Go Back   Home > Forums > Source & Line > Digital Source

Digital Source Digital Players and Recorders: CD , SACD , Tape, Memory Card, etc.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 13th March 2003, 02:17 AM   #21
diyAudio Member
 
Join Date: Oct 2001
Location: .
I'd have to disagree with you on the 'way more difficult' bit.
While it can take a few chips if one insists on using TTL/CMOS logic, a CPLD or FPGA provides a flexible one chip solution.

ray.
  Reply With Quote
Old 13th March 2003, 02:51 AM   #22
diyAudio Member
 
Join Date: Jan 2003
Location: Audubon, PA
Quote:
Originally posted by grataku
random noise + random noise= more noise or less noise?
Actually, random noise + random noise = more noise. However, we also have more signal as well. Consider two current output DACs in parallel. The first DAC has an output current

I1 = Isig + Inoise1

and the second DAC output is

I2 = Isig + Inoise2

where all values are RMS.

When uncorrelated noise sources add, the resulting noise is

sqrt(Inoise1^2 + Inoise2^2)

If Inoise1 and Inoise2 have the same RMS value, then the total noise is

1.414*Inoise

so

Itotal = 2*Isig + 1.414*Inoise

Therefore, although the noise increased by a factor of 1.414, the signal increased by 2 so the SNR improved by a factor of 1.414.
__________________
Mike
"Never confuse motion with action." B. Franklin
  Reply With Quote
Old 13th March 2003, 03:33 AM   #23
grataku is offline grataku  United States
diyAudio Member
 
Join Date: Dec 2000
Location: -
Mike, that's exactly what I was saying. Noise adds but coherent signal adds more so the s/n increases, anyway in the interim I think I got your earlier post, although staggering would require a much more complicated circuitry than simple parallel.

What does quantization noise look like? Excuse my ignorance, but what happens if you send a signal to one dac, a signal 180 deg out of phase to another dac take the outputs from the two dacs and subtract them? Wouldn't that get rid of the quantization noise?
  Reply With Quote
Old 13th March 2003, 04:03 PM   #24
diyAudio Member
 
Join Date: Jan 2003
Location: Audubon, PA
"Quantization noise" is more accurately called "quantization error". Quantization error is actually introduced during the A/D conversion process - it is the difference between the quantized (digital) waveform and the analog input waveform. An ideal A/D converter basically "rounds" the analog input waveform to a finite set of discrete quantization levels - imagine rounding a real number (infinite precision) to an integer (finite precision).

If certain assumptions are made (mainly that the input signal is constantly changing and is not synchronous to the sampling rate) then the quantization error can be viewed as a uniformly distributed random number which can range from -0.5 LSB to +0.5 LSB (where LSB is the size of one quantization level of the ADC). Quantization error from sample to sample is uncorrelated (i.e., the quantization error at sample N is independent of the quantization error of all previous samples). Given these assumptions, the quantization noise looks like white noise with RMS level of LSB/sqrt(12).

Since quantization noise is introduced in the A/D conversion process, by the time it gets to the DAC it is actually part of the signal. Nothing can be done at the DAC end to eliminate it - at least not without affecting the signal as well.
__________________
Mike
"Never confuse motion with action." B. Franklin
  Reply With Quote
Old 14th March 2003, 07:44 AM   #25
A 8 is offline A 8  Sweden
diyAudio Member
 
Join Date: Mar 2003
Location: Gothenburg
I decided to test "piggyback" two pairs of PCM1704K. I could only compare between a single and 2 in parallel (I'll do balanced later)


I believe my set-up is reasonably neutral and transparent.
I out from the DAC chip goes directly (20 cm of double shielded cable) to the +input stage transistors (2SC3722 and 2SA1455K) of the power amp (LC Audio Millenium run on batteries with 0.5F elyts and added MCap PP decoupling at every appropriate transistor (50uF on the legs of the power trans)).

The volume is controlled by low value resistors to ground and the -Input on the power amp, test volume load (DAC load) between 100 to 500 ohms.

The speakers are Dynaudio Contour modified using metallfilm resistors, CFAC coils and MIT MultiCap RTX.

I realize I have a somewhat different set up for the IV conversion and that this might play a role in my findings...

