Pearl and analogstage news.

My analog stage is in action for some weeks now and the sound is better than i could expect. I have put the output through an attentuator, like aleph L, into the aleph 30. I have a coppling cappacitor in the input, i know that its not neccesarry but i had to change the phase because the circuit is unbalanced and so it turns the phase 180 degrees. With the cappacitor its right again.
Now i get the liquit rythm i was missing before and the distance to analog is getting smaller again.
Thank you Mr. Pass for publishing no project manuals from your DAC!
The housing for the pearls are finished too and all the three are at there new homes. My friends cant stop listen to records anymore. A pearl makes even a ortofon T3000 redundant. The sound improvements are the same in every equipment they sound cleaner, nicer with more resolution more rythm and more color. The only point missing is still some more gain. An ortofon MC3000 has just 0.15 mV output. I think about the possibilities for a third gain stage. I dont want to change the power supply, so i dont get negative voltage, thats limiting the possibilities. Maybe DC coppling is a solution, or the easy way another pair of Jfets without cascading, it guessi it could have more gain than 10 DB.

Im also looking for an alternative to the 9Volt batteries, because you never know if they are empty. Maybe a precion voltage source like the LT1019N8-10 could do the work, it has a very good ripple rejection, but its expensive.
The picture show the analog stage as its running. I would like the nice looking Pearls in their housings too, but i havent got a digicam now.


  • dsc00397.jpg
    88.6 KB · Views: 606
One way to ge more gain is to simply use more devices in parallel. That is what I did.

Input stage is 8 parallel 2SK170
Output stage is 4 parallel 2SK170

Source resistors were reduced to 18 Ohms without any issues (but then again that was to be expected since devices were tightly matched).

Input load resistors was also made larger, and PSU voltages upped to match these larger load resistors and resulting currents.

R23 was changed to reflect the "correct" input to filter. Tip: R23+R19 should be the same as Wayne put in. However, as has been noted, the alignment of the passive RIAA filter is not in line with the reference (Lipshitz). Then again, it is good enough according to my simulations and measurements. However, next time, I will use a Lipshitz alignment.

I did some major work on the power supply side, changing the active input to multi-stage passive and adding secondary cap multiplier to the output stage as well. May I suggest ca. 1 uF film or higher? I used such bypassing everywhere, including electrolytics. It does add to the cost since such caps can set you back about $1 or more each depending on your taste. My system thus came in at a price point $30 higher which was acceptable to me. I did not try without such bypassing.

In the Pearl, the voltage dividers are not always bypassed. This is something you might want to add. R21 and R28 are the first places you should consider using bypass and in my opinion is where a capacitor to ground will do minimal damage if not a lot of good :)

Reinhard, im concerned about the feedback in the power source, i thought all the effort with the differencial regulation and the 9V batterie is just because off the need for a clean supply and than i wondered if the ono doesnt need the effort. But i guess i missed something, if the ono power supply is clean enough and i dont put this into question, so than colburn must have had another reason for the pearl solution. I like too make some changing like the LT1019N-10 i mentioned yesterday, its interesting if there could be an improvement.
Petter, i did the same 6 sk170 in the first and for in the second stage and i substituted R19 with 1K it brings about 4DB more gain. Why do you think R19 shouldn´t be changed, it didnt make any sonical differences, just more gain.
I like to try bypassing R20 and R21, but did you really mean R28?, i wonder if this would be usefull. Increasing the supply voltage a little bit could help too. I tried some bypassing with filmcaps but it sounded bad in the high frequencies, sometimes electrolytics do a better job, a more balanced sound.
Did you try the second stage without capmultiplier? Im not shure about the influence it has on the performance, i though about multiplying the cappacity instead of using a cappacity multiplier.
Its allways worth to try things out, but i dont like it so much to change finished circuits because it often damages the optical appearance.:)
You are right, it is R20 and R21. I did not mean source of input transistor (shame on you Wayne for having such low quality bitmap)

I too changed R19 to 1K to double gain for given current. This alone means that power supply needs to be increased. Since I also run double current through it this also necessitates higher supply voltage. I can't remember what I did with the output load resistor -- the factors I considered there was output impedance versus gain. Do you want higher output impedance with higher gain or not?

The sum of R19 and R23 defines the output impedance (simplified) of the input stage. The filter is sensitive to such impedance. So reduce R23 by 500 and you are back to intended alignment.

I did not try second stage without cap multiplier.

I have about 54dB of gain. My balanced version will have much more.