another one, Chinese design CS43122 DAC, what do you think?
Nov 23, 2005 at 3:54 AM Post #16 of 34
Quote:

Originally Posted by jerb
do the chips come presoldered?


I have no idea...
anyone explain the bad things mentioned by glassman, please?
 
Nov 23, 2005 at 4:34 AM Post #17 of 34
Well, this DAC does use 4x OPA627, that should be worth the price of admission right there
biggrin.gif
 
Nov 23, 2005 at 5:53 AM Post #18 of 34
Quote:

Originally Posted by Teerawit
Well, this DAC does use 4x OPA627, that should be worth the price of admission right there
biggrin.gif



well you didn't buy the kit for the parts, did you?
biggrin.gif

would anyone explain to me why the design is good or bad?
 
Nov 23, 2005 at 9:31 AM Post #19 of 34
The PCB design is essential to the performance of the DAC, and frankly this one does just about everything wrong

It is not just a matter of putting audiophile parts on a PCB, that will only get you part of the way there
 
Nov 23, 2005 at 9:33 AM Post #20 of 34
Quote:

Originally Posted by Glassman
very poor design decision placing DAC and receiver on socketed adapters.. no decoupling and the PLL loop filter network too far from the chip.. well at least the PCB in red looks fantastic
tongue.gif



I'll agree to Glassman,. It looks fantastic but I'm a little worried about EMC. This design may be a very good noise transmitter in the band > 100-500 MHz. Putting things on sockets are also a typical no-no.

Input wires are tied together with the outputs. Not to recommend. It's a virtue to keep inputs clearly separated from the outputs and also analog circuits well separated from the digital circuits.
 
Nov 23, 2005 at 9:40 AM Post #21 of 34
Nov 23, 2005 at 12:19 PM Post #23 of 34
Quote:

Originally Posted by diablo9
you mean the DAC chip should be directly soldered on the board? but people always socket the op-amp right?

anyone else could explain why the 2nd part could be bad?
confused.gif



The datasheets of such digital chips are extremely clear. Those are high frequency devices and for optimal operation, you must place your decoupling capacitors as close as possible from the pins of the IC. Which is near impossible with sockets and only big caps (good smd ceramic capacitors are a must here). Btw, layout has become very important for the last generation of opamps too. Some don't like sockets much. If the designer of this dac is not even able to understand that, I'm affraid he messed up the whole grounding and alimentation scheme. I'm not gonna throw him the stone, I did some big mistakes in my first pcb layouts too.

I cannot say this will sound bad. The schematics are well thought, the parts are quality parts. But they are implemented in a very poor fashion. Considering the price of the kit, I could buy it and design my own pcb for it. Would still be interesting pricewise and probably better sounding. If you can get the chips without sockets of course.
 
Nov 23, 2005 at 4:42 PM Post #24 of 34
Quote:

Originally Posted by diablo9
you mean the DAC chip should be directly soldered on the board?


Yes.

Quote:

but people always socket the op-amp right?


When you are only passing a 20 kHz bandwidth signal through a fast op-amp, you may often get away with cutting certain corners. Those same compromises may bite you in the butt if you were to use the op-amp's entire bandwidth. And since any strong RFI source is enough to effectively use your op-amp's bandwidth, sometimes this amateur equipment malfunctions, and most of the time the audiophile never finds out why.

It's like the old story: The patient says, "Doctor, it hurts when I bend my elbow like this." And the doctor replies, "So don't do that, then." That advice is good enough for audiophiles when their equipment misbehaves in odd situations, most of the time.

You cannot take such a laconic attitude when using a fast digital chip. This particular DAC chip wants up to a 42 MHz master clock. You have no choice but to work in the land of high bandwidth here. The chances of failure from ignoring good decoupling advice are much higher.

Quote:

anyone else could explain why the 2nd part could be bad?


I'm not sure. A PLL is a very fast circuit (necessarily so) so perhaps it's an RFI radiation question. The longer the trace, the less effective the decoupling network.

Another thing that worries me about this design is that the receiver chip (CS8414) is not listed on Crystal/Cirrus's web site. I don't even get an obsolescence message.
 
Nov 23, 2005 at 5:28 PM Post #25 of 34
Quote:

Originally Posted by tangent
Another thing that worries me about this design is that the receiver chip (CS8414) is not listed on Crystal/Cirrus's web site. I don't even get an obsolescence message.


It is in the dropdown on the main page (look for cs8413/14). Here's the datasheet: http://www.cirrus.com/en/pubs/proDatasheet/8413-4.pdf

They do say not to use it in new designs, but it is still a current chip.

AudioNote, Pass Labs, and just about everyone else puts the cs8414 on adapter boards and drops it in a socket. The reason is is that it is a drop in replacement for the cs8412 which is a DIP-28 chip, so if there really was problem with this DAC, you could siply pick up a cs8412 and use it instead. The 8414 will receive higher sampling rates, but it is fully backward compatable, which seems to encourage adapter boards and sockets. Thus, lots of people offer it as an upgrade. Also, I don't recall the datasheet saying not to do this. I can't speak to the DAC chip itself, but the receiver is fine. Indeed, I have done it too in several DAC's (Cal Audio, 1543 based, etc.) with no discernable problems.

Also, the opamps used in this project are not high speed ones. They are opa627's, so they should be fine as well.

-d
 
Nov 23, 2005 at 5:40 PM Post #26 of 34
it's not a question of working / not working, but how well it's working.. $30 DVD player plays CDs fine, but is that really what we want?

also, when designing the adapter, there should be pads for ceramics.. then it's fairly ok, but I haven't seen any..
 
Nov 23, 2005 at 6:01 PM Post #28 of 34
Quote:

Originally Posted by Glassman
it's not a question of working / not working, but how well it's working.. $30 DVD player plays CDs fine, but is that really what we want?

also, when designing the adapter, there should be pads for ceramics.. then it's fairly ok, but I haven't seen any..


I understand what you are saying, but really, if it is good enough for Peter Qvortrup and Nelson Pass, it's good enough for me too.
 
Nov 23, 2005 at 8:17 PM Post #30 of 34
well, it would be better to design it in such a way that the return currents for the different sections (Pll = analog supply, digital supply = roughly the rest) of the cs8414 don't mix too much.
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 2 (members: 0, guests: 2)
Back
Top