high performance 25W PowerAmp

Precise:
no current mirror, no cascode (of any kind), also the current source / or sink should be omitted - and be clear about why the original was designed inverting (we can only guess, or well estimate) ... at least a regulated or at least stabilized power supply for the complete first stage, Class B OPS as diamond-setup.
 
Are you interested in the L.Stellema design, a real-life 2024 new version?
I posted my circuits just for your inspiration, personally I have no intention to built anything like that.
Finally it is your job to prep up that original schematic for something worth to built for you.

If you have a good updated design, please post it, I will be pleased to put it in the sim.

availability of components
Most parts I used are standard parts except the output transistors, I chose them randomly from the MJ list in Microcap because I have no 2SC & 2SA models.

I can not react to all objections, it just does not make sense.

Just one more:

At that stage it is of absolutely no relevance how I would keep the 1A bias stable.

(Then I suggest retaining the three assemblies IPS, VAS, OPS as far as possible.)

My ops is basically the same.

I do not like the VAS for various reasons, so I changed it.

The ips... in my opinion the two followers give no advantage, so I omitted them.
The author claims that his inverting ips with input at virtual ground has benefits, I doubt that. If it was true, everybody would built inverting opamps stages. So I set it back to normal.

The L.Stellema concept shows strengths in the idea
What are those strengths?
 
Yamaha C2x final stage of output buffer

yamahac2x.jpg
 
I posted my circuits just for your inspiration, personally I have no intention to built anything like that.
Finally it is your job to prep up that original schematic for something worth to built for you.
You're probably right about that.

Unfortunately I already knew your suggestion, in 1990 I saw this form in a commercially very successful amplifier, of course not exactly 1 to 1. So no inspiration, just a memory.


But nevertheless, a big thank you for your efforts.
 
Most parts I used are standard parts except the output transistors, I chose them randomly from the MJ list in Microcap because I have no 2SC & 2SA models.

I can not react to all objections, it just does not make sense.
My comment refers to your floating point values /numbers,
On the basis of which one could assume that you are only pursuing the matter for fun - and so it was.

If you have a good updated design, please post it, I will be pleased to put it in the sim.
I have explained my next steps. Of course I will report on the progress and also post the final version (i.e. “Tim”) here in the forum. I will gladly accept your offer so that we can compare my reality with your simulation: Measurements on the object vs. MicroCap simulation. How do we handle the models you will need for this task?

The ips... in my opinion the two followers give no advantage, so I omitted them.
The author claims that his inverting ips with input at virtual ground has benefits, I doubt that. If it was true, everybody would built inverting opamps stages. So I set it back to normal.
As I am not looking for disputes, it will probably be better not to make any further comment on your opinion.

For the time being, I will deliberately leave the entry stage as it is at its core.
So the performance thd wise will be very limited to begin with. I know, there is the "less is more" faction...
For my own amplifier I already have different plug&play frontend boards, could be that finally I will prefer a slightly downgraded version.

Are we playing rhetorical games?

Right now, as I write this, I'm listening to Henry Mancini's Peter Gunn - with an ancient, restored and slightly modified Technics SU-V50 hifi amplifier. Perfect - and this Squire belongs to an extremely low THD community.


But even a simple NAD 3020i can produce an excellent THD curve, mine does (but it's no longer in its original condition - although it hasn't been morphed).

So please don't put me in the wrong place.
My favorite would be the legendary distortion-free amplifier, the ideal.

;)


All the best,
HBt.
 
For my own amplifier I already have different plug&play frontend boards, could be that finally I will prefer a slightly downgraded version.
I'm glad for you, I don't go to that much trouble.

With regard to Tim, you are of course absolutely right, if you forbid yourself the good things, then you can't expect so much from the result.
He should just work, that's enough for me.

THD below 0,03% in all circumstances, perfect. I make no secret of that.
 
Excuse me wahab,
but the “early” objection no longer works for me - for me it's nothing more than a rhetorical evasion.

Just say that you prefer h22e, i.e. an rce of 55k555 ohms, to 16k666 ohms ...
I have an original, old Siemens data book in front of me.

#
In plain language (no attack),
Your explanations are not enough for me personally. But now the thread is easily drifting into a turbulent and controversial star chef number - and I don't want that.

What i see is that if i replace in the differential the 65V BC546 by 45V BC550C THD increase by about 50% (mainly H2).

Of course that s with very low distorsion, so this doesnt change things much if any at all, but still, i m more comfortable
with 65V devices than 45V ones in designs that use +-25-30V rails, that the higher Vce devices produce less distorsion
is a pleasant feature on top of better reliabilty.
 
  • Like
Reactions: 1 user
What i see is that if i replace in the differential the 65V BC546 by 45V BC550C THD increase by about 50% (mainly H2).

Of course that s with very low distorsion, so this doesnt change things much if any at all, (...)

Apart from that,
the question now is: why is H2 actually imperceptible, but nevertheless smaller (in the sequence C -> B -> A types of the same family) - according to your observation, determined on the basis of the simulation.

This has absolutely nothing to do with dielectric strength, even if we should, or rather must, avoid collector-emitter breakdown!


What can we do? Take the data and calculate:
local voltage gain (without global negative feedback, i.e. in the open loop view /case) 1 to 1.22 to 1.311 times (C to B, C to A).

Who would have thought that?!



HBt.
 
Perhaps I was not clear enough, so I will repeat the following fact:

BC - 546, 547, 548, 549, 550 is the NPN family
BC - 556, 557, 558, 559, 560 is the complementary PNP family


Vce_max is not the parameter that led to 'wahab's' observation, but the subdivisions into A, B and C bags are responsible for it. As we had already disguised with the "Early-Number throw-in".

