Simplistic NJFET RIAA

Member
Joined 2006
Paid Member
In this case we are using the fet as a ccs, not an amplifier.... so I believe we do not run into any nonlinearities.... the issue might be that we can have lower noise at higher Id, but if I choose a lower value for vref resistor, I will loose Zo in the shunt so it is a compromise.
 
In this case we are using the fet as a ccs, not an amplifier.... so I believe we do not run into any nonlinearities.... the issue might be that we can have lower noise at higher Id, but if I choose a lower value for vref resistor, I will loose Zo in the shunt so it is a compromise.

Sure - sorry, I was thinking about the JFETs used for the 1st & 2nd gain stages ("sitting underneath" the CCSes).

Regards,

Andy
 
I am using 30kr for vref so I need around 1mA to get 35Vout.

I chose a K117GR with 3.6mA Idss... the trimmer on the gate reduces this value to the correct Id.

Thank you Ricardo. Thus K117's Idss the lower the better, right? ;)

I missed the point: why did you lower R11 from 47k to 30k? If I'm not wrong you double up the value of C1 too.

P.S. trimmer on the source, not gate I guess
 
Last edited:
Member
Joined 2006
Paid Member
Thank you Ricardo. Thus K117's Idss the lower the better, right? ;)

I missed the point: why did you lower R11 from 47k to 30k? If I'm not wrong you double up the value of C1 too.

P.S. trimmer on the source, not gate I guess

Yes, I use low Idss k117 for this purpose.

I reduced R11 to get higher lower wondering in output voltage and it seems the higher current on the ccs fet reduced it's operating noise.... it works , but I am on the limit now... can not go lower than 30k, bypassed by a quality foil 10u cap... lowering the vref increases Zo

Yes... trimmer on the source...
 
Last edited:
Member
Joined 2006
Paid Member
The folded runs at around 80% in first stage.

I should report last mod.... 4 x k170 with 12r for Rs proved floppy in the bass and noisier tan the 2 x k170 build so I reduced it to 3 x k170 with 4ohm for Rs... R14 went to 1k and now I have 63dB instead of 65dB total gain....

Now playing with the BiB.... very happy with the results... full on sound, snappy, detailed, low noise....

PS: this build is very sensitive to magnetic fields from the TX.... had to place the tx at 90° and 1 meter away.... very rewarding results in terms of bass definition... induced hum can kill bass deinition
 
Member
Joined 2006
Paid Member
Back to the BiB

Once more Vref cap proves it's significance.... After listening for some time with a 330uF cap over the vref trimmer and finding that something could be done to improve highs, I replaced that HQ EL cap by a smaller value MKP... incredible difference in terms of detail... highs are more fluid and fast and that translates to better discernibility in the bass also.

Strangely, I do not notice any increase in noise by reducing this cap value from 330u to 4u.....
 
Member
Joined 2006
Paid Member
In fact there is less hiss with only one 4,7u film cap than with the 330u EL.... but there seems to be a stability issue because the shunt sags when I touch the heatsink and also when I rotate the preamp volume to the max.

Now with two 4,7u in parallel, the sagging disapeared and sound is even better.... more detail and better spatiality... the problem is the hiss came back.... it seems the big cap in this place is prone to collect noise (like an antena)
 
Member
Joined 2006
Paid Member
Yes... maybe that is the way to go... anyway I learned that too small capacitance over resistive vref can compromise output current capabilities... because it sags when I touch the heatsink (the heat sink is totally isolated so it must be a magnetic happening).

It puzzles me that it also sags if I turn the volume up.

This volume thing is quite strange... untill 90% all is ok and the noise is really low, that it sags at 95% and than it wakes up again at 100% but with a lot of noise.

Maybe increasing the ccs current might cure it.... but it allready runs at 300mA.

With two 4,7u there is no problem but noise is much lowder than with the 4,7u alone.