BIG FET BLOAK: new layout attempt and a few questions...

ORIGINAL TOPIC:
http://www.groupdiy.com/index.php?topic=5866.0

Hi, I allways wanted to give a try to this little DOA, while I never did until now.
Since Tamas, nor Fabio seem to appear on the forum anymore, I have sketched my own layout for it based on this schematic of the big fet bloak:
http://twin-x.com/groupdiy/displayimage.php?pos=-4692

here is the layout :
http://twin-x.com/groupdiy/displayimage.php?pos=-4691

and the traces for home-etching (single-sided):
http://twin-x.com/groupdiy/displayimage.php?pos=-4690

after several readings of the entire topic during the past years, I tried to make it so the biasing diodes are barely thermaly coupled to the power devices (read: close), and there is room behind the output transistors to mount these  on a shared "heatsink" (I plan to use a 1.5mm thickness plate or L shapped piece of aluminium for it).
Feel free to : analyse, criticise, and use this layout.

I plan tu use it in a classic 312-like circuit and I need your advices for this, because the "recommended circuit" drawn by Tamas is no longer aviable and some infos are still unclear to me.
I will use 10k for feedback R and a 10k C pot in serie with 100 ohms for Rgain. I'll try 10pf in parallele with Rf.

First question: what minimum value would you recommend for Cgain (cap in serie with Rgain to ground), and what minimum voltage rating.

Second one: do I NEED an output coupling capacitor from DOA out to 600:600 output transformer ?
and if so, what value would you recommend ?

Third one : for +/- 24v supply, what  current per channel should the PSU be able to safely provide ?

and fourth : I have a bunch of 1:10 input tx's, do you think they'll be too high a ratio in this circuit ?
(I remember Tamas recommended 1:3 to 1:7 for the task )

Thanks a lot for your comments and help.

Laurent.

i have like 6 of those so far only tried 2 and whatever the volts i use them on they got hot...really hot.

Ok. So, large heatsink....

Laurent.

couldn't you up r7/r8 a little?

Yes BIG heatsinks.
They sound great.
Coupling cap 1000-4700uF.
Same for Cgain.50V.
1:10 should be fine.
I think~ 35mA.

Hi Laurent - any luck with this so far?

Hi,
Still have to save time and collect a few parts. Hopefully in a few days or weeks this will move forward.
Anyway, thanks for the advices and interest.

Laurent.

d'accord! Allez-y !

I might have a few of these boards if I dig to the bottom of the closet and put up with the ole' lady screaming: NOOOOOOOO!!! All that s@#$@t has to go back into the boxes now!

Hi, Tamas, nice to see you were still lurking around here!
Any advice or memories related to my concerns in the first post ?
and what about the heat problem ?

Here is what I recall:
- When you do the darlington on the outputs you will need four diodes to do the biasing.
- DO NOT thermally couple the diodes to the output transistors! Makes things worse.
- With 22 ohm (1W) resistors on the emitter of the output transistors I never had to use a heatsink.
- 10pf to 22pf accross Rf was ususally enough to tame the amp.
- Yes, you will need an output coupling cap beause there is always a DC offset (200mV minimum).
- I would use at least a 220uF cap, but you can use larger ones if you need hear down to 5Hz.
- Calculate at least 40ma per BFB DOA.
- The idle current will be all over the place between 20ma and 40ma.
- If you use the 2SK170 on the front you can have any ratio you want 1:2 and up. That FET is very quiet.
- Cg should be 470uF and up, but it can be low voltage like 10V or 16V.

Thanks !
Now, excuse another stupid question : how should I calculate the value of a resistor to put in place of D7 for a given supply voltage (+/- 24v in my case) ?

Thank you for your help.

Laurent.

The Gainbloke v4 uses a 15k resistor instead of D7 when using 24v rails.

man,, i have around 20 of those guys here and never did anything with them.

I'll some searching and see if I can came up with a use for these guys.
anyone need some?

I just made some boards based on this design, but with my own custom layout (i.e. HAD to be single-sided to be able to CNC) and they turned out pretty well.
They are for 2 x 2-stage BigFetBloke microphone preamplifiers I'm building: