CF-FET V2.0 front-end - going high-tech (SMD)

Status
Not open for further replies.
Hi Russell,

I use the German flux paste - see the picture. Something pretty similar in Mouser stock:
>NO CLEAN FLUX PASTE<

With regards to the FETs - good point. I think, I will have to remove a couple of the output plugs to be able to use some lower angle while soldering them, just to have more "massive" contact with the iron. I use a regulated-temperature iron, so normally, if I slightly increase it - it's enough for soldering such "areas".

Cheers,
Valery

Here's one with solder in it. It works good with an iron or in an oven.

SMD291AX10 Chip Quik Inc | SMD291AX10-ND | DigiKey
 
Hi Sam, it's possible to modify the OPS for N-channel only - I will drop the schematic update - but it will slightly suffer from additional distortion - not too much though, as the front-end is a relatively high loop gain one.

Cheers,
Valery

Thank you, Valery. This news sounds great, and if the increased distortion is low order or even harmonics, I am sure the amp will sound fantastic.
 
Done, tested!

OK. Soldered the FETs, pluged in, measured, listened - perfect!
Both channels are exactly the same. Surprise - no output "plop" at power on. Practically zero offset (some millivolts) right from the start. VAS FETs are very lightly warm.
Excellent :up: Great experience 😛

Guys, who are interested in building those - green light now 😉

No energy for Shuntie tonight - will solder them in the nearest days.

Cheers,
Valery
 

Attachments

  • DSC03849.JPG
    DSC03849.JPG
    306.3 KB · Views: 385
  • DSC03853.JPG
    DSC03853.JPG
    370.5 KB · Views: 371
Just noticed an interesting thing - didn't really notice it yesterday.

VAS runs at the higher quiescent current, than the one in TH version.
This one runs at 28mA now. Both channels. Well, fine with me, I don't even want to decrease it, OPS bias pot has got enough range to adapt to that and with such current it drives the OPS MOSFETS perfectly - their high input capacitance does not really matter any more.

But why? All the values are exactly the same in TH and SMT versions respectively. Must be some differences in Vbe / Vgs between the TH and SMT versions of the transistors.

Well, Nelson Pass runs those fets at 40mA. No problem. But I did not expect an increase for more than twice as much, comparing to the initial TH version...

Plays beautifully anyway 😉
 
Just noticed an interesting thing - didn't really notice it yesterday.

VAS runs at the higher quiescent current, than the one in TH version.
This one runs at 28mA now. Both channels. Well, fine with me, I don't even want to decrease it, OPS bias pot has got enough range to adapt to that and with such current it drives the OPS MOSFETS perfectly - their high input capacitance does not really matter any more.

But why? All the values are exactly the same in TH and SMT versions respectively. Must be some differences in Vbe / Vgs between the TH and SMT versions of the transistors.

Well, Nelson Pass runs those fets at 40mA. No problem. But I did not expect an increase for more than twice as much, comparing to the initial TH version...

Plays beautifully anyway 😉
Nice work Valery!
 
Thank you, Valery. I will start laying out the PCB. I believe one can play with the value of R65 and even parallel a Diode and Cap for optimal performance. Is only Q19 to be mounted on one of the Output MOSFETs and is LED5 a Red one? Thanks again.

Yes, the LED is the red one. Q19 is mounted on the local heatsink with the drivers (Q30, Q31), Q23 is mounted on the main heatsink with all the MOSFETs. It's not necessary to mount it right on one of the MOSFETs - just touch the heatsink. I recommend to use the original Slewmaster layout as a basis for your PCB development - output section's arrangement is good there. Just modify the fragments that have changed (this is what I would do).
You can play with the value of R65 and even parallel a Diode and Cap for optimal performance, however it simulates pretty well even as is.

Cheers,
Valery
 
Status
Not open for further replies.