I'd say to keep the digital ground plane separate from the analog ground plane(like do not run them on top/bottom of each other) but just used as a shield/to-contain the relay signals. Think of it as a means to couple any emissions into the digital ground and not radiate to others. Another way is to have the two ground planes cross but only at right angles and a small area in which to couple into each other.
The other way is to do it like i did on the portable, a 4 layer with a dedicated power/ground plane and fills on top/bottom too so the whole pcb is one big common point ground. in your case because of remote power supply(not on the pcb) you want to keep these ground returns separated.
The other way is to do it like i did on the portable, a 4 layer with a dedicated power/ground plane and fills on top/bottom too so the whole pcb is one big common point ground. in your case because of remote power supply(not on the pcb) you want to keep these ground returns separated.
Wonderful job , Jeff !! Looking over the PDF , I see you have every digital/
analog input covered.
U341/342 are the toslink , right ?
Your feature set is akin to the XDA-2 emotiva preamp/dac. I HAVE NO IDEA
how they can sell it for $269USD !! Chinese slave labor ?
They use the PDxxxx supplies , TI dac , OPA op-amps - same deal.
Analog is IC version of my line stage. NO Jung regulators , just TI IC
ones.
WOW , does diptrace output the PDF. Kindhornman just sent me the
software.
OS
U341/342 are Toslink. With the LCDuino front end only U341 will be available to use with the Arduino software as is. I plan to use only one digital input on mine but the other inputs will be connected for use if wanted.
Anything will output to PDFs. Install Foxit Reader(just say No to the Ask Toolbar crap) and it gives you the option to print anything to a PDF.
It's all about Z. The buffer outputs most likely less sensitive than the dac -
buffer inputs.
It is , however - best to keep separate.
Very high Z (like the OPS predrivers) , actually needs physical distance or
a "guard trace/track" of clean earth or regulated voltage.
OS
I'm rethinking the layout a bit. I may move the serial expanders up onto the daughtercard and rearrange some components a bit. If I put all control parts up on the daughtercard they can be swapped out for whatever anyone likes for controls. This also would free up some pins for more options on the main board. and possibly give me some more room to separate the digital components from the analog better.
When I have this issue , I look to an OEM with a similar featureset.
Emotiva XDA-2 USB DAC/Digital Preamp/Headphone Amp Page 2 | InnerFidelity
I pieced together a few photos , hard to find a real closup of the
whole Emotive PCB. They use a larger board , the row of heatsinks
are all IC regulators for both the digital and analog.
headphone out/line out is LMExxxxx. Cirrus logic cm6631 USB/ad1955dac.
OPA/5534 dac output buffers. Same as the pitchfork.
I'm surprised you actually have multiple analog inputs , as well. XDA is
just toslink / analog out.
So , they have it split to a - (below photo).
- big 10 X 4" PCB (usb/dac/relays/attenuator/line buffer/regulators).
- Unregulated PS PCB with just RF/common mode filtering (diodes/caps/ac filter only).
- Micro is the small board all by itself (near the main trafo) , it gets its power
from the unregulated pcb and most likely has a good reg. onboard.
- The long thin front PCB with user controls/display.
I thought the XDA would make a good comparison . It reviews well ,
the designer is a real person who loves music (and a good price).
Strange how the pitchfork evolved so similarly.
OS
Attachments
I've come across a couple posts now saying the WM8805 can be a pain to run in software mode. In hardware mode it automatically detects what frequency it's receiving signal in. When you switch it to software mode it becomes dumb and you need to tell it what frequency it is. Sounds like another good reason for one digital input.
Rsavas, have you had any experience with this IC? Is this an issue we should be concerned with?
Rsavas, have you had any experience with this IC? Is this an issue we should be concerned with?I've got a revised layout with all the controls moved onto a plug on card. The control card is pretty dense. I had to go four layer with some parts on the bottom. It will likely be much less crowded with more advanced controls.
Attachments
From what I understand, the datasheet isn't much help on the subject either. People just use the WM8805 in hardware mode with single input. It's cheaper than the WM8804 and works well. There was some other similar devices suggested. I'll look into it more.
You mean fun in a good way or should I be looking for another chip? They're actually cheap enough to be able to run a couple in hardware mode and just select one through MC74VHCT244AMs.
"Fun in a good way"
When you are reading about others having difficulties programming this chip, are they using SPI or i2c mode?
I do not believe this statement to be true,
"In hardware mode it automatically detects what frequency it's receiving signal in. When you switch it to software mode it becomes dumb and you need to tell it what frequency it is."
When you switch to s/w mode the default settings are not the same as in h/w mode, would be more accurate. Thus you must configure the chip correctly. With 40 registers it is not simple, requires one to fully understand the chips operation.
I do not believe adding more WM8805 chips is the answer to be able to have more than one spdif channel!
Edit, well it is the answer, if you want to run without a controller that is.
When you are reading about others having difficulties programming this chip, are they using SPI or i2c mode?
I do not believe this statement to be true,
"In hardware mode it automatically detects what frequency it's receiving signal in. When you switch it to software mode it becomes dumb and you need to tell it what frequency it is."
When you switch to s/w mode the default settings are not the same as in h/w mode, would be more accurate. Thus you must configure the chip correctly. With 40 registers it is not simple, requires one to fully understand the chips operation.
I do not believe adding more WM8805 chips is the answer to be able to have more than one spdif channel!
Edit, well it is the answer, if you want to run without a controller that is.
Last edited:
From what I've been reading we would need to externally measure input signal frequency and tell the WM8805 what that is. Here's a quote from Linuxworks in another forum.
"it does not detect the samplerate in software mode since the wolfson spdif receiver is braindead in software mode and this is why almost everyone takes the easy way out and uses hardware mode on the spdif rx. note, this has nothing to do with the dac chips; its wolfson 8804/5 receiver that causes all the software control problems. so far, I have not heard of anyone doing a software control of this (sort of) nice spdif rx chip. it can't easily tell 176 from 192 and that's its core problem (in software mode). but in hw mode, it locks onto everything and its wordclock out is accurate. what you would have to do (which most people would choose not to) is to freq-count the wordclock in software to 'know' the real samplerate, then use that info to program the dacs.
I have arduino code that does this (which I plan to use in a later spdif switch project) but there's no simple direct way without using an extra controller, which everyone admits IS a drag."
Yes, that's what makes it a contact sport!
I'm going to recheck the schematics a couple more times, then order some boards. I've managed to rework the control board onto two layers. It just seems wrong to pay more for that little board than the big DAC board.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.