The diyAudio Firstwatt F6

6L6

Moderator
Joined 2010
Paid Member
After a quick blessing from our technical, spiritual, and menu advisor, Nelson Pass, I am glad to present the preliminary discution for the diyAudio Mosfet F6.

F6_DIY_SCH.gif


Cool, huh? :) This is made for IRFP420, or a similar Mosfet that has a Vgs in the 4V range. This is a beautifully simple amplifier!

F6_Super_long.png


Member Didiet78 has done some preliminary layout, (Thanks a million! It looks wonderful!!) this is the current thinking - 2 PCB, a left and a right, in order to keep the input transformer as far back as possible on both sides, and to have the Mosfet mounted low on the heatsink. The PCB will be designed to only mount the specified Jensen transformer, but I'll look into the possibility of a second set of pads, if possible, to mount a Lundhal or something... No promises.


The diyAudio store is also looking into stocking the transformers, which would make the whole process quite easy.

The other consideration in this project is the PSU, by which a shielded transformer is going to be very necessary, mounted as far from the input transformer as possible, and possible in a screening can as well. Or, of course, a separate box for transformer, diodes and first capacitor bank. Other than that, everything is straightforward. The diode, cap and resistor bank are the 'Firstwatt vernacular', I.E., look at any Firstwatt amp and use that PSU. The diyAudio PSUv3 boards will work beautifully, as will many other PCB.

The floor is open for discussion. :D
 
Last edited:

6L6

Moderator
Joined 2010
Paid Member
Factory_F6.jpg


This is a photo of the factory F6, the layout should imitate this as best as possible - you can see the input transformers way at the back, and the power transformer up front.

Of course the trick will be using the 4U deluxe chassis, it's only 300mm deep, so the power transformer will be much closer. Again, use a shielded toroid, and perhaps a can as well.
 

6L6

Moderator
Joined 2010
Paid Member
The far transistor needs to be over the centerline of the 5U case's heatsink - it is a two-peice extrusion, and this way each Mosfet will be on it's own heatsink.

The 4U is a single extrusion, so it being off-center means little.

A bit closer to the transformer is not a bad idea however... :)
 

PKI

Member
Joined 2011
Paid Member
I don't like the absence of a big ground trace on the board. Personally, I prefer taking speakers output from the board.

BTW, has Nelson mention how it measures compare to the first F6 with SS in the output. I know that subjectively people preferred Mosfet version better, but what about distortions?
 
6L6. by inner transistor, i'm thinking of the transistor closest to the transformer. move it 40mm closer to the transformer so they are 120mm apart from each other. maybe that wont work with UMS with the 400mm chassis? on the other hand. 400mm sinks will be in-efficiant anyway, as the baseplate is not thick enough to transfer the heat from only 2 devices placed so close together.
 
After a quick blessing from our technical, spiritual, and menu advisor, Nelson Pass, I am glad to present the preliminary discution for the diyAudio Mosfet F6.

The floor is open for discussion. :D

It appears that the two zeners will be starved, and operated in their soft and most noisy region. There is 1.8 mA through the 10k resistor feeding the zeners. They will have about 0.8 mA each. The other 1 mA will be diverted through the bias pot.

Usually you want a minimum of at least 5% or more of the max zener current. For a 400 mW zener, this would be about 4 mA, about 5 times more than the actual 0.8 mA here

Also, how about a negative supply pilot led also?
 
The one and only
Joined 2001
Paid Member
It appears that the two zeners will be starved, and operated in their soft and most noisy region. There is 1.8 mA through the 10k resistor feeding the zeners. They will have about 0.8 mA each. The other 1 mA will be diverted through the bias pot.

You are welcome to put more current through them, but any noise is not going to make it
through the 10K + 1,000 uF RC filter, and if it did it would be taken down another 20 dB
or so by feedback.

:cool: