A New approach of Volume Control: AAVA (Accuphase Analog Vari-gain Amplifier)

Status
Not open for further replies.
Jump to Latest
tiefbassuebertr

tiefbassuebertr

Member
Joined 2005
tiefbassuebertragung.de
#1
about
http://www.accuphase.com/cat/e-250_e.pdf
I read:

How AAVA-II works:
AAVA-II operates by feeding the music signal to a V-I (volt-
age - current) converting amplifi er where it is weighted in 16
steps [1/2, 1/22, ..., 1/215, 1/216].
The 16 current steps are
turned on or off by 16 current switches, and the combina-
tion of switch settings determines the overall volume. The
switching operation is controlled by a CPU according to the
position of the volume control knob. The combined signal
current forms a variable gain circuit that adjusts the volume.
Finally, the combined current is converted back into a music
signal voltage by an I-V (current - voltage) converter.

Are there any experiences from this kind of a level attenuator without any resistors in the signal path?
 

Attachments

Last edited:
#2
This whole complicated circuit (well, more complicated than two resistors anyway) contains no resistors? Does it use switched capacitors instead? Or are we to believe that a whole set of p-n junctions can be matched so accurately that the outcome is more linear than a resistor? Is not a current switch a resistor, of sorts?

This sounds like more straining at gnats and swallowing camels!
 
#4
A digtally controlled attenuator. There are a couple of chips that do the same. Not sure if this is realized in discrete form or if it is on an IC.

Of course the music signals flows still through resistors, as with any gain control chip or any relay controlled attenuator. I respect Accuphase very much, but I don't see the big advantage of this, compared to other simple, established solutions.
 
#6
Might this be the patent?
JP 2004104269 SIGNAL SWITCH AND VARIABLE GAIN AMPLIFIER PROVIDED WITH THE SAME
The switch comprises a fixed current drive circuit having a voltage-current conversion gain for converting an input signal voltage applied to an input terminal to a current to output, a current-voltage converter circuit for converting an output current from the drive circuit to a voltage to output, a plurality of signal switching circuits connected in parallel between the input terminal and the input of the current-voltage converter circuit, each composed of a series switch circuit connected between the input of the current-voltage converter circuit and the output of the drive circuit and a short-circuit switch circuit connected between the drive circuit output and the ground, and a switch control circuit for controlling the switch circuits.; The switch control circuit changes over the switching of switches in the plurality of parallel-connected signal switching circuits singly or multiply with specified time intervals one after another to control the gain
Click to expand...

Its in Japanese :scratch1:
 
#7
This is interesting approach, but I think it's too complicated for volume control. Maybe it's really as good as they claim, but it's very bulky and expensive.So this is for some rich audiophils, wich don't know what to do with their money.
 
#8
I've looked at this before and think this type of volume control is really interesting, I would love to make one. If someone gave me a schematic I would design a PCB.

I like it because:
- it's so precise compared to everything else
- lower distortion than PGA2310 etc
- Simple concept
- No click sounds. Pretty much all relay volume controls click when all relays flip (like at step 64/128 etc)

There is some more info here on page 2:
http://www.axissaudio.com/pdfs/C-2400.pdf

Mark Levinson use the same concept too:
http://www.marklevinson.com/image_library/32AB_lo.jpg

bbp built one a while back. He used DG413 analog switches but these days there are better ones like DG611 or ADG1221. He also measured he version and achieved higher specs than the PGA. Here are his threads:
http://www.diyaudio.com/forums/solid-state/127997-solid-state-volume-controls.html
http://www.diyaudio.com/forums/soli...igh-precision-digital-control-attenuator.html
http://www.diyaudio.com/forums/vend...control-module-ultimate-attenuator-audio.html
 
#9
I was thinking about designing something similiar, but have no time currently. Electric circuit seems not to be a big deal. More interesting is interfacing switches to volume know. I was thinking about absolute one turn encoder and some simple uC, like 8051. Is this a correct idea? Or anyone has more simple way to do it?
 
#10
barmanekm said:
More interesting is interfacing switches to volume know. I was thinking about absolute one turn encoder and some simple uC, like 8051. Is this a correct idea? Or anyone has more simple way to do it?
Click to expand...

You are on the right track. Personally I would use an arduino and a standard encoder since there is a much bigger range available as opposed to absolute encoders.
 
#14
#17
The advantage of converting the signal into a current, passing it through either an MOS switch to ground or an MOS switch to virtual ground and then converting the current into the virtual ground back into a voltage again is that the on resistance changes of the MOS switches cause practically no distortion. The switches work at a constant VGS, which means that their on resistance doesn't get modulated much by the signal, and any remaining on resistance modulation doesn't affect the current anyway if the current sources are nearly ideal.
 
#18
tvi said:
Might this be the patent?
JP 2004104269 SIGNAL SWITCH AND VARIABLE GAIN AMPLIFIER PROVIDED WITH THE SAME


Its in Japanese :scratch1:
Click to expand...

When you click on the link to the B2 version, you can download it and look at the simplified schematics at page 11. The fact that the patent's text is in Japanese doesn't matter much, as patents are in general unreadable in any language.

What isn't clear to me is how they keep the offset under control. The slightest offset in the voltage to current converters will give ticks or thumps when switching.

Another issue is noise. At least with a conventional negative-feedback voltage-to-current amplifier, the noise of the bias current sources ends up in the output current, so you need current sources with lots of emitter degeneration to get a nice, low noise floor. Then again, a conventional negative-feedback voltage-to-current amplifier would have too much offset anyway, so they probably do something different.
 
Last edited:
Status
Not open for further replies.
Top Bottom