Simplistic NJFET RIAA

diyAudio Chief Moderator
Joined 2002
Paid Member

Attachments

  • salasriaa2sk170bl.jpg
    salasriaa2sk170bl.jpg
    76.8 KB · Views: 52,532
www.hifisonix.com
Joined 2003
Paid Member
Salas, if you want something simple and reliable with adequate performance, just use a good quality op-amp like a 5534.

If you want to use JFETS to improve your MC/MM listening expereince, I don't think this circuit is the best one - there are better JFET circuits - see JC-1, Dennis Collins et al.

Good luck
 
There's already some stuff using that concept, L'Pacific RIAA, since I cannot find a schematic for this one, I link you to Mad_K's variant of it

http://www.diyaudio.com/forums/showthread.php?threadid=28223

Have fun, Hannes

EDIT: oh and I almost forgot: the Pearl by Wayne Colburn (Pass Labs) is another example of a 2 stage Njfet-riaa pre.

EDIT2: bonsai, I cannot find a schematic for the Dennis Collins'pre - would you please point me to one? Thanks!
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Bonsai said:
Salas, if you want something simple and reliable with adequate performance, just use a good quality op-amp like a 5534. Good luck

Had 5534 2 stage with passive RIAA for many years when I was doing solid state. Very effective. Had original Signetics and synergistically sounding caps and resistors. Best I have heard is my mate's Steve Bench modded 12AY7 parallel & ECC88 cascoded. I just have 2SK170BL left overs from my recent DC coupled Pass B1 version and I am monkeying around a bit. :)
 
diyAudio Chief Moderator
Joined 2002
Paid Member
I saw in that link that some put the first stage to second stage coupling capacitor after the passive RIAA and say they got better subjective performance. I guess because they use mainly small ones and the response is linear after the RIAA stressing less. I will have to listen for that position exchange. They also had several other circuits made with highly endorsed op amps etc. but some prefer the ''Pacific'' class for better subjective musicality. I guess that my experiment to be, falls in that class. Interesting. Some cascoding mods that I saw are also good to test. I did not see Rsource bypassing for more gain as I do aspire though. Maybe sounds bad, or it is noisy. I have some Dynavector 20dB silver step up if things prove difficult. Nice pass time concept non the less.
 
I use a mad_k with ska optvol ,good sound for me

I will try at first stage 170bl with 33r to ground for lower the second armonic dist.

at 24v there is a possibility of battery and use small series caps...what I found that this simple phone are very sensitive to quality component for now I use :
all R dale rn55 non magnetic
riaa caps vishay MKP1837
for series caps c3 66nf I use clarity caps Sa 0,1mf
outcaps 1mf Sa //mundorf 0,47mf silver-oil without r parrallel 100k
psu caps near jfet mundorf 22000mf

I will order more exotic part and report

this circuit are not so good for mm because have higth input capacitance near 300pf and no more 3-4mv at input ,Iuse denon dl-160 mc 1,8 mv
 
diyAudio Chief Moderator
Joined 2002
Paid Member
kevinkr said:
Take a look at Richard Sears JFET phono stage design, works well and sounds pretty good.

http://www.triodeguy.com/other_projects.htm

Note that you will need to match jfets regardless of what you decide to build.. Their transconductance varies all over the place.


Now that's a decent one. SRPP to drive the properly impedance sized RIAA, BJT follower for output. Kevin you dropped in here to remind me of glass...:smash:
How does it compare to decent vacuum?
 
diyAudio Chief Moderator
Joined 2002
Paid Member
nicoch46 said:
I use a mad_k with ska optvol ,good sound for me

I will try at first stage 170bl with 33r to ground for lower the second armonic dist.....
I will order more exotic part and report
this circuit are not so good for mm because have higth input capacitance near 300pf and no more 3-4mv at input ,Iuse denon dl-160 mc 1,8 mv

In my design to experiment with I included Rs trimmers and switchable cap bypasses so to listen for best sonic balance and gain, looks it will be practical since you mention degeneration.

Hmmm no feedback circuits can cost a fortune if looking for best parts to cook. I would do that in best tube RIAA, I don't know if the ''Pacific'' class is the ultimate circuit to throw big bucks in. Have to listen to a rather generic one with Wima caps first, to gauge its potential.

I looked for margins in my simulator, I gave it a bit better headroom for input, less current, big drain resistors and B+. But the 300pF + input cabling is really a problem except Grado MM. Miller bites there, it can be BJT cascoded though if destined for typical mm. I have in mind the Denon DL-160 too, just on a Technics 1200MKII and 1K input resistance.
 
Administrator
Joined 2004
Paid Member
salas said:



Now that's a decent one. SRPP to drive the properly impedance sized RIAA, BJT follower for output. Kevin you dropped in here to remind me of glass...:smash:
How does it compare to decent vacuum?

Not really sure as I have not heard it in an A/B situation. Given the designer's orientation to detail and serious listening tests I suspect it would be about on par with a middle of the road tube set up. (His tube stuff is very, very good.) Certainly better than most SS designs I've heard and more open sounding than the typical cookbook op-amp based implementation which it was intended to better. Results are very dependent on the amount of effort put into building it. (Jfet choice and matching, passives, layout, etc.)
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Did some simulations, calculated some Zo, modified the RIAA for that, added an output buffer. Gives a lot of gain for 12mA IDSS. I have some strong BL ones. See where it gets in real life though.
One of my DIY friends has a very low 0.25mV Ortofon MC on a Walker turntable. We have modded a Steve Bench double 12AY7 cascoded by ECC88 schematic. It is input driven by Lundahl X20 step up. Sounds best so far after 4 other tube RIAA circuits. One was a Thorsten Loesch. I guess I will have to make the attached configuration and listen in that system VS the glass so to know where it stands.
 

Attachments

  • riaas2k73db.gif
    riaas2k73db.gif
    6.9 KB · Views: 15,554
Hi Salas

I play with Rload , 3k close to much the sound , usally I like 23k , but in my system i think 8-15k range are optimal
(a switch // r for two setting, soft or racing )


Pacific+ dl160+P3 compete with last Pedja dac.....not so bad!!!

for me is exceptional result, this dac are nos , trafo have 7 separate seconday, psu all discrete regulator ,1541 are S2 ,no feedback,no output caps ,all quality component.....
 
This is good concept...
and it will be working better then any OP amp riaa
expetialu ones with feedback riaa circuit...
try to move the C1 and R9 from input of
Q2 to outout of Q1
in my tube riaa that was the move to linearise totaly db curve...
try to calculate propper passive netvork
You can use morgan jones example
(the very few are right...)
I am not shure for this one...
Use passive inverse riaa network for simulation
not only straght signal...
try haggerman inverse riaa
cheers
You will be amazed
 
diyAudio Chief Moderator
Joined 2002
Paid Member
I calculated (AES Lipshitz) the RIAA by subtracting the given input stage's Zout from R10//R5. That is why it looks a bit odd for values given it's displayed 36k input. I will have to check real life accuracy of course.
Had C1 where you suggest initially, and it was a 4.7uF. Then I saw some mods in the Pacific links that they moved it before Q2. Why is that exchange? They claim its for better.

*Can it be that in your tube RIAA the shunt was miscalculated and was changing the curve?