CFH7 Amp

Member
Joined 2014
Paid Member
I never said they are genuine but they work. Also, some humble workhorse actives aren't going to be faked. BC547/557, BC550/560, J110, J310, 2SC1845 etc. I measure my caps with LCR meter and they are within spec. Always on the low end so 2200uF is really 2000uF.
What's the point of wasting time and efforts on building amps with inferior
parts? Price wise, it's cheaper to buy them from Best Buy.
It's about building amps that are better than that, and this can't be done
with fake parts from Aliexpress...
 
Founder of XSA-Labs
Joined 2012
Paid Member
What's the point of wasting time and efforts on building amps with inferior
parts? Price wise, it's cheaper to buy them from Best Buy.
It's about building amps that are better than that, and this can't be done
with fake parts from Aliexpress...

Not going to argue about this. Just letting people know there are options. I have about a dozen amps that would tell you otherwise about how good Chinese sourced parts can sound. Many components actually originate from China (according to my genuine parts compliance statement from Digikey) so doesn't surprise me.

You can look at my amps (more than a dozen and counting) and even listen to them - most made with AliExpress sourced parts. Some parts I got from Digikey, but I found no difference in sound quality. And I am quite certain they sound much much better than any amp from BestBuy :)

http://www.diyaudio.com/forums/solid-state/295286-virtual-audition-very-simple-quasi-mosfet-amp.html

580661d1479476675-virtual-audition-very-simple-quasi-mosfet-amp-ss-amps-2016-review2.jpg


Here is the Aksa/Ranchu quasi amp in Dacz layout. A beautiful sounding amp and 100% AliExpress sourced (except for heatsink (Dell Pentium CPU cooler and PCB from PCBway.com, also in China).

575037d1476638634-very-simple-quasi-complimentary-mosfet-amplifier-img_5410.jpg
 
Last edited:
CFH9

Now i can't believe that is something to do with oscillations .
100nf parallel to 2200uf removed.
Capacitors parallel to feedback electrolytics removed.
One pair of output fets removed.
A separate 100uf//47nf for each fet installed directly on fet pin.
Zobel removed.
Rails has been traced again,using analog and digital scope.
32 KHz square wave 15v RMS ,i cant see something wrong:confused:
Then putting the scope between power GND on board and signal GND...clear visible sinusoidal.Exactly what the input signal is but at a lower level.It is 15mV when output is 2V.
Pictures,1,2,3 32KHz
>> 4,5 The rails
>> 6.7 The rails
Picture 8, 15V RMS 32KHz
Picture 9 The probe between POWER GND & SIGNAL GND on the board.
 

Attachments

  • DSC08649.jpg
    DSC08649.jpg
    297.6 KB · Views: 226
  • DSC08651.jpg
    DSC08651.jpg
    307.1 KB · Views: 222
  • DSC08654.jpg
    DSC08654.jpg
    290.4 KB · Views: 218
  • DSC08656.jpg
    DSC08656.jpg
    312.6 KB · Views: 47
  • DSC08665.jpg
    DSC08665.jpg
    309.1 KB · Views: 51
  • DSC08676.jpg
    DSC08676.jpg
    244.1 KB · Views: 59
  • DSC08671.jpg
    DSC08671.jpg
    234.3 KB · Views: 51
  • DSC08670.jpg
    DSC08670.jpg
    236.4 KB · Views: 44
  • DSC08667.jpg
    DSC08667.jpg
    236.9 KB · Views: 50
Last edited:
Now i can't believe that is something to do with oscillations .
100nf parallel to 2200uf removed.
Capacitors parallel to feedback electrolytics removed.
One pair of output fets removed.
A separate 100uf//47nf for each fet installed directly on fet pin.
Zobel removed.
Rails has been traced again,using analog and digital scope.
32 KHz square wave 15v RMS ,i cant see something wrong:confused:
Then putting the scope between power GND on board and signal GND...clear visible sinusoidal.Exactly what the input signal is but at a lower level.It is 15mV when output is 2V.
Pictures,1,2,3 32KHz
>> 4,5 The rails
>> 6.7 The rails
Picture 8, 15V RMS 32KHz
Picture 9 The probe between POWER GND & SIGNAL GND on the board.

Congrats! It's clean. :up:

But if you are shorting the 10R lift resistor then the sinusoidal shouldn't be there. There should be 0V if the signal ground header is connected to ground.
 
Now i can't believe that is something to do with oscillations .
100nf parallel to 2200uf removed.
Capacitors parallel to feedback electrolytics removed.
One pair of output fets removed.
A separate 100uf//47nf for each fet installed directly on fet pin.
Zobel removed.
Rails has been traced again,using analog and digital scope.
32 KHz square wave 15v RMS ,i cant see something wrong:confused:
Then putting the scope between power GND on board and signal GND...clear visible sinusoidal.Exactly what the input signal is but at a lower level.It is 15mV when output is 2V.
Pictures,1,2,3 32KHz
>> 4,5 The rails
>> 6.7 The rails
Picture 8, 15V RMS 32KHz
Picture 9 The probe between POWER GND & SIGNAL GND on the board.

Only after you have debugged problem I remember.
Borys made good suggestion, how to lift input gnd on CFA amps in 100W Ultimate Fidelity Amplifier thread.
 

Attachments

  • open inpur str gnd.PNG
    open inpur str gnd.PNG
    105.2 KB · Views: 223
  • open input gnd together on board.PNG
    open input gnd together on board.PNG
    106.9 KB · Views: 217
  • shorted in  star gnd connection.PNG
    shorted in star gnd connection.PNG
    106.2 KB · Views: 207
  • shorted inp together on board gnd.PNG
    shorted inp together on board gnd.PNG
    105.5 KB · Views: 204
Last edited: