Simplistic NJFET RIAA

[massimo]

Originally Posted by RCruz

Hi massimo

Can not see the mod... would you care to explain in plain words ?

Quite easy..... Can you see the large cut? This drawing from the French mag L'Audiophile explain it better.

Pic A (without dotted line) is the std body, pic C the mod I performed

[Salas]

Originally Posted by massimo

I already extended the pcb to host the 4 input critters
It may become useful in a while.

I am going to use a common Rs and high B+ when I will do that. For the time being you can either make pads for a 2200uF capacitor parallel to R2 so you can have what I have now, namely the standard MC version with bypassed R2 that will go back to 56dB for circa 0.5mV MC by just disconnecting the cap, you can even put a mini pcb switch there to its ground leg, or gear up for the following in the pics but for low MC only. Needs some moves with some values to go lower gain. No instant switching capability or a little desoldering for two components. But better noise.

[massimo]

Oh, oh.....Q6 now 2N5459: different pin out
Than R9 added, C1 much bigger (if Silmic, it will be huge).... a deeper revision needed.
R3 no longer 43k? 47k as in the MM version?
R2: if you use a common resistor for the 4 critters, they must be absolutely identical (difficult to match) while a splitted resistor allows for some imbalances. Do you think it's a real advantage?

Shunt: back to v.1? no v1.1?
Q3: BC550? No MJE340?
C4 reduced value?

[Salas]

2N5459 looks better for Vt. *But we can't take advantage since it is 25V device, so 2SK170 in the end. 39V there. On TO-92 sink.
R3. Different Rload before it, must compensate.
R2. Syn08 has shown that it works with even relaxed matching in his HPS3. You can use 4x1R before they meet R2 so they can share current even easier if in doubt, since you make PCB. They can always be jumpered if it ends up the same.
R9. Helps to avoid any noise by feeding the cap in the cascode base, especially with relatively low impedance leds, and low Vp 2SK.
BC550 q3 in shunt higher hfe, if using sink it can take it. C4 220uF terminates OK with 9140 and 46V.

I saw difficulty in simulation for 1.1 at such a voltage. So I can't say do if I don't first.

All I know for sure right now is that the standard 56dB works with the 103R with bypased R2 at 36V V1.1 with 8.2k buffer's emitter load. And that is all I can surely recommend for now.

[Salas]

Originally Posted by nicoch46

Salas
searching around the web I see people use 140 to 180r to 500r to 840r

burn in min 15h

After some little run in I tend to settle around 150R now it has opened up a bit. Sounds heftier around that load. Its system dependant, I would say 150-270R should be the play area for those with a 103R. VTA is best with the cart's body parallel to the record's surface and only a tiny bit nose down. Almost a hair. Here it is in 1/48 octave resolution.

[Tham]

I change Rload to 100r and have 2200uF R2 bypass for my 0.2mv 2 ohm Supex 909. With little burn in time, I hear better clarity, firmer bass and better seperation/layering . It may change with more run in though.

[Salas]

What 2200uF? I like Nichicon FG, Chemicon KY, KZH. For industrial caps.

P.S. Did you fix shunt V1.5?

[Tham]

I had 16v Pana FM there...those are what I have in my parts bin.

I replace almost all jfets/bjt and still could not revive it. It must have been one or both burnt mosfets. One other think that I realised is that when it failed, it passed Vin as Vout to the load. This could potentially damage the load because of over voltage. Maybe a over voltage protection is required to be implemented.

[Salas]

Pana FM very good. Maybe your mids fleshed out too, now.

For V1.5 I don't know details since Iko did that version, maybe you can ask his help on the relevant thread...Or make directly his V2 R5?

[Tham]

The mid certainly fleshed out more now because of optimum loading of the Supex .

Hope to build the second one soon with better parts (the 4 // 2sk170 version).