My new design or updated old one

[Mooly]

What I would now is print the layout to scale and see how it looks.

Some of the large pads (if it were me) have a bit to much "hole". I'd make the hole smaller while keeping the same pad size. When you come to solder you want the component legs to be as near to the copper as possible. If you are drilling by hand then I find 1mm or even 0.8mm holes a good size for the PCB. The hole in the pad is just needed to start the PCB.

[astankov]

Big holes on the pads are because of my software. I can't change the hole diameter without changing the the pad size. I also see that the holes are too big. I use ExpressPCB - any idea how to fix that?

[Mooly]

Sorry no, I use Diptrace where you right click a pad and alter its properties.

[astankov]

Yes today was the actual build but schematic was different. After an not very successful build with the EF buffered VAS I decided to give a try a cascoded VAS. Also I moved the trimmer for the DC offset and OS is Quasi now. So I build the schematic on the picture on veroboard , powered up and listened. All connections from the board to the drivers and output transistors was with long wires and incredibly there was no any hum and buzz. Sound was not bad but I tested one channel only and dont have full impressions.

Now several questions : I will use this schematic with +/- 50 V supply. BD139/140 cant be used so with what to substitute them? I dont like MJE340/350!
Will one pair IRFP260N be enough to use safe with 4R load?

Any other comments on the schematic are welcome.

Best Regards!

P.S. Here in Bulgaria we have a really bad snow now. I wasnt seen something like this from years!

PS2: Oops I attached a wrong picture! Actually the only difference with real schematic is that there is no 100R resistors in the driver bases also 10pF in NFB is not absolutely necessary no problem without it - test was without.

[Mooly]

For real world domestic use I think a single pair is OK, even on -/+50 volts. There is a world of difference in sustained high power testing into 4 or 8 ohm.


If you really want to drive it as a lab amp into 4Ω or 8Ω at maximum output then multiple pairs are advised.

Single pairs are often sonically better too... the capacitive loading is that much less.

Drivers... hmmm depends what you can get. How about the 2SA1209/2SC2911 or second choice, the 2SA1668/2SC4382.

Just a typo on C10. T'Other way up
Output inductor ? You never know what someone might hang on the end of it.

[astankov]

Thanks for the comment , Mooly!

Yes C10 drawn in wrong way. I can get 2SC4793 and his brother are they suitable for VAS and drivers? Output inductor is also good to have it but how to make it?

Is there something in the schematic that can be made better or is OK like I make it?
I'm asking because still testing the amp and if someone can tell me what can be made better will test and post the result.

Best regards!

[Mooly]

The 2SC4793/2SA1837 should be good.

Output inductors are pretty non critical in absolute terms.

This is quoting D Self... "air cored... around 6uH... 20turns of 1.5mm copper wire... around 1 inch diameter". You can also add a 10 ohm carbon 1 watt resistor across the coil for damping.

It's very non critical. It goes after R25/C13 of course.

You could also try (on listening tests) adding small series output resistor of around 0.1 to 0.33 ohms. By raising the output impedance the speaker impedance comes into play modifying the overall response. Some like it, some don't. You must listen to it carefully and critically.

Edit... picture of output inductors from my amp. I wound these round an old electro cap as a former.
Output Relays

[astankov]

After listening tests I increased the driver stage current to near 50mA. It gives a much better detail and HF response also added the new values for the higher supply voltage , output coil , gate protection zeners and I made better more understandable drawing.

So any comments?

Regards!

PS: Why when I put an emoticon in the text after submitting post it disappears? Is this happening only with me?

[Mooly]

typed out and and as emoticons.

The increased driver current will certainly make for better hf perfomance.

It would be an interesting experiment to compare the HEXFET's with Laterals sonically in an amplifier of this topology.

Edit...... emoticons.... yep, just you

[astankov]

another icons test.

Will be interesting for me too to compare V-FETs with L-FETs .

What browser do you use? I use Firefox latest version.

Edit.....now have emoticons but not all one is missing.