The rce-understanding then automatically leads to the cascode at some point.


But in this project, the key question is how to optimize the T.I.M. with the simplest means?.
 
Last edited:
BC546A/B perform both better than the BC550C, and both have lower beta but also higher Early voltage.

I can only agree with you.

Are you still simulating the original Stellema?
No, because there s no point simulating what is badly thought, as said either you ll end with a blameless or if you want
a symmetrical VAS that will end as a symmetrical input dual differantial, in the first case this will require 14 transistors
and in the second 15 or even 13 since active CCS are not needed.

Both will perform very well in respect of the Stellema amp whose IPS output cant even swing at lower than 0V,
wich is total non sense when it comes to drive a symmetrical VAS, not sure that the author even noticed it,
otherwise he would have done things differently.


Stellema use Rails below +/-24Vdc, and as Ed stated right at the beginning, it must be a regulated PSU.

The crucial point is rather the delta v_ce.
You ll never know, when powering the circuit nothing is settled at the final values from the start.
 
BC546A/B perform both better than the BC550C, and both have lower beta but also higher Early voltage.

That is correct dear wahab!

But you don't seem to be interested in why. Perhaps you would like to keep this very trivial secret to yourself. Otherwise, we humans often lose some of our basic skills or acquired understanding of a subject if we sit in front of the simulator too much (we seem to leave the thinking to the machine, it seems to be omnipotent).
I use SPICE in a very targeted way, I use this tool consciously and scrutinize the output. I don't even work without plausibility checks and the respective why question. The practical test must also (still) be passed.

But let's get to the point and get back to the Siemens BC-54x (including 550) family, divided into three classes, namely A, B and C. The same applies to all members of the classes, they bring an r_ce, a beta and an r_be in their satchel.


Class A:
rbe 2.7kOhm; beta 200; rce 55.55kOhm
Class B:
rbe 4.5kOhm; beta 330; rce 33.33kOhm
Class C:
rbe 8.7kOhm; beta 600; rce 16.66kOhm


In the national youth competition, the classes are now to show what they can do in the "VAS" discipline. The framework conditions are the same!
Results,
on average the respective class members threw the sling ball OLG far:


1st place
Class A with 386.1 OLG points

2 st place
Class B with 359.32 OLG points

3 st place
Class C with 294.6 OLG points

The crowd of spectators was impressed by the incomprehensible fact that the fat, thin, strong, weak, tall and short riders in classes A, B and C were equally distributed.
 
  • Like
Reactions: 1 user
(...) and both have lower beta but also higher Early voltage.
I hope you're not deliberately trying to mislead young readers, dear wahab?


Let's be clear once again:
your original conclusion, and therefore your deduction that a greater dielectric strength is the cause of your observation, is wrong.

Flatter branches of the characteristic curve of the 1st quadrant of a BJT to be described automatically lead to a larger EARLY voltage. That is correct!


We are only interested in r_ce at the operating point, but also in S or gm, i.e. the slope ... 74.1mS applies to all members of class A (73.33mS for B; 78.96mS for C) ... all practically independent of the value V_ce_max of the individual.



#
You know wahab, I don't feel like this kind of communication, it's too exhausting and extremely annoying.
 
No, because there s no point simulating what is badly thought, as said either you ll end with a blameless or if you want
a symmetrical VAS that will end as a symmetrical input dual differantial, in the first case this will require 14 transistors
and in the second 15 or even 13 since active CCS are not needed.
For my part, I know both alternatives inside out, but it's not about these two topologies - which you can choose, but don't necessarily have to.

If you find the approach, the idea of the 1980s feature article so extremely bad that you never tire of disqualifying it, the question must be allowed: why don't you simply improve the design in detail instead of drifting off in a destructive direction?

Please, you don't have to answer this question.

It is a great pity that nothing constructive (from your expertise) can actually be expected. I regret this very much.


Both will perform very well in respect of the Stellema amp whose IPS output cant even swing at lower than 0V,
Assuming she can't do it, without any questioning, she can't(!) - does she even have to be able to do it if she couldn't(?).

wich is total non sense when it comes to drive a symmetrical VAS,
What follows is an impedance converter, nothing more and nothing less! An emitter follower in push-pull, followed by ...

not sure that the author even noticed it,
otherwise he would have done things differently.
Rhetorical games - as I have already said :cry:; none of us can know what L. Stellema would have thought or done if he knew that..!

Perhaps the published proposal is actually a stopgap solution, perhaps his professor said to him after a laboratory experiment: "my student, your experimental setup is so great, we can make an academic publication out of it" - perhaps.

Who knows exactly and who cares at all ;)?

#
Finally, I would be happy if we all (those who are interested!) could make the project a success together. I can do it on my own, but then why do we need a public space, a DIY forum?


All the best,
kind regards

HBt.
 
From now on I will refer to the project name Tim, an offshoot of the aforementioned article placement /schematic.
I think it's good if we refrain from any disparagement of an unknown person and his placement /schematic. Other circuit ideas that deviate from the actual topic of this thread should be discussed in their own threads.


HBt.
 
You ll never know, when powering the circuit nothing is settled at the final values from the start.
There is a totally trivial solution for this too - and I suspect you know it too.

You pay attention to the switch-on sequence and the vast number of soft start options etc. - just one example.




#
Now I release the game again, Tim is on the ball (and is on his the way to the goal).