Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Digital Line Level DACs, Digital Crossovers, Equalizers, etc.

Do you really tie Analog GND and Digital GND at pin chip level in DAC/Receivers
Do you really tie Analog GND and Digital GND at pin chip level in DAC/Receivers
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 11th January 2019, 05:04 PM   #51
Markw4 is offline Markw4  United States
diyAudio Member
 
Join Date: Jul 2016
Location: California
Quote:
Originally Posted by ygg-it View Post
@Markw4
@batteryman

It depends on if you have separate bridges/transformer or not.
Sure. Then your problem statement should include that you have some existing coupling at the power supply end. Then you could decide if you need to do some engineering work at that end to get your system performance where you want it.
  Reply With Quote
Old 14th January 2019, 06:20 AM   #52
chris719 is offline chris719  United States
diyAudio Member
 
Join Date: Jun 2004
Location: Connecticut
Do you really tie Analog GND and Digital GND at pin chip level in DAC/Receivers
Quote:
Originally Posted by 5th element View Post
Surely it depends on what it is you are trying to accomplish, the absolute lowest EMI, or isolation?

It could very well be that the lowest impedance return path would be right where the supposed inductor is suggested to go. But if the currents are allowed to flow there the digital noise disturbs the linearity of the analogue stages.

The point of the inductor isn't to minimise EMI it's to deliberately stop current from flowing where it's going to reduce performance elsewhere. I'm guessing that this is seen as a trade off. Increase EMI in a low power circuit but improve absolute performance elsewhere.

Whether or not this design process is strictly necessary is another thing entirely.
You need to ask yourself why high frequency currents are deemed to be a problem, when they return in the most predictable manner. Only the I2S and I2C/SPI interface currents are flowing there anyway. There is no problem to solve unless you have done stupid stuff like put a DSP in the middle of the analog side of the board. In fact, low frequency return currents are going to spread on the reference plane to a much greater degree.

Click the image to open in full size.

Of course, there will be lower frequency content in the return as well, but an inductor isn't going to help with that... maybe cap stitching the planes would work if you were trying to isolate from low frequency noise but still maintain an RF connection.


Connecting the planes with an inductor is a BAD idea.

Grounding Converters | Analog Devices

"But they must be joined at the package(1) and nowhere else.

(1): And with the lowest possible impedance – do not separate them with resistors, inductors or ferrite beads. "

"sometimes a converter data sheet may recommend that AGND and DGND are connected to system analog and digital grounds respectively. When a data sheet tells you to do this it is incorrect, and you should ignore it."


Grounding of Mixes Signal Systems

Last edited by chris719; 14th January 2019 at 06:42 AM.
  Reply With Quote
Old 14th January 2019, 09:00 AM   #53
KSTR is offline KSTR  Germany
diyAudio Member
 
KSTR's Avatar
 
Join Date: Jul 2007
Location: Central Berlin, Germany
Quote:
Originally Posted by ygg-it View Post
Is this correct?
No. There is no need for separate supplies or supply GND returns. You simple make the impedance of supply feeds to the local (on-board) regulators high impedance (series resistors will do in almost all cases) so that GND connections under the DACs dominate the scene. And were do we put the supply connector on the board, etc? Avoiding to elaborate here I just recommend everybody get a copy of H.W.Ott's "Electromagnetic Compatibility Engineering", this is the best available textbook on EMC and Signal Integrity and answer absolutely all questions about grounding in this thread as well as the gazillion of similar threads. It actually covers every single detail you need to know from basic electrical circuits theory (most design errors start right there) to the layout of a proper PCB to the design of a complete flaw-less product with supplies, enclosure, internal cabling etc.
  Reply With Quote
Old 14th January 2019, 09:15 AM   #54
Markw4 is offline Markw4  United States
diyAudio Member
 
Join Date: Jul 2016
Location: California
Sometimes it makes sense to use separate supplies because analog needs +-15v and digital needs 5v or 3.3v. Doesn't usually make a lot of sense to drop +15 to +3.3 with onboard reguators, for various reasons. However, those separate power supplies exist in the first place for reasons other than grounding. Since they do exist, their separation can be used as is convenient.
  Reply With Quote
Old 17th January 2019, 01:04 PM   #55
ygg-it is offline ygg-it  Italy
diyAudio Member
 
ygg-it's Avatar
 
Join Date: Feb 2008
Location: ITALY
Default built!

Hello, just built the PCM56 board but not tested yet.... Four PCM56 in parallel, left and right channel.
I've decided to create two separate power supply boards with separate transformers, bridges and regulators, one for analog circuitry (+/-5V) and one for digital (+/-5V).

