Simplistic NJFET RIAA

Member
Joined 2006
Paid Member
I found higher Vgs in my latest batch of BF245A the other day so rule it out, stick to K117GR as an alternative. Its the different tempco curves in various types that make them behave differently for drift under certain VDS and current bias.

Just finished k117gr mod on V12r vref.... much faster catch up indeed.... almost instant to 34.5v.... I wonder how you know all this stuff.... You seem to be able to pick the best parts for their function but also for their sonic posture with ease. Do you pick those by reading the datasheets or what ?

I ask this because in my case, I can only pick the parts after experimenting a lot ... after I know what a part sounds like, I can predict it's effect on a future build but this requires a lot of work and experience.... are you that experienced ?
 
Last edited:
Member
Joined 2006
Paid Member
What is Rg (signal source resistance) in the datasheets ?

Noise is comparable between k170 and k117 in the vref, but I see noise starts to rise with Drain currents below 1mA in the case of k170...

Soundwise, k117 seems to behave better at low freq but lacks the "fantasmagoric" high freq presentation of the k170....

Maybe I could use a low idss k170 with higher drain current (2mA) but in that case I would need a 17k resistor instead of a 30k for vref... what should be the backside of this choice ? How low can I get before Zo starts to rise and be noticed in the very lows ?
 
diyAudio Chief Moderator
Joined 2002
Paid Member
K170GR is even worse for drift there. "Fantasmagoric" can be a bit of thermal modulation artifact or its more capacitance modulation harmonics. Better stay with the faster stabler Vref FET and compensate elsewhere for subjective stuff. Like in Riaa curve and various AC couplers. Or in the reg parts itself.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
For that series they don't show but in other more proper for CCS fets they show tempco zero cross curves like attached (J201). Alas most such fets are much more demanding than the Vbe only VDS we use them in those regs plus they are much more noisier usually. I could see a 1.2R making an LMC losing 10dB noise floor with a J201 for instance.
 

Attachments

  • Snap 2013-04-05 at 00.15.51.png
    Snap 2013-04-05 at 00.15.51.png
    112.4 KB · Views: 285
Member
Joined 2006
Paid Member
So that is wha i need now.... I felt the 4 x k170 lacked bass grip and is noisier than the 2 x k170 build.

Inittially I wondered if that noise came from the 12ohm Rs used... I was now recalculating the folded ccs resistor to lower these Rs to 3.9ohm... but now I know better... where can I get some good 389 ... ampslab ?