However in this set-up parallel dacs sound MUCH worse, upper midrange and high range completely looses transparency and precision and gets gritty and harsh.
Really not much more to say, it really sounds bad. The only thing that may have improved was upper bass or perhaps that was the only area my ears could stand, the level obviously increases.

There is no way I am going parallel.

I am pretty confused why others find it better
  Reply With Quote
Old 14th March 2003, 08:09 AM   #26
A 8 is offline A 8  Sweden
diyAudio Member
 
Join Date: Mar 2003
Location: Gothenburg
Forgot another piece of info that might explain my findings...as you probably figured there are no analog filtering except what is done by the poweramp and speakers, the power amp has a -3dB of 500kHz.
Prior to the dacs the signal is upsampled to 24/96 and then oversampled to 768kHz.
I run the dacs directly on a 24.576kHz clock.
  Reply With Quote
Old 14th March 2003, 08:25 PM   #27
diyAudio Member
 
Join Date: Jan 2003
Location: Audubon, PA
Quote:
Originally posted by A 8
I realize I have a somewhat different set up for the IV conversion and that this might play a role in my findings...
Sounds are using a passive I-V converter - a 100 to 500 ohm resistor to ground? I think that you may get some strange interactions between the paralleled DACs in this case. Parallel DACs running into the summing node of an opamp-based I-V converter would be effectively isolated from each other.
__________________
Mike
"Never confuse motion with action." B. Franklin
  Reply With Quote
Old 14th March 2003, 09:08 PM   #28
Cobra2 is offline Cobra2  Norway
DIY !
diyAudio Member
 
Cobra2's Avatar
 
Join Date: Apr 2002
Location: Stavanger, Norway
Send a message via Skype™ to Cobra2
Default PiggyBack

Did you try "the Monarchy Audio" way?
(they used only output from the "piggyback" chip)
Attached: PCM63-K)

Arne K
Attached Images
File Type: jpg pcm63 piggy back2.jpg (90.4 KB, 762 views)
  Reply With Quote
Old 15th March 2003, 03:20 AM   #29
BrianL is offline BrianL  United States
diyAudio Member
 
Join Date: Mar 2002
Location: USA
"paralleled" DACs, as stated by Mike, can enhance s/n) by
3dB for every doubling of the number of converters.
(this concept is very appealing to Burr-Brown/TI and Analog Devices...)

However, unlike what some have said (including an
audio converter guy, Bob whats-his-name at ADI),
distortion is NOT reduced, in general. To reduce distortions
would require that the nonlinearity errors in the various
bits be random. In practice, any bunch of converters
you buy for this sort of project will likely be from the same
lot and thus be very likely to all have their errors skewed
in the same direction. Thus they'll be highly correlated.

On the other hand, buying the premium grade of any DAC
is a very good guarantee of getting better linearity and
hence lower distortion.

One common practice that seems guaranteed to mess up
the DAC's linearity is to use a resistor on the output of
a current-out DAC for the I/V conversion rather than an
op-amp summing junction. Saw some curves that a
friend did recently and the resistor version was pretty
pathetic.
__________________
bel
  Reply With Quote
Old 15th March 2003, 06:46 AM   #30
A 8 is offline A 8  Sweden
diyAudio Member
 
Join Date: Mar 2003
Location: Gothenburg
Default Passive IV conversion

-Yes I am using passive conversion.

Right or wrong I found it to be sonically prefered.

Before settling for the passive variant I used the opa627 as I/V converter and my perception was that I actually had lower overall distorsion with the 627.

I lost on more subtle things like "body" and 3-D feel, "drive"
Particulary the interchannel time precison (not measured, just perception) was significantly better using passive conversion which convinced me to go passive as it really pulls you into whatever you listen to.

Michael
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Parallel with different chips! Matthewong Chip Amps 18 2nd September 2011 02:17 PM
gainclone chips in Parallel steven344 Chip Amps 2 21st November 2005 04:52 PM
Elegant way to feed parallel data-input DAC chips ? Bernhard Digital Source 27 18th April 2005 09:28 PM
Active Feedback on Parallel Chips Stocker Chip Amps 8 4th May 2004 09:40 PM
can I parallel 7812 Chips PaulHilgeman Parts 16 6th May 2003 09:05 PM


New To Site? Need Help?

All times are GMT. The time now is 04:26 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2