A/D Grounds are joint at each chip level as manufactures are suggesting: I unfortunately formed in this way multiple small ground loops due to the parallel DAC configuration, but this can be also seen as a Ground network plane, hope with some benefits.
Analog and digital signals are fully separated in the board. Each chip's power supply is filtered with a RCLC network (R damping= 1 ohm, C=22uF OS-CON, L= ferrite bead, C=100nF or 220nF ceramic for digital, polyester for analog).
That was the best I could design for 8 chips in 10x15 cm perfboard. I would not expect any digital noise at all. I'm just a little worry if the 1 ohm damping series resistors will badly influence the sound or dynamic (but without the simulation has shown a huge resonance peak in the 1MHz area!)
Sockets are mandatory as I need to select the eight DAC chips one by one for the lowest THD (PCM56-S)
Pics are four but board is just one.
Attached Images
File Type: jpg PARALLEL PCM56.jpg (599.5 KB, 99 views)
File Type: jpg BOT BOARD.jpg (640.4 KB, 99 views)
File Type: jpg TOP BOARD MARK.jpg (600.8 KB, 96 views)
File Type: jpg BOT BOARD MARK.jpg (700.7 KB, 93 views)
  Reply With Quote
Old 17th January 2019, 07:09 PM   #56
Ken Newton is offline Ken Newton  United States
diyAudio Member
 
Ken Newton's Avatar
 
Join Date: Apr 2003
Location: Eastern Pennsylvania
You weren't joking about using Perf board.
  Reply With Quote
Old 17th January 2019, 07:21 PM   #57
MarcelvdG is offline MarcelvdG  Netherlands
diyAudio Member
 
Join Date: Mar 2003
Location: Haarlem, the Netherlands
Quote:
Originally Posted by ygg-it View Post
Analog and digital signals are fully separated in the board. Each chip's power supply is filtered with a RCLC network (R damping= 1 ohm, C=22uF OS-CON, L= ferrite bead, C=100nF or 220nF ceramic for digital, polyester for analog).
That was the best I could design for 8 chips in 10x15 cm perfboard. I would not expect any digital noise at all. I'm just a little worry if the 1 ohm damping series resistors will badly influence the sound or dynamic (but without the simulation has shown a huge resonance peak in the 1MHz area!)
Instead of a resistor in series with the inductor, you could try an RC series network (or a cheap aluminium electrolytic with substantial ESR) in parallel with the capacitor. The frequency-dependent losses of the ferrite bead may also help to some extent, although most ferrite beads are still mainly inductive around 1 MHz.

Edit: I didn't realize yet that your caps are already electrolytics... Did you model their losses in your simulation?

Last edited by MarcelvdG; 17th January 2019 at 07:23 PM.
  Reply With Quote
Old 18th January 2019, 10:13 AM   #58
ygg-it is offline ygg-it  Italy
diyAudio Member
 
ygg-it's Avatar
 
Join Date: Feb 2008
Location: ITALY
Quote:
Originally Posted by MarcelvdG View Post
Instead of a resistor in series with the inductor, you could try an RC series network (or a cheap aluminium electrolytic with substantial ESR) in parallel with the capacitor. The frequency-dependent losses of the ferrite bead may also help to some extent, although most ferrite beads are still mainly inductive around 1 MHz.

Edit: I didn't realize yet that your caps are already electrolytics... Did you model their losses in your simulation?
Hi, the resonance in the MHz area to smooth with the 1 ohm resistor is formed by the ferrite plus 100/220nF capacitor, not by the big 22uF electrolytic cap.
  Reply With Quote
Old 21st January 2019, 07:01 PM   #59
nattawa is offline nattawa  Canada
diyAudio Member
 
Join Date: Apr 2010
Location: Mississauga, Ontario
I think the nut in the shell is to not allow digital current carrying GND conductors, be it power supply return current or digital signal return current, to become any part of the analog circuit. To achieve this goal we need to examine the physical current loops and make GND plane slits or separations accordingly. It is also important to place ceramic SMD bypass capacitors closets across the power pins and the associated GND pins of the chip, so that the capacitor shunts a current loop to its minimum physical size with respect to the higher frequency components in the current.
  Reply With Quote

Reply


Do you really tie Analog GND and Digital GND at pin chip level in DAC/ReceiversHide 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

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
What Chip is inside these 2 Digital (Toslink+Coaxial) to Analog (2RCA) Converters? spaceman5 Digital Source 7 29th September 2018 10:53 AM
shield ic chip at analog ground or digital ground? KevinLee Digital Source 2 17th July 2012 01:15 PM
Job Listing: Electrical Engineer, Analog & Digital Board Level Hardware Developer Aurasound Everything Else 9 12th March 2011 01:31 AM
Add analog outputs to cheap digital receivers stef1777 Digital Source 4 24th November 2006 08:52 AM
PCM & Clock for Digital Receivers? gevorg Class D 0 4th November 2006 08:50 PM


New To Site? Need Help?

All times are GMT. The time now is 09:39 PM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 15.00%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Copyright ©1999-2019 diyAudio
Wiki