140v single rail Nmos amp..

Status
Not open for further replies.
the amplifier's speaker output terminals.

The R+C at the amp's terminals completes the HF load that the amp needs to see and it also attenuates interference coming in through the speaker cables.
The Zobel alone does a very bad job of attenuating cable induced interference getting to the -IN input.
The Thiele improves on this significantly.
the Pi improves attenuation further still and seems not to interfere with sound quality since it is well separated from the amp output by the full Thiele Network.
 
the amplifier's speaker output terminals.

The R+C at the amp's terminals completes the HF load that the amp needs to see and it also attenuates interference coming in through the speaker cables.
The Zobel alone does a very bad job of attenuating cable induced interference getting to the -IN input.
The Thiele improves on this significantly.
the Pi improves attenuation further still and seems not to interfere with sound quality since it is well separated from the amp output by the full Thiele Network.

I am not familiar with Thiele network, how can it be implemented on this amp?
 
a Zobel R+C from amp output to amp power ground plus a L//R from the amp output to the amp's speaker terminals = Thiele Network.

Thiele also moves the Zobel from before the L//R to after the L//R

The values of the components change depending on speaker impedance, amplifier's required HF load and which Thiele Network is being implemented.

My idea of the Pi filter is combining the two optional Zobel locations either side of the L//R. I have seen this very occasionally in other designs. It was when I read Cherry's paper on the Thiele that I came to realise what I had seen and started experimenting with it.
 
Hi,

Single supply rail with zero thump using two series caps in the rail in stead of one. Why no thump, because capacitors are charged in series and juntion is at rail/2, there is no voltage developed over speaker.

Regards

Nico
 

Attachments

As Hugh explained. R5 = R6 (the NFB resistor); R7 = R8 could be anything say 47K; C2 = C3 say 10 uf to couple input ground to amp DC ground. Your amp should work fine.
 

Attachments

  • pic.JPG
    pic.JPG
    16.4 KB · Views: 226
Thanks Nico,

#44 is brilliant, we have seen it before on a headphone amp you did on my forum, and it gets around all the switch-on problems. However it does place the speaker lines at +70V wrt ground, which could be a problem in a PA/sound reinforcement scenario.

Thanks again for the input, Nico, much appreciated,

Hugh
 
Thanks Nico,

#44 is brilliant, we have seen it before on a headphone amp you did on my forum, and it gets around all the switch-on problems. However it does place the speaker lines at +70V wrt ground, which could be a problem in a PA/sound reinforcement scenario.

Thanks again for the input, Nico, much appreciated,

Hugh

Hi Hugh, would you agree that the output of the amp at steady state is at half supply and that the center of the caps are at half supply. Why would it be different for P/A use?

PA amps also uses a Push Pull output arrangement, this configuration will also work for a transformer. In fact it would work for any push pull amp regardless if it is single or split supply and ultimately protects the speaker agains DC regardless of the fault.

Nico
 
Hi,

Single supply rail with zero thump using two series caps in the rail in stead of one. Why no thump, because capacitors are charged in series and juntion is at rail/2, there is no voltage developed over speaker.

Regards

Nico

Hi, First time that I have seen this arrangement and seems good. But for the proposed PSU I am going to use 30 pcs of 1200 uf/200v capacitors in parallel with a CLC form. so there will be no center tap point. Else, I would add 2 10,000 uf/75v in series then tap the speaker terminal on the center. Will this be okay?

With this will still need a DC protection?
 
Status
Not open for further replies.