Heatsinks for Bosoz?

Status
Not open for further replies.
Jump to Latest
#83
Kari,

I have used this sort of design for my first prototype, but as soon as I
had need to replace one of the mosfet specialy the balanced pair, I was in trouble ; even to access the screw it's difficult... and to reinstall the new mosfet even harder...

You could find a way to arrange the diff pair to be more accessible:
by changing the heatsinks or with a little move ...
put the diff mosfet pair lying flat on the circuits board...
moved them to the lower part, so you can place a hole {in the other heatsink} to access the screw...
place the CCS under the pcb...

Better use a pref board and have a try, I think you will see what happens.

Regards.
Alain.
 
#84
Thx for your thoughts. Still i think if you lift the heatsink with the fet still attached to it, this should not be a problem.

I could ofcourse turn the fets 90degress and move the ccs down but then the board size is going to get a lot bigger. I'll play around with some ideas on a protoboard, but as it is know i'm not going to change the layout.

Thx,

Kari
 
#85
It remains one problem on your last design, if you want to replace the central or left mosfet ; you have to remove the two others and heatsinks to do that ; a best way would be to give better access to each mosfet to be able to replace them... {I have changed twice since installation}
Click to expand...
I dont see that as a big problem. You can unsolder the hole unit including the heatsink and pull it out, mount the new mosfet and solder the unit back in, or am I missing something there?
The issue is propably the alignment of the fet?
But ofcourse, it would be easier to be able to access the Fets directly.
Just out of curiosity, what happened to the fets you changed? I didn't manage to fry any yet.
To get another thing clear, I guess the diff pair needs matching, right?

Steen🙂
 
#86
steenoe,

I have changed about 4 pairs, even matched at 0.01 volts
there are sensible differences...
Starting with series of matched MosFets IRF 610 same lot code
and same 0.01 volts values there are sonic differences
so I have unsoldered them a lot of times... and installed the heatsinks
with screws not with solder to go quicker!
I am just trying to help here.

A+

Alain.
 
#88
steenoe,

Better buy 25 minimum, if you match them within 0,01 volts
there are differences in transconductance between the pairs
that you choose ; have fun.
For my part I bought a tube of 50 since the A75 and used the unmatched for tests prototypes, for CCS's etc.
If you can get series of 4 to 6 well matched it's fun to change only one
in the diff pair to get best sonic pairs!!!
{also have same gain, in the right and left side of the diff. pair}
I didn't blow any of them, but some already used and not of higher sonic
quality have been trashed.
The protos have worked in the 60 to 125 degree range, depending of the
CCS negative voltage ; and with bias as high as 80/100mA each...

Regards.

Alain.
 
#89
If you can get series of 4 to 6 well matched it's fun to change only one
Click to expand...
Alain, from this I understand you go to the extremes to get the best sound🙂 I propably wouldnt have taken it that far, but now you did plant the seed😀 Come to think of it, I remember Nelson once mentioned the special synergy that can happen with a particular set of Fets, I suppose thats what you are talking about? I did consider that unacheivable for a layman like myself😱 I think there is 50 610's in a tube, right?
I am not confident, that I could do something like that anyway:xeye:

Steen🙂
 
#90
question on input/output locations on bosoz pcb

I noticed on Kari's bosoz pcb that the input and output traces
are close to together because the holes for the input and output
leads are near each other. Will this cause any problems with
feedback oscillations? Shouldn't they be further apart?

One reason why I ask is that when I recently built my aleph 5
I had some distortion because my input and output leads
were too close.

John
 
#91
It shoudn't be a problem.

What about the idea of adding cascoding as metalman did? This could be done
pretty easily and if you dont want to "cascode" you'd simply bypass it with a jumpwire.

Any thoughts?

The boards are 99% ready, i just need to do some final cosmetic work to the silkscreen.
 
#92
And could look something like this.......
 

Attachments

  • boz2.png
    boz2.png
    32.2 KB · Views: 1,271
#95
Have been watching this thread without comment

But now I have to jump in to say

very nice job!

it could also be made single layer (with only two jumpers)

could you post a schematic ?

I might consider trying a proto using printer transfer film.

but ha who am I kidding I allready have an aleph 30 , a

1541 dac and a TQMT speaker project all underway.
 
#99
Update:

While my mother in law desided to come for a visit i desided to go for a business trip so i haven't updated the board until now.

I still need to figure out the advantages of cascoding so i'm going forward with the design i started with. The sinks are now turned 90degress, i've added a 22uf cap to reduce noise from the zener and added a groundplane on the toplayer.

I'll have a look at the pwrsupply this weekend too.

I should get some estimates on pricing beginning of next week.

/Kari
 

Attachments

  • boz_v2.png
    boz_v2.png
    31.9 KB · Views: 1,057
Status
Not open for further replies.
Top Bottom