Which Slone's cct? I've built the "Optimum Power Amp" of Slone's and it doesn't work at all.
Mike's design is good (Sounding good), I think. It has low odd harmonics, and topping even harmonic.
Mike, could you tell more about the sound? Is it fatigue or not?
Mike's design is good (Sounding good), I think. It has low odd harmonics, and topping even harmonic.
Mike, could you tell more about the sound? Is it fatigue or not?
Hi, Mike,
Just notice something in your design. The CCS in differential is bigger than the one in VAS? (100ohm-330ohm)?
Usually VAS has more than differential, isn't it?
Just notice something in your design. The CCS in differential is bigger than the one in VAS? (100ohm-330ohm)?
Usually VAS has more than differential, isn't it?
Yes, with ccs.I built the schematic which you post at " single diffential or dual differential" .Oh! It is a exciting thread
thanh said:
Yes, "hiding". If a frequency is loud enough, the human ear is
not able to detect near frequencys below a certain threshold.
This mechanism is used in mpegsound to discard "unnecessary" data.
lumanauw said:
From the 1st time i tried CFP in the diffamp in simu, i started wondering
why not everybody uses them ? Maybe because of thermal stability ?
Or because you need PERFECT matching of the devices ?
Or because its patented ?
And the price: 4 cents each part make 16cents extra cost, wow
that's expensive !
But a strange world, where a resistor has the same price as a BJT.
Mike
lumanauw said:
Thanks !
It's not fatigue, its more like exciting as i discovered so many
new details in the music i've never recognized before. But
my speakers are not really highend... It seems to me like a
"perfect" amp, its just playing, no colorizing, softing etc.
I was surprised that just changing/replacing the amp can
improve the sound so dramatically.
lumanauw said:
As i use small bjts i can't use more current in the VAS, they would
heat up. The current through the whole diffamp is 4.8ma, makes
~2 ma in the cpf-bjts. The VAS has about 1.8ma.
But as i use tripledarlington, the vas does not need high currents.
I just wanted to use maximum possible current without stressing
these small bc546's.
lumanauw said:
It nearly kills 3rd harmonic, because the currentchanges in the
diffamp are reduced to a minimum. The slightest currentchange
produces large outputs. And it automatically balances the diffamp.
The circuit itself generates a large amount of 2nd harmonic.
Sounds like the perfect solution, but at least in my simus this
circuit makes the diffamp too slow and only large cdoms stabilize
the whole circuit. So the result is a perfect amp with 1khz signal,
but a disaster at 20khz.
Openloopbandwidth reduced below 1khz...
I believe that many amps suffer from this circuit, this might be a
reason why so many amps sound so boring/soft, hiding any detail.
That's at least my problem with my yamaha, it does sound soft, what
is not bad, but its softing everything and the result is some sound
without any separation.
That's only an assumption, i haven't built an amp with currentmirror
yet. Maybe you need some RF-transistors for the currentmirror ?
For a subwooferamp i would recommend the currentmirror.
Mike
I read this stuff in D Self book. If I'm not mistaken he said that wider openloop bandwith will have better audible sound.
Is there anything like openloop bandwith up to 20khz, or usually only around several khz?
He sugested to put R 220k parrarel with Cdom, to widen the openloop bandwith.
How true is this?
If Open- loop bandwidth is enough large ,TIM will be zero
http://www.angelfire.com/ab3/mjramp/golopid/golopid5.html
lumanuaw! The schematic which I posted doen't work with 100mv input signal😀
http://www.angelfire.com/ab3/mjramp/golopid/golopid5.html
lumanuaw! The schematic which I posted doen't work with 100mv input signal😀
lumanauw said:
I believe good openloop-bandwidth is one of the most difficult part
in designing an amp, as you want a stable (non-oscillating) and fast
circuit. I believe that it's possible to achieve openloopbandwidth
up to 20khz, but very hard. I think you need a very simple circuit
like the aleph or else. My cct goes up to ~10khz, due to the small
cdoms (10pF), and the "high" current in the diffamp.
I can't say which openloop bandwidth is common, it depends on so
many things. But it seems easy to build an amp with below 1khz.
I need trying the 220k...
Mike
Hi, Mike,
Please do so in Sims. I also want to look at the effect of 220K parrarel with Cdom. I have found one commercial unit (quite expensive) that do this.
I think it has to do with gain-bandwith product. Make amp with high gain, but low bandwith or we can create moderate gain but have wide bandwith.
Please do so in Sims. I also want to look at the effect of 220K parrarel with Cdom. I have found one commercial unit (quite expensive) that do this.
I think it has to do with gain-bandwith product. Make amp with high gain, but low bandwith or we can create moderate gain but have wide bandwith.
MikeB said:
Welcome to the club !
Of course the device that interfaces with the speakers - which are hard to feed (compared to a pre-amp that have a 47k ohms load for example) - makes a lot of difference.
I had the same experience as you with sound improvement by amp and it is quite refreshing to hear the music closer to reality!
Fab
Just aware of something. Non-global feedback power amp is a power amp with flat open loop up to 20khz, isn't it?
lumanauw said:
yes, it is supposed to be that way.
Your previous post is also correct "Make amp with high gain, but low bandwith or we can create moderate gain but have wide bandwith." There are other thread on this subject but I believe that everybody does not agree on the goal to obtain (high open-loop gain or high BW). Like anything in life, it is a question of compromise.
Fab
There are 2 design that I want to know how it sounds. First is a non-global feedback design. Does this design have good bass (I always tought tight bass have to do with enough feedback).
Second is the error-correction output stage (like Hawksford). Does this gives better audible sound? Why there are experts who say to skip this error correction, and use parrareled output stage instead?
Second is the error-correction output stage (like Hawksford). Does this gives better audible sound? Why there are experts who say to skip this error correction, and use parrareled output stage instead?
lumanauw said:
lumanauw
I am not one of the experts you are referring to but I believe that parrareled output stage has a much lower output impedance and less thermal variation by devices for same amount of power because of current sharing. So, the error correction may not be needed and the error correction itself may not be perfect I suppose.
Fab
thanh said:
Hi thanh !
Are you familiar with the AC-Sweep function ?
I create the openloop by cutting the feedback, and connecting
the '-'input of the diffamp to gnd. Then careful adjusting of the
DC-Offset is necessary, sometimes i get resistorvalues like
536.417ohm, but it's done quick in a few minutes.
Then i set the input to some 0.1mv, and check with the AC-Sweep
the output. The Output divided by the input is ~openloopgain.
This means if output is 14v, openloopgain would be 1:140000.
With the AC-Sweep you also see the bandwidth, but don't get
shocked, transistors are slow ! I had one circuit with -3db @ 500hz !
Mike
- Status
- Not open for further replies.