Current FB Input Stage Idea (power amp)

[lineup]

There are plenty of way to arrange input stage in a current feedback amp.
I have seen many. Standard is 4 transistors in a Diamond type figuration.
But also used is 2T and a bunch of resistors in a network.

The way I show in diagram uses 2T and 2 Current sources.
I have never seen this way before, so this I have figured it out myself.
But to me this method is rather obvious, so I think this must have been done before.

The heart is within the smaller rectangle.
2 current sources creates 0.6V across R3 and R13, 47 Ohm.
To get a good gain in first stage I use rather much current, 14 mA and 47 Ohms resistors.
R8 and R7, 150/10 Ohm sets the Gain of amp. Here we have x16
Current Feedback comes to R8.
I keep feedback resistance low in order to get low distortion.

VAS is only a pair of BC556/BC546 working at 5 mA

Output stage in the bigger rectangle is very standard.
Quiescent current for MJL3281/1302 is at 50 mA

Amplifier shows good performance. 25 Watt in 8 Ohm.
What is special for CFA amplifiers is the speed.
And this one shows stable -3dB at around 4 MHz
So it is very fast.

 

[lineup]

Yeah!
I know.
When peole see something different from 2 transistor diffferential input
and the Blameless concept
they get nervous.

Their brain is not used to much more than blameless.
So CFB amplfiers must be suspect.


One nice thing about Current Feedback Amplifiers is
that they are symmetrical.
There was a time meself liked the usual 2T diff pair. Like blameless.
But lately I just love Symmetrical Amplifiers.
John Curl likes sym, too.

Happy Greetings
/lineup

 

[wahab]

That is similar to this one , which date from about 1977..

Designed with Otala s "ideas" in mind.
I would say, to the extreme, since open loop
gain is drastically reduced by use of degeneration and vas
loading in the aim to display a 20khz open loop BW...

Funny is that increasing the open loop gain by increasing
the impedances that load the vas allow vast reduction of
THD...

 

[fotios]

Hi Lineup. I missed you my friend. How are you doing?
Hope everything is fine with you.
Fotios

 

[moschfet]

Hello lineup,

this is a very interesting circuit. My experience with amplifire in current-feedback-topologie ruled me to layout very tidy.

Because of the oscillations, the amps burning faster, than you can pull the plug.

Do you want to build it up in real?

 

[ostripper]

Yeah!
I know.
When peole see something different from 2 transistor diffferential input
and the Blameless concept
they get nervous.

Their brain is not used to much more than blameless.
So CFB amplfiers must be suspect.


One nice thing about Current Feedback Amplifiers is
that they are symmetrical.
There was a time meself liked the usual 2T diff pair. Like blameless.
But lately I just love Symmetrical Amplifiers.
John Curl likes sym, too.

Happy Greetings
/lineup

They are nice .... at cancelling out even order harmonics , leaving the "nasties" (odd orders) for us to hear. They have to be very low in distortion by design just so we don't hear what is "left" (the nasties).

Your amp is very similar to member SYN08's VSOP , which is a stable proven design (he made the prototype). It has very low distortion , but what distortion it does have is all odd order. Just an observation.

Electrocompianet actually uses a single ended stage to drive a symmetric stage . I guess one cancels out the evens while the other cancels out the odds.
OS

 

[hahfran]

There are plenty of way to arrange input stage in a current feedback amp.
I have seen many. Standard is 4 transistors in a Diamond type figuration.
But also used is 2T and a bunch of resistors in a network.

The way I show in diagram uses 2T and 2 Current sources.
I have never seen this way before, so this I have figured it out myself.
But to me this method is rather obvious, so I think this must have been done before.

Yes yes...but...current fb has poor CMRR so a separate stable PSup for the predriver is inevitable. Without that you can be sure the amp starts to oscillate at recovery from overdrive ( when supply voltage drops) . If you provide a separate stabilized supply such CFBs excel with high open loop bandwidth and little compensation problems - unlike Voltage feedback where you'll need to compensate the open loop bandwidth down and down again. Old buddy Mr. Williamson who is somewhat famous for the first HiFi tube amp, the Williamson amp, had noticed that and acted accordingly.
In replicas of that design you'll for sure find stabilized supplies for the pre-stage.

Hans Dieter

 

[hahfran]

Electrocompianet actually uses a single ended stage to drive a symmetric stage . I guess one cancels out the evens while the other cancels out the odds.
OS

That is correct. It is indeed possible to have a predriver single ended that has
a particular thd and a particular distortion spectrum driving a sym power stage
that as also its thd and spectrum and these two together have less thd than
each of the stages. That depends much on open loop bandwidth and lead lag compesation but it works as such that the harmonics spectrum remains constant with rising frequency of the signal.

 

[assembled]

What about using something like THS6012 and a discrete output stage?