Mpp

[janneman]

Originally Posted by Joachim Gerhard

Hello Jan !
i really whould like to meet you some day. maybe i just drive by if you care.
by the way i started to buiild your Pax anplifier and may need some advice from you.
How could i buy this issue of AudioXPress ?
Yes Malcolm is the most creative design genius i ever met and we did a lot of exiting stuff together. unfortunately he is far away and i am not so flexible any more then i was 15 years ago. i hate that driving on the left side ! i am a poor driver anyway.

I'd be glad to help with the paX!
I can send you a copy of that AX, please send me a mail address to jan at linearaudio dot nl.

Now go to bed

jd

[anatech]

Hi Joachim,
A warm welcome to diyaudio! I'm looking forward to your future posts, which are far beyond me, technically speaking. Some of us are here to learn, so please bear with us.

-Chris

[Joachim Gerhard]

Thank you Chris !
I hope i do not disappoint you and will give my best.
By the way i just have looked at Jan Diddens interview with my Mentor Malkolm Hawksford and i was reminded that an I/V Stage in a DA converter has a very similiar task like an MC phonostage so the MPP could be used for that too.
when i find the time i will shurely try it. it is high speed and has no voltage feedback to rely on so the extremely poluted and wideband signal from the DA can not drive it into slew limiting temporaryly disabling the feedback loop.

[Joachim Gerhard]

Today i whould like to post some measurements.
I will show noise and distortion of the MPP prototype, the Quantummusic Goldstandart and the Aspen Paris.
Usually i start to measure at -50dB/U, that whould represent cruising level of a typical modern MC cardridge in my case a Lyra Tiatan i with 5.5 ohm impedance putting out 0.5mV at 5.0cm/sec, zero to peak, 45 degrees.
-50dB/U is a level of aroud 3mV so what we see in this measurements is the low level resolution of the compared design`s subjectively it is not that easy because human beeings can listen into the noise.
a sinewave with -30dB level under white noise can still be heard.
sharks are particular good at that to be able to detect far away prey in a noisy environment the phenomenon is called Stochastic Resonance. is it to this day not fully understood and has attracted some researchers in the past years.
i am using stochastic techniques in the commercial version of the Goldstandart and the effect can also be used to lower the noisefloor. this is particular important in the goldstandart because it is a balanced differential design and as John Curl explained so convincingly in his Blowtorch thread such a circuit has a theoretical noise disadvantage of 6dB against a parallel symmetric design with the same component count
that leave me of cause with the option to use more components, i hope to good advantage
but jokes aside, i am not able at that moment to penetrate my -150dB/V noisefloor and can not measure better then -140dB distortion so can not tell you what the resolution under the noise is.
i am shure clever people at Audio Precission or Rhode&Schwarz know how to measure distortion under the noise and when somebody here on the Forum has an idea i have an open ear. so speak about the ear.
my option is to listen so i will try to discribe the sound of thoose phonostages subjectively too. i think my music system differentiates quite well and i will later offer more details of my listening room and my components.
so here are the first measurements:
the Paris at -50dB/U
the MPP at -50dB/U
the Goldstandart at -40dB/U the Goldstandart is my lab sample, the commercial version has 4,5 - 6dB less noisefloor
i measure with a 1kHz sinewave, i did a sweep to 20kHz too to detect any rising high frequency distortion but there was nothing to worry about with the coosen designs
ceaveets: The MPP was measured with the 75usec equalisation in place
that explains the lowered noisefloor in the treble
I measured the Goldstandart at -50dB/B but the distotion was
completely swollowed by the noisefloor so i omitted that
measurement
I will give you some time to digest this and will comment on the measurements in my next post
this will also include measurements at higher levels

[Joachim Gerhard]

From left to right:
Goldstandart lab prototype -40dB/U magenta
MPP prototype -50dB/U grey
Paris - 50dB/U green

[Joachim Gerhard]

Here are more measurements with higher volume of -40dB/U
i included the Goldstandart again
from left : Goldstandart, MPP, Paris

[Joachim Gerhard]

Here come measurements @ -30dB/U
again from left :
Goldstandart, MPP, Paris

[Joachim Gerhard]

Sorry, the left picture is not the Goldstandart but the Generator again.
I can not find the measurement at the moment and will post it later.

[Iain McNeill]

Hi, I'm sure you know this already, but time averaging can lower the noise floor. Taking two time records and adding increases the signal by 6dB but the noise by only 3dB. Each doubling of time records buys you 3dB reduction in noise floor.

Anyway, please continue. I'm interested

[Joachim Gerhard]

Here you will see how the prototype of the MPP is made.
I use two groundplanes. One for the Input Stage and one for the Intrumentation amp.
I connect these planes in series and then to a single "star" that is also connected to the soft steel box and the ground jack.
I decouple locally with 330uF Panasonic ECA and 0.1uF Vishay/Röderstein MKP1837. The promo literature says that the ECA series is made for high quality audio but my measuements show nothing special. there are lower impedance
devices awaillable like Panasonic FM or Rubicon ZLH that should perform even better. Some people may prefer to bypass this arangement with a ceramic capacitor to lower impedance a higher frequencies even more but i had no stability problems although this design has very high speed and reaches several megaHz. If you use a ceramic i whould recomment 10nF as NPO or COG. Lesser varieties may produce distortion so can Tantals.
I supply the prototype with 9V alkaline blocks so this is not really a high end solution more a kind of worst case scenario. the noise under 1 kHz can be lowerd at least by 10dB with a more sofisticated powersupply.
this setup i only use for my own pupse so when the design is finished i whould like to put the input stage on a copper piece with the transistors thermocoupled and mechanically decoupled with soft springs. i also will put the MPP in a nice box so please do not take this prototypes as products.
On the other hand this kind of prototype construction gives quick and good results and the circuit can be changed quite easyly.
So enjoy the photos.
from left : Parallel Cascode Stage with FET constant current source,
Open box with working channel on the left side and the
rudimentally inputstage with MAT02/03 on the right side,
closed box
P.S. in fact a did a small aditional groundplane over the Servo section with 10uF ECA series connected to the Instrumentation Amp groundplane