DIY Sony VFETs OS2 Official thread

Jump to Latest
#141
R1+R2 total 0.94R, and drop 1.2V, so I=1.2V/0.94R=1.28A

19Vdc across 2SK60 so P=19Vx1.28A=24.3W

16V-1.2V=14.8V across IRF250 so P=14.8Vx1.28A=18.9W

By the way, an infrared thermometer can be unreliable for measurements on shiny, reflective surfaces.
 
#142
So, the MOSFET and the VFET are close. I suppose I will construct a channel and place one channel on the heat sink and see how it does, then the second channel. One heat sink if possible, two if necessary. This would be on the bench before assembling in the chassis.

Thank you for the calculations. I think there is the 'crikey' finger thermometer, too. I hope NP hasn't patented his finger thermometer.
 
Last edited:
#143
As a quick guess, I'd expect your 4650 heatsink will get very hot quickly with just one channel on it. Two channels is likely out of the question without some forced movement of air. A slow moving PC fan might work in the chassis. Tread carefully.
 
#145
Well, all of these heat sinks in the Sony Vfet amps are the same through the TA 8550 series. In the 8550, the heat sink holds six VFET's operating at about 100 watts relatively high bias class AB, which should mean that it should disperse up to 100-200W or so of heat, so 90 watts or so could be within spec. for one. The idle power draw of the TAN 8550 is 260 watts into two heat sinks, so I would speculate that a good bit of that dissipation would be in the heat sinks. Even the power consumption of the TA 4650 is 270 watts. There was one article in the day that measured the 4650 at 60 w RMS per channel with about 1 percent THD, so it was like many of the high end products in those days underspecified. I know there are issues of rate of heat flow and dispersion etc. but it is worth a bench test and I will report if I catch on fire or blow myself up in the process. I would give it 50/50 that it is OK with one heat sink and certainly OK with two, and maybe 3/10 that I complete this project and get something working, but I hope it will be fun and interesting along the way.
 
Last edited:
#147
I have totally confused myself now.

I got the “N” channel set from the third round : boards say OS2
I have 2SJ28 VFETs.

1) should I use the resistor values from third round OS2 schematic (for another VFET)
2) can I (should I) use the resistor values from the earlier round N channel VFET
3) does it matter?

Capacitors look to be the same (bypass caps a bit different, but I don’t think importantly)

The chassis will be the last part to get here, so I have time. Stuffing Bulark boards tomorrow. Maybe dreadnoughts next weekend. I’ll build the OS2 boards when I get the schematic sorted.
 
#149
Thanks, my dyslexia can be hazardous. And juggling all the other stuff for this build.

Just confirmed that my OS2 boards are “P”, matching the VFETs I put in that box.

Just replace “N VFET” with ”P VFET“ above please.

I ordered a set of “N VFET” boards when Jason announced that there were kits available after the initial surge. I plan to build those into a bigger chassis, with dual PSUs, and Mark Johnson’s balanced input cards.
 
#150
DIY Sony VFET pt 1 (2SJ28) was single ended follower. OS2 is optimized for the lower power 2SJ18 by utilizing a mu-follower complementing a follower (MUFF), which gets a bit of output power from the mosfet, to raise the output power level to about equal that of the 2SJ28.

You can run the 2SJ28 with the existing OS2 parts and have the 2SJ28 as a MUFF.

Or since you have the 2SJ28, you can run it in pure single ended mode, as pt 1, and take advantage of its higher power capability. To do that jumper R2 on the board with a wire, and place two 1.5 Ohm 3W in parallel in R1 position. All other OS2 parts remain.
 
  • Like
Reactions: Bones13
#155
I am planning to build the P channel into one of the special VFet chassis, when they come out. I am also planning to build the N channel into a 4U chassis with dual PS, and balanced input cards. I would be happy to swap out some VFets for the boards 😎.
 
You must log in or register to reply here.
Top Bottom