diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Digital Source (http://www.diyaudio.com/forums/digital-source/)
-   -   Dac Tips - grounding solution (http://www.diyaudio.com/forums/digital-source/208585-dac-tips-grounding-solution.html)

Stefan0815 11th March 2012 11:43 AM

Dac Tips - grounding solution
 
3 Attachment(s)
Two years ago, i bought a DAC PCB with AD 1955 and AD 1896. With my knowlage at this time, the board looked perfect desinged:

Every chip has it's own little voltage regulator backside of the pcb which are supplyed by LT 1084 Regs. All the voltage regulators were "devided" by a little inductor. Good parts, a discrete regulator for the analogue stage and all the stuff what you need to get your boards well sold on ebay.

A few day's ago the AD 1955 chip burned out and with a new chip it is not running (all voltages were ok!). Finally i ruined the PCB at the bottom of the AD 1955.

As i locked to the shematics and the board again, i noticed the "strange" grounding. As you can see in the Schematic, the Digital Ground is connected with the Analogue Ground. When you look to the pcb, you see that both are not direct connected, the two way's the current can flow to the grounding are different (As you can see in 5) but there are still connected! What do you think about this way grounding?

The second thing is the connection between the AD 1955 grounding and the I/V and buffer stage. The only connection is through an inductor in the Powersupply ( 4 and 1). The whole analogue stage has its own groundplane at the top of the pcb and the grounding at the bottom.

I think there must be a direct connection between analogue ground of the AD 1955 and analogue ground of the IV stage. I know that the IV stage is differential, but the CMRR of the opamps is limited and it would't be good if there is much noise and a little bit of hum bettween those two grounding points. I had always a little bit of humm (100Hz) in the signal, but i don't know if it was caused by the powersupplylanes laying next to the pcb. The carging currents of the big caps must be very high.

In 3 You can see my try to seperate the analogue ground of the AD 1955, but i don't know if this kills the dac finally or the failor before.

So should i buy this board again? Or should i swap to another diy pcb or another ebay PCB?
I realy like the desing, but as i know now is, that those chinese guys who desing make big mistakes sometimes (i remember a totaly wrong connected regulator on the early CS 4398 boards). And the mistakes were always in the details (on my CS 4398 board was no division between analogue and digital powersupply).
If i had the money, i would by a Benchmark Dac ;-(

Chris Daly 11th March 2012 12:33 PM

Try to identify in Ohms what is the difference if any between the ground areas, ground areas looking different or separate may just be conventional star grounding which is good practice.

Cheers / Chris

Stefan0815 11th March 2012 01:06 PM

With my normal DMM it is not possible to notice any difference between the grounding points.I think the differences are smaller than 100mOhm.

Chris Daly 11th March 2012 01:28 PM

So its star grounding,of sorts one area by one area. Next thing would be to check all of the LT 1084 Regs
LT1084 - 7.5A, 5A, 3A Low Dropout Positive Adjustable Regulators - Linear Technology are up and running.

Re large caps,horrible! my approach would be to use transistor based cap multipliers.

Cheers / Chris

Stefan0815 11th March 2012 01:46 PM

I think it's not worth investing to much time in this PCB. Every AD 1955 costs 12€.
The problem is now, that there a a few broken iron points, because I desoldiered the AD 1955 twice.
The Audiodata went into the AD 1955, the bias of 2,5V is measurable at the outputs but the AD 1955 went very hot.
The whole board costs around 100€, so i won't invest to much time in finding the failure (it's more fun to work with my Revox A77 instead). I can seperate the powersupply and use it in another project.
I don't think it is a stargrounding because of the desing at the bottom.

Am I right saying, there must be a direct connection from AD 1955 ground to the I/V ground (not the long trace in green?)

Chris Daly 11th March 2012 02:36 PM

Quote:

Originally Posted by Stefan0815 (Post 2941687)
I think it's not worth investing to much time in this PCB. Every AD 1955 costs 12€.
The problem is now, that there a a few broken iron points, because I desoldiered the AD 1955 twice.
The Audiodata went into the AD 1955, the bias of 2,5V is measurable at the outputs but the AD 1955 went very hot.
The whole board costs around 100€, so i won't invest to much time in finding the failure (it's more fun to work with my Revox A77 instead). I can seperate the powersupply and use it in another project.
I don't think it is a stargrounding because of the desing at the bottom.

Am I right saying, there must be a direct connection from AD 1955 ground to the I/V ground (not the long trace in green?)

The schematic shows Dgnd for AD1955 is with a chassis type symbol, different to other ground.... but other gnds are probably also leading back to Chassis gnd eventually. Part U10 needs to be researched to ask What is it ? and analog Agnd appears to be treated differently I would say because of the mute relay switching in and out, and also referencing to the PSU components related to I/V op amps.:)

Stefan0815 11th March 2012 03:21 PM

The Schematic is to be handled with care. On the PCB there is not a big difference between AGND and Chassis ground. The area around the PCB name is DGND and connected directly to one of the AGND Ground.

If i would have desinged the PCB i would connect AGND and DGND near to the AD 1955 with a little inductor and not via a long pcb trace as in 4.
Do I miss something? For me the desing seems to be wrong!

U10 and so on are all little 1117 Regs.

SoNic_real_one 11th March 2012 03:31 PM

1 Attachment(s)
Don't just belive all stories that are vehiculated here, read what real engineers are recomending (manufacturer). Don't add any inductors or resistors between those grounds! Don't drag them all the way to chasis with wires (inductors)!
Digital ground and Analog Ground pins of the DAC need to be connected as close together as possible at the DAC cip analogue plane (on the same PCB):
Analog Devices | Solving mistery of AGND and DGND
Read all, including the italics warning notes at the end!!!
Or for ADC (similar concepts):
Analog Devices: Analog Dialogue: Ask The Applications Engineer - 12

Stefan0815 11th March 2012 03:59 PM

Thank you very much for those pdf's!
Stefan

Ken Newton 11th March 2012 07:00 PM

I suggest that you not worry yourself about such design intricacies as specific grounding scheme. If you enjoyed the way your DAC sounded in your own system before it failed, then that's all that really matters. All else, except for price and reliability, is academic, IMHO.

As has been said, the proof of the pudding is in the tasting. Not in the recipe or specific preparation.


All times are GMT. The time now is 08:48 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