"Compact" VFA

Status
Not open for further replies.
Jump to Latest
vzaichenko

vzaichenko

R.I.P.
Joined 2010
#1
Compact Amplifier - Singleton CFA

Here goes the final "as built" design and the latest measurements:
Compact CFA - Refined

I can't change attachments here - this is a very early draft, so see the attachments at the link above.

My power supply is +/-52V DC.

A rather simple, inexpensive, reliable, great sounding power amp.

Cheers,
Valery
 

Attachments

  • @Compact-VFA-01.jpg
    @Compact-VFA-01.jpg
    495 KB · Views: 1,648
  • Compact-VFA-PCB-in-progress-01.JPG
    Compact-VFA-PCB-in-progress-01.JPG
    220.4 KB · Views: 1,524
Last edited:
#4
vzaichenko said:
Hi All,

I was thinking about slowing down with new developments for a while and focus on building some earlier designs and refining some older PCBs.
However... a friend of mine asked for something rather compact yet high quality, utilizing the old slow MJL21193/21194 (I have also got some of these in stock).

Here is what I came up with. Nothing "revolutionary", mostly proven arrangements. Interesting for me (I've tried push-pull and single-ended with CCS load VAS topologies initially), this kind of balanced VAS with bootstrapped load showed an excellent simulated performance.

It's not live-tested yet (PCB design is in progress). Trying to squeeze the whole thing onto a 3"x5.5" inch board.
Power supply is +/-50V DC, quiescent current if 90mA per output pair.

It will also work with modern output pairs, like NJW3281/1302 of less powerful NJW0281/NJW0302, or more powerful Sanken MT-200 2SC2922/2SA1216 (although these ones will require certain PCB rearrangement due to their wider cases).
If somebody would like to experiment with alternative PCB designs, etc. - you are welcome.

Cheers,
Valery
Click to expand...

I have "live tested" on my last modular "supersym" (EF2).
Slight issues = Thump on startup ,
"un symmetric" performance - I used 2 diodes to drop the bootstrapped side rail voltage down as the bootstrap will swing higher than the active side , and the B-S resistors are rail dependent - I used about 10K between the 2 for 70V with 7ma VAS.
Had to drop to 6K to bench test at 40V.

Otherwise , the bootstrap actually sounded better than the CCS'ed version.
I like your CFP input /differential VAS scheme.

PS - did you read my mind ? .... A bootstrap was the next IPS I was planning. 😱

OS
 
#5
vzaichenko said:
Hi All,

I was thinking about slowing down with new developments for a while and focus on building some earlier designs and refining some older PCBs.
However... a friend of mine asked for something rather compact yet high quality, utilizing the old slow MJL21193/21194 (I have also got some of these in stock).

Here is what I came up with. Nothing "revolutionary", mostly proven arrangements. Interesting for me (I've tried push-pull and single-ended with CCS load VAS topologies initially), this kind of balanced VAS with bootstrapped load showed an excellent simulated performance.

It's not live-tested yet (PCB design is in progress). Trying to squeeze the whole thing onto a 3"x5.5" inch board.
Power supply is +/-50V DC, quiescent current if 90mA per output pair.

It will also work with modern output pairs, like NJW3281/1302 of less powerful NJW0281/NJW0302, or more powerful Sanken MT-200 2SC2922/2SA1216 (although these ones will require certain PCB rearrangement due to their wider cases).
If somebody would like to experiment with alternative PCB designs, etc. - you are welcome.

Cheers,
Valery
Click to expand...

Nice! I've got a couple boxes of MJ21195/6 steel can TO-3s waiting for a home. Looks like a good candidate.
 
#6
ostripper said:
I have "live tested" on my last modular "supersym" (EF2).
Slight issues = Thump on startup ,
"un symmetric" performance - I used 2 diodes to drop the bootstrapped side rail voltage down as the bootstrap will swing higher than the active side , and the B-S resistors are rail dependent - I used about 10K between the 2 for 70V with 7ma VAS.
Had to drop to 6K to bench test at 40V.

Otherwise , the bootstrap actually sounded better than the CCS'ed version.
I like your CFP input /differential VAS scheme.

PS - did you read my mind ? .... A bootstrap was the next IPS I was planning. 😱

OS
Click to expand...

Hi OS,

I see slightly asymmetric clipping here - exactly, as you say, bottom half swings a bit higher - however, clipping is rather soft, so I don't consider it as a big issue.

We think pretty close sometimes 😀

Cheers,
Valery
 
#7
PCB

Done!

Who wants gerbers? Please send me your e-mails.
Jeff - I know yours 😉

Q1, Q9 are expected to have a local heatsink.
Q15, Q16, Q17 go to the main heatsink.
Small output transistor cases are shown - MJL21193/21194 have got the bigger ones - anyway, the pinouts are the same, only the holes for them will go a bit wider...

Cheers,
Valery
 

Attachments

  • Compact-VFA-PCB-01.JPG
    Compact-VFA-PCB-01.JPG
    246.5 KB · Views: 1,128
  • @Compact-VFA-PCB-01-Silk.jpg
    @Compact-VFA-PCB-01-Silk.jpg
    208.8 KB · Views: 1,044
#9
borys said:
Nice simple amp.
I do not why but I like a lot simple design's.
The fine tune resistors ratio R20 and R29 in lower part of the VAS can give nice sound resaults.
Waiting for the pcb.

The input j-fets can be substytuted with BC546 ??


Regards
Peter
Click to expand...

Hi Peter,

jfets are much more preferable here - with their high input impedance they allow good balancing of the circuit, also this IPS topology gives better linearity and sounds good. So... I would recommend to leave it as is 😉
Unless you like to experiment 😛

Cheers,
Valery
 
#14
LSK170 are good ones, but in this particular design they appear to be at the edge in terms of maximum G-D voltage. 2SK117 will feel better here.
Well, I could add a cascode above the input pair, but I tried to keep this design minimalistic 😉
 
#16
Single-ended IPS/VAS CFA - same board size

Another idea came to my mind 🙂

This is a really cool one. I put it here because it sort of shares a similar OPS and PCB arrangements. However, in terms of IPS/VAS it's a completely different creature.
Single-ended, with high-current GNFB, coming to the low-impedance point (input jfet source).

Rather symmetric, clean clipping. Low distortion.
Low cost components. OPS may vary (even MOSFET option is possible).

Cheers,
Valery
 

Attachments

  • @Compact-CFA-01.jpg
    @Compact-CFA-01.jpg
    526.8 KB · Views: 600
  • Compact-CFA-PCB-01.JPG
    Compact-CFA-PCB-01.JPG
    282.9 KB · Views: 493
#17
vzaichenko said:
Done!

Who wants gerbers? Please send me your e-mails.
Jeff - I know yours 😉

Q1, Q9 are expected to have a local heatsink.
Q15, Q16, Q17 go to the main heatsink.
Small output transistor cases are shown - MJL21193/21194 have got the bigger ones - anyway, the pinouts are the same, only the holes for them will go a bit wider...

Cheers,
Valery
Click to expand...
PDF please😉
I have some 2sk146,can i use them?
Mary Cristmas:xmastree:
 
#20
PDFs

Hi Thimios and All,

Sorry for delay - just got back home 🙂
Here they are.

Cheers,
Valery
 

Attachments

Status
Not open for further replies.
Top Bottom