.jpg
    SSA Balanced Input.jpg
    239.6 KB · Views: 732
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.