The first plot is plain awful.
Perhaps you can elaborate a little. R9, R13 and R14 (original typology) set the input impedance for the buffer (and amplifier) which you have just dropped dramatically...
Also, the circuit you show is incomplete. By Hypex's own admission "EMC/ESD countermeasures are left out for clarity". These presumably include 100R/100p input filters to chassis to filter inbound HF. How does "EMI/RFI and noise from Hypex Amp and power supply" get into the audio from this path? Presumably the circuit also omits for clarity bypassing of the op amp supplies. Are Ce and new R13 necessary if proper EMC/ESD countermeasures and proper supply bypassing are in place?
The other thing I find interesting is that there are no DC blocking capacitors, bipolar ideally in the region of 47u - 470u (the bigger the better). If placed ahead of R13/14 (original circuit) then, perhaps counterintuitively, one would likely find that noise increases with lower values for R13/14.
Last edited:
Nothing on the NC500 datasheet has anything to do with the separate buffer. The buffer was intended for basic testing of the NC500 for OEM's. However some decided to use it in their commercial products. But that being said I would like to hear more details on the testing procedure. Because those OEM buffer boards are dead silent in my system.
Hi 3lviz this is all very interesting.
Not sure what the rest of your system is perhaps you can you tell us.
I found the Sparkos a bit dull and the Burson just right in my system.
Also when you have finished all your testing could you list out all the op amps with a brief summary of your findings and your order of prefence.
Last edited:
IMO, the LME49860 is somewhere between those two discrete opamps. There is just a slight amount of hiss (ear on tweeter) tho, but it's under my personal threshold of too much noise. Worth a try, but could be too dark sounding for your taste...or then again, maybe its just right...😉.
I don't want to put all the opamps in order, since some of them are tested less than others. Only some get to the longer listening, some just seem to be worse than the previous option. Here is the complete list so far:
LME49860
OPA2132
OPA2134
Sparkos
Burson V5
994-enh-ticha
LM4562
OPA627
NJM4580D
OPA797AR
OPA797BR
OPA2228
NE5532
Iirc, that should be all.
My setup
Streamer: PC/USB
Source: MSB Analog Dac + Diamond power + Volume control + Premium QUAD USB2
Amp: Nord One Up (with RCA)
Speakers: Marten Bird 2
Last edited:
Thanks Julf for sorting that out for me. It also helps me understand some of the previous discussions.
Its a busy old night on here with these two threads firing away! Anyway Nerdman (Sonic Imagery) provided me with the schematic a few weeks ago after contacting him re the noise with the 994 installed which does settle down. They do sound superb a noticeable improvement in depth of soundstage and well I just love listening to them. Could measurements provide a definitive answer to what sounds good? If so which one? I shall try the mods in due course and if necessary incorporate into a Rev C of the board. Took the Nord UP to the HiFi Wigwam show the weekend and the reaction to my room was incredible Lots of "best sound we have heard so far" Even had 5 individuals come back and camp out in my room in the afternoon.
Boggit, thats good news 🙂! However i'd like if you tested the rev C with other op amps also. So is it going to be board for the 994 or possibly a more generic fix other opamps as well?
Rubbish. It is the "recommended design" for the input buffer.
The application hint goes on to provide a model for estimating the currents the input buffer must provide and alternatives to the recommended design.
The designer of the amp module has provided his recommendations as to the design of the necessary input buffer. That doesn't mean those recommendations have to be followed but to say that there is nothing in the NC500 data sheet which applies to the input buffer is rubbish.
The discussion here is pertaining to the recommended design as outlined in the NC500 data sheet.
Sonic Imagery have suggested some changes. Most of these deal with stabilising their op amp in this application. No surprise an unstable op amp led to greater distortion.
As for the 47k resistors being the predominant source of noise, I struggle to believe this. They're not even in the signal path. In stark contrast to the op amps and feedback resistors....
The discussion here is pertaining to the recommended design as outlined in the NC500 data sheet.
Sonic Imagery have suggested some changes. Most of these deal with stabilising their op amp in this application. No surprise an unstable op amp led to greater distortion.
As for the 47k resistors being the predominant source of noise, I struggle to believe this. They're not even in the signal path. In stark contrast to the op amps and feedback resistors....
Last edited:
Great news Colin! I also like the 994's the best once the noise was taken care of. Your new revision C board will likely be outstanding.
Yes this is the big question, if measurements are critical, which measurements should we be looking for? Perhaps Richard can provide some insight to why his 994 opamps sound so good.
Ok be argumentative just for the sake of it. There is a section in the NC500 data sheet pertaining to recommendations with respect to the missing input buffer. The designer recognises OEMs like to design their own but has provided recommendations nonetheless - in the data sheet. So it is incorrect to say that nothing in the data sheet relates to the input buffer. Furthermore, the discussion at hand is in relation to the (recommended) design ... included in the data sheet.
As for your image, it looks like you pulled it from the list of images of the Phison Audio PD2 for which you are a distributor. Your design or Messrs Anderson?
"Recommended" just like ketchup is recommended for use on a hot dog.
My "Recommended" buffer has slightly different circuit topology.
Yes, but that is a piece of a 7,5k machine.
Let's get back to earth.
Depends on what earth you are on. Some peoples earth is a $50 set of computer speakers connected to the analog outs of their netbook.
We are all in the same earth. 7.5k is ~2000 Big Macs (using your example🙂).
That's a lot of food 🙂.
The point was that a 7.5k pre, to drive 1k amp is a little to much. Isn't it?
Lets get back to finding the best solution for an input buffer for boggit's amp, which is a very interesting subject (in this thread). And your contribution on this, was significant so far. Keep on