PPA Project Announcement
Jun 27, 2003 at 1:44 PM Post #301 of 340
Quote:

Originally posted by morsel
There was almost no difference between using a cheap unregulated wallwart, a regulated Elpac WM080 wallwart, and an LM317 based regulated supply.


Can you say a little more about this? Were there differences between the Elpac and the LM317 PS? Thanks and welcome back!
 
Jun 28, 2003 at 8:09 AM Post #304 of 340
hm, i did a bit of reading, and the meta42 v3 would not be better-sounding than the ppa ... so never mind about that part. any other changes? ie, if i were to have a maxed-out mint (heh) at the moment, would it be good for me to upgrade to a ppa, or a meta42? my sources aren't great anyway, but i do plan to get much better headphones.
 
Jun 28, 2003 at 10:13 AM Post #305 of 340
The difference between the Elpac and the LM317 supplies on PPA noise was small enough that I didn't bother to take notes. I'm guessing it might have been a couple db difference at around -110db down. Perhaps Kurt recalls the specifics.
 
Jun 28, 2003 at 6:06 PM Post #306 of 340
The JFET isolated rails and separate VGD for the Opamp rails, as used in my Pocket Amp, in concert with the Ground channel are incorperated in the PPA to improve the Amps Power supply rejection Ratio and prevent unwanted feedback of Load current noise to both the sensitive opamp rails in addition to not contaminating the Signal ground. this also has the advantage of almost removing the quality of the DC power supply from affecting Both Performance and most importently Sound quality.

BTW the sonic advantage I heard using Three vs one Op Amp rail is an inprovement in sound stage. this was noticable on my Sony MDR-V6/7506 Phones. These sonys are not Knowen for soundstage and with just one Isolated Op Amp supply the image was confined Between the ears but with Three Isolated Op amp rails the image extended to about 1 inch from the outside of the ear cups, Moreover the Image also had depth and extended to the back and lower base of the Neck vs just between ears within Head. Outside sounds also seemlessly blended with the Music. Nice.
 
Jun 28, 2003 at 11:04 PM Post #307 of 340
Quote:

Originally posted by ppl
The JFET isolated rails and separate VGD for the Opamp rails, as used in my Pocket Amp, in concert with the Ground channel are incorperated in the PPA to improve the Amps Power supply rejection Ratio and prevent unwanted feedback of Load current noise to both the sensitive opamp rails in addition to not contaminating the Signal ground. this also has the advantage of almost removing the quality of the DC power supply from affecting Both Performance and most importently Sound quality.
....


I am very curious how many more dB of PSRR this technique can produce. Any results on this?
 
Jun 29, 2003 at 12:00 AM Post #308 of 340
Quote:

I am very curious how many more dB of PSRR this technique can produce. Any results on this?


I just did some informal testing with a signal generator injecting ripple on top of one rail with a 10V supply. I increased the ripple signal until I got at least 2mV on the other side of the JFETs, then measured the ripple voltage. (This was to ensure that I wasn't straining the lower limits of my millivoltmeter.)

The rejection across the JFETs is frequency dependent. At 120 Hz, the ripple is attenuated by about 15dB, at 1 kHz it's 32 dB, and at 10 kHz it's 29 dB. This suggests a ramp upward towards a plateau, which is a neat complement to an op-amp's declining PSRR curve.
 
Jun 29, 2003 at 2:08 AM Post #309 of 340
Quote:

Originally posted by tangent
I just did some informal testing with a signal generator injecting ripple on top of one rail with a 10V supply. I increased the ripple signal until I got at least 2mV on the other side of the JFETs, then measured the ripple voltage. (This was to ensure that I wasn't straining the lower limits of my millivoltmeter.)

The rejection across the JFETs is frequency dependent. At 120 Hz, the ripple is attenuated by about 15dB, at 1 kHz it's 32 dB, and at 10 kHz it's 29 dB. This suggests a ramp upward towards a plateau, which is a neat complement to an op-amp's declining PSRR curve.


This was the Intent of using this topology. If a lower Idss FETs were used the rejection would be even better. The PN4392 is a rather stiff device for the AD-8610 However is good if Higher Icq op amps are used like the AD-843. I like to use the 2n5486 with the AD-8610. Let us also not forget the Psrr of the Buffer on the Ground channel. had more room been available on the PPA board i would have used the capacitor Multiplier like i did on my older Portable Amps. The jfet only was used in my Pocket Amp because of lack of room on the Pocket Amps board as is the case with the PPA also. I may play with the PPA board and see if i can incorperate the additional cap multiplier's on the rails with out upsetting other Requirements.
 
Jun 29, 2003 at 2:50 AM Post #310 of 340
This may have been addressed before on one of the previous 16 pages but I really don't feel like reading all 16 pages... perhaps you can understand.
wink.gif


How hard/easy is it to adapt the PPA board for home use ONLY with a true dual PS? It seems as though a lot would have to be done to the PS section to change it. Is it as simple as wiring it up, or a tad more complicated, requiring parts changes, etc.?
 
Jun 29, 2003 at 9:03 AM Post #311 of 340
Due to the Highly intergrated nature of the PPA's internal power supply a single External Supply such as a qualty wall wart or a regulated Conventional supply on a singel rail (Positive & ground) will preduce as good of results as using a split rail supply. the PPA is designed to be a portable that has the sonics of the Better Home only units. this required an aproach to power supply design that is Highly optimized for this Amp.

so if one were to do as KTpG Sugested then three Regulatedl ow current dual rail supplies are required one for each op amp Channel. then another singel polarity supply just for the output stage. In no way should an attempt be made to tie any supply connected to the output stages to ground.

While the PPA is a DIY Amp the amount of custom user changes are kept to a min. as jeffreyj pointed out in another Post I would not want someone to undoo all the effort put into this Highle Optimized Amp and thus create a less than Ideal Performer.


Wher are those "No user servicable parts labels at Hee.. hee..
 
Jun 29, 2003 at 12:10 PM Post #312 of 340
You could configure a PPA with just one set of op-amp rails, remove all the TLEs, jumper the TLE-G position, tie PS ground to the ground plane and the PS V+ and V- to the appropriate places. That should work. But, to what end? I doubt it'd perform as well, and the premium for the dual supply is likely more than what you saved by removing the TLEs.
 
Jun 29, 2003 at 12:30 PM Post #313 of 340
Quote:

Originally posted by tangent
....
The rejection across the JFETs is frequency dependent. At 120 Hz, the ripple is attenuated by about 15dB, at 1 kHz it's 32 dB, and at 10 kHz it's 29 dB. This suggests a ramp upward towards a plateau, which is a neat complement to an op-amp's declining PSRR curve.


Coss is starting to come into play around 10kHz, then, but the suppresion is still impressive at this frequency and seems to be very much worth the additional components. Nice!
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 2 (members: 0, guests: 2)
Back
Top