Jan, I agree if the controller is just doing control of the PGA. If you are driving a display and doing a whole lot of other things, then it's not that easy. Since the controller is also powered off a separate transformer and power supply, isolation also makes sense from this perspective as well. BTW, even if I use something else here (e.g. relay volume control), I will still isolate the digital and analog sections.
Bonsai said:
Depends. If you update the display when the level changes, or a different source is selected, you can still go to sleep. I still think it is not very helpfull to isolate 'digital hash' if it isn't there. But, YMMV.
Jan Didden
I tried WM8816, PGA2310, CS3310, PGA2320 is(IMO) best.
Practical resuts:
http://www.bvaudio.sk/PGA/PGAE.htm
Controler is driving display, PGA, input selector (relays) and receiving IR.
Practical resuts:
http://www.bvaudio.sk/PGA/PGAE.htm
Controler is driving display, PGA, input selector (relays) and receiving IR.
BV said:
Very good. As long as there's 3 zero after the comma, no worry!
BTW What sound card was that?
There is a funny peak at around 1300 Hz (for the 1kHz measurements) and 13kHz (for the 10kHz measurements) in the right channel, any idea where that comes from?
Jan Didden
Sound card is Lynx L22. Acording peak at 1300Hz- measurement setup was unbalanced, sound card is internal in PC with CRT display...Here is to many ways for injection of unwanted signals.
Great set of results - so much for all the crap about how bad the PGA's are. Use optos. The noise will go down even further 😀
Bonsai said:
Doesn't matter, the PGA's can't sound good. They have more than three legs!
Seriously, I played extensively with the PGA2311 and the results were very similar (noise) or slightly better (THD). PGA2311 max output should though stay under 2V.
I love my PGA2320 volume controls (thanks error401). I am a cloth-eared mid-fi though, so my opinion doesn't count. 😱
You are talking about feedtrought from control (digital) inputs to analog signal?
Noise background have nothing to do with this.
BV said:
Well, these anecdotes that opto's cut down noise crop up all the time. Problem is that people don't really think about it but just parrot what they once heard or read, even if it was in a totally different context.
Jan Didden
if the problem with digital breakthrough only happens when the PGA is set to provide gain, then why not simply provide the gain ahead or after the PGA, and use it only for attenuation?
Alternatively, couldn't you put RC filters on the digital lines to reject high frequency crap without being so high as to distort the digital control signals ? These are only low speed anyway... so i would imagine an RC filter would work quite well.
Alternatively, couldn't you put RC filters on the digital lines to reject high frequency crap without being so high as to distort the digital control signals ? These are only low speed anyway... so i would imagine an RC filter would work quite well.
jaycee said:
The digital control lines are at DC. They ONLY change state, and are digital active, when you change volume. After that, put the controller to sleep, shut down the oscillator, and everything is back to DC again. What's the point??
Jan Didden
Janneman, no need for the lecture - I've done enough low noise work (T/C amps etc) to know that isolating the digital and analog domains plays dividends in keeping sensitive analog circuits noise free. What's your beef?
Suppose for some reason you are *not* having the controller go to sleep ? eg, you are polling something regularly like a rotary encoder, or updating a display with a VU meter...
Also surely the filtering would help keep the noise out when volume is being changed?
Also surely the filtering would help keep the noise out when volume is being changed?
You cannot aplly RC filter for CLK signal, it must have defined transient time, for WM8816 it is max. 100ns...jaycee said:
Bonsai said:
You're right about that. It's a good idea to isolate ground loops if there is a chance that ground currents in one part (digital) invade the ground on the other (analog) part. But that has nothing to do with digital noise of the 2 or 3 control lines into a VGA, because there isn't any. It is an issue for instance in a DAC where active high speed signals and analog signals co-exist side by side and simultaneously.
And I still feel that when the ucontroller sleeps all the time except when you change volume, there is no issue with ground currents because their ain't any.
Jan Didden
hi all,
i am very interested in this subject, to date i have never considered using a chip for volume adjustment, never liked the specs. however some of the later generation devices dont look bad at all. But which one to try?
i would like to ask for a vote of types that have been tried by folks and their recommendations. this would be extreamly useful.
i am very interested in this subject, to date i have never considered using a chip for volume adjustment, never liked the specs. however some of the later generation devices dont look bad at all. But which one to try?
i would like to ask for a vote of types that have been tried by folks and their recommendations. this would be extreamly useful.
- Status
- Not open for further replies.