http://www.diyaudio.com/forums/analogue-source/129126-simplistic-njfet-riaa-839.html#post3221512
It seems quite easy looking at the schematic: http://www.diyaudio.com/forums/analogue-source/129126-simplistic-njfet-riaa-802.html#post3167648
Maybe changing R13 and R2.....
PS: Thank you Vgeorge
Last edited:
Would a "test" version of one (not stereo) of these circuits be a good way of identifying good low noise versions of the 2sk170 ready for assembling the first stage of the stereo versions?
Which version would be more selective in identifying the slightly noisier jFETs to be put into the second stage, or not to be used at all?
Which version would be more selective in identifying the slightly noisier jFETs to be put into the second stage, or not to be used at all?
post8621 link1 took me to post8601 !!!
post8621 link2 took me to post8601 !!!
Both links took me to the same wrong post !!!!!!
post8621 link2 took me to post8601 !!!
Both links took me to the same wrong post !!!!!!
Maybe for noise only.... of course other aspects like sounstage width and imagery could not be evaluated and these are the strongests points of thiese designs.
http://www.diyaudio.com/forums/analogue-source/129126-simplistic-njfet-riaa-161.html#post3167648
post8368 has no sch and no link
post8368 has no sch and no link
Last edited:
I am talking about noise selection.
Many report that the K170 and F862 have different noise.
For a high gain Riaa stage maybe the noisy jFETs should not be used in the first stage.
If this is the case, then we should be selecting out the noisy devices so that they don't get used in the first stage.
Then we can fine tune the gains/EQ to get the final sound required.
I believe it is a good idea to compare ifferent types of jfets for the input, but as using different jfets with different transcondutances would need also different values for Rs and other parts, I do not know if it would be conclusive.
Anyway, I never detected differences in noise using 2SK170BL. In my builds I believe the differences in noise come principaly from the value and type of Rs used.
In this case, using two k170 in parallel with much lower Rs really made the difference.... This 60dB build is even lower noise than the bjt build I did previously.
Anyway, I never detected differences in noise using 2SK170BL. In my builds I believe the differences in noise come principaly from the value and type of Rs used.
In this case, using two k170 in parallel with much lower Rs really made the difference.... This 60dB build is even lower noise than the bjt build I did previously.
The folded cascode first stage can give times hundred. If enclosed in copper shielding and better be powered by floating means, maybe a battery array, so to even isolate the ground it refers to, it could do for amplifying the equivalent input noise of JFETs for a handy nVrtHz characterization jig.
Salas,
are you saying use the post8011 sch as a pre-amp and then add a test fixture before it?
Or
Remove Q1 and use this position as the DUT location?
are you saying use the post8011 sch as a pre-amp and then add a test fixture before it?
Or
Remove Q1 and use this position as the DUT location?
Ok... this seems a nice idea... maybe I could build something like that... maybe it could be used to feed a pc soundcard so we could "see" the noise instead of merely try to determine differences by ear.
I need something to read about folded cascodes... would you point me a good link ?
I need something to read about folded cascodes... would you point me a good link ?
I would use 3 female pins base for exchanging Q1 as DUT, a higher source resistor to accommodate wider IDSS, say 10R factoring its noise in. It would take centering at 1/2 the LEDs + Vbe with the trimmer each time, or a steady resistor and variable PSU. By comparing the total amplified output noise each time, the DUT could be relatively characterized.
Forget ear, we are talking instrumentation now. For ear you can just exchange Q1 in a full phono on a pins base in a given system and go stick at the tweeter for hiss, is the most practical I can recommend.
That's the answer I was hoping for.
Best (ear) noise performance in the actual circuit it will be assembled into.
@RC
Folded cascodes are used mainly in op amps. You can look for some Analog Devices datasheets and they are referred in electronics papers about cascodes in general.
Folded cascodes are used mainly in op amps. You can look for some Analog Devices datasheets and they are referred in electronics papers about cascodes in general.
Hi Salas
What do you think about the 4391 fets ?
http://www.datasheetcatalog.com/datasheets_pdf/2/N/4/3/2N4391.shtml
What do you think about the 4391 fets ?
http://www.datasheetcatalog.com/datasheets_pdf/2/N/4/3/2N4391.shtml
Those are high IDSS fast choppers that would lead in high source resistance values to control the voltage drops for logical supplies in a circuit like here. Low gfs hence noisier than other options. Good for sampling and fast switching or low drift CCS in an amp's VAS or something.
No go
Found those here: http://www.diyaudio.com/forums/analogue-source/200447-folded-cascode-hijinks-pn4391.html
Found those here: http://www.diyaudio.com/forums/analogue-source/200447-folded-cascode-hijinks-pn4391.html
Last edited:
Wrench had a series of great concepts basing on American chopper FETs for MM. That linked one maybe I missed, I will read it again to refresh. Not best applicable for circa 60dB gains and noise levels established here IMHO. Still folding is just another circuitry technique not specific parts related. Can be done with anything and as everything in electronics must be judged to best integration and spec. Roender had it in his amp's input stage with Japanese FETs for instance, and so on and so forth.