DIYHiFi.org

Hi Guys,
this my last test but 2 different disc 1Khz at 0db, diagnostic test cd and another one is original Technics test cd. R4=100, R2=120 and R7=3.3K the upper CCS = 10.1mA and lower CCS = 8.58mA.
Here with technics disc

Maxpou

here with dianostic disc cd

I cannot comment about the cause of the different results, but -80dB and the clean spectrum with the Technics disk test is good. I bet it sounds better than most existing Opamp I/V's..... Jocko deserves a "thank you" isn't it? ;-)

Don't you hear the lack of bass slam?
I have built several versions, even with JFETs running high current....
The only plus I found is lack of glare.....

Combined with a 1-bit DAC that also lacks slam I fear bass is amputated.....

Hi,

the measurements already look fine. A clean decay of the harmonics with a dominant K2 could be expected.
The bias currents are now close to my SimA. What is the DC-working point voltage at Riv? It should be ~7.5V.
The sim showed a higher THD-figure of -72.6dB and a clean decay of higher harmonics and a Vriv of 7.5V
But of course the sim was done with 0dBfs not -3dBfs. Corrected for this, the sim and reality seem to be quite close.
As suggested in the same post, try to increase bias currents even more to the values of SimC.
Maybe You can reach the magical -90dB limit

jauu
Calvin

ps. I found that a ´lack´ of bass was due to insufficient capacitance right at the JFETs supply inputs. 1000µF or more right at the pins saves the day

Hi,

I´d suggest at least 100µF at the supplies for the IV-part. Then adding series resistors of 560Ohms going to the JFETs, decoupled right at the JFET pins with 1000µF.
A reduction of the source resistors to 4.7Ohms increases the bias current of the JFETs from 5mA (27Ohm) to 7.2mA for a SK170BL with an Idss of 8mA.
JFETs like current and typically sound better with loads of it. Also the output impedance drops which improves behaviour into lower load impedances.
With all these, the JFETs then see approximately +-10V supply voltage which leaves more than enough Vds left even at full output voltage stroke.
And it keeps the heat dissipation of the JFETs at nearly the same level as with 14V/5mA supply.
So one gains on all important parameters, with just a slight increase in part number count.

jauu
Calvin

Hi Calvin,
I have 6.5Vdc at Riv and for simC I tried it, I have 7.95 at Riv. The result here. Maxpou

Hi Max,
Congratulations for your progress, good job !
I have two short observations to make. No, make it three:
1. You drop a significant voltage on Riv. Search for "excess noise". AN-348 from AD may help here...
2. A lot of mains related components... Not having those can't make it worse...
3. The rise between 10k - 20k doesn't look 'dangerous' yet but suggests what is coming after it...If your amp becomes too warm or sounds 'compressed' already at lower volume levels when compared with a cleaner source, then you'll have to think about making a better filter.

One advice (it certainly proved useful in my case): try to measure using the same methods and tools other CD-players / source components and then try to corelate the differences between measurements with your listening impressions and/or with the changes you made in your circuit...