a wonderfull sziklai

[AndrewT]

4pr @ 150W each is 1200W of devices.
Divide by 6 to get 200W of maximum output power.

200W into 8r0 from 4pr of c5200

[kml88]

okay ... thank you, for how many pairs of 5200 and 500W for the scheme which is right? sorry with my english

[AndrewT]

Kml,
what does post201 say?
Why have you sent me a duplicate PM asking the same question?
What are you playing at?

[kml88]

I just want to understand more about the power of this, sorry

[catalin]

Hello, I had the ideea to amplify the miller capacitor(if the transconductance permits) by a common base stage (to have gain without phase modification) and now we gain at least 2 things :
1. The miller cap can be decreased , we gain bandwith and SR ;you can try in your amps .
2. The miller cap in DC has almost 0 potential ,from here the noise of the miller cap is reduced .
I tried on my amp and now it is very stable with the 12pF cap .Also on square wave signal .

For the moment I don't want TMC because it introduce bad behaviour with capacitive loads .

[AndrewT]

Hi,
nice idea. Keep us informed of how it works when listening and when speakers work the output stage.

How did you determine the values of R12, R13, R22 & R23?
How did you determine the values of C1 & C5?

[catalin]

Hi Andrew ,
In listening test I don't see difference until now,maybe a little on the treble ,being more clear only on some parts of musical pieces .But it is just a personal opinion or a placebo-effect .I need to listen more to detect difference.

R12 is set to 10k to linearise a little the thermal deviation from BD139,we have there 0.2 mA through 10 k .In place of R13 practical there is a adjustable resistor with 5 kohm value or 1k series with 2.5k adjustable .It is shown in previous schematic ,version 4.7 and of course that is setting the current through the final stage .The resistors is a divider and is setting the voltage on the base of Bd139 .The current through BD139 is constant to 4.3mA so the divider can be calculated for the desired voltage Vce .

R 23 is setting the current of the follower bjt therefore to have the smallest noise but also with good thd so 1k for 1mA is set .Also R23 depends by R7.

R22 is there to set the maximum current through the follower and from here the maximum voltage injected in VAS .In transient signal or square wave the follower is reaching very high current which can destroy it .Therefore the maximum current is set to 3 mA with R22 .It is just a follower protected .

C1 is coming from C>1/(2*pi*f*R) where f is the lowest frequency desired and R is the 10 K input resistence .But if we want to have the gain of the filter almost 90% from the 0 db then we shoul choose a value of 6 times bigger .
I choosed a 5 hertz frequency not for the gain because I don't hear that f but I wanted a linear phase at low f .Therefore I choosed a C=6/(2*pi*5*10k)~19uF .

The other filter R11+C5 must add a small load to the preamplifier/output opamp,etc .The reactance of the cap will be 1/wC where w is 2*pi*f .F is choosed 20khz .In this case the reactance of C will be 1/(2*pi*f*C)~80k .
Therefore the smallest capacity, the smaller the load will be at high F like 20 khz , but we should not decrease it too much because we need also to limit the bandwith of the input signal .But ..

For a SACD there are some test that prove that a 50V/us per 100W power is required, therefore the input filter should not add a degradation of the SR .I calculated the filter to limit the SR at almost 100V/us .If Rail to rail voltage of the amp is 70 volt and the amp can perform 100V/us then the t rise/fall is under 1us so also the input filter should not slow down under 1 us .I choosed
a 0.6 us delay at input therefore if Time=R*C then C=T/R so 600pF would be the result .But also here to have the output voltage at 90% from 0 db then a 6 time smaller cap is good .The 6 number comes from the exponential function for 90% .
Also another advantage to choose a filter with high R and low C is that the noise and distortion of the cap is smaller compaired with a greater capacity .
I always prefer to have the smallest possible capacity by increasing the resistance in the circuits because the distortion of cap can't be simulated and also the resonance of their armatures .

[Stee]

I'm successfully using a Chinese CFP offering outstanding stability thanks to a simple resistors separation between stages

VERY IMPORTANT: use final devices with very low hFE

LJM-Assembled MX50 SE Power amp kit Stero Amplifier boards | eBay

[catalin]

Hi Stee ,Here wasn't issues from the final stage .If the drivers are fast enough the CFP stage has no issues .Here the issues was comming from the fast voltage amplifier stage behind the final .Because was difficult to compensate with small capacity like 12pF with normal Miller compensation .Because with this topology I wanted to have a slew rate about 100V/us .
But now is possible with this compensation.

[Stee]

ok but the speed is not the only quality
Motorola tries to use these then tell me

check out the inverting input too
is much more natural
(with toroidal 50VA 10:1 230V-24V)