[ble0t]

OK, here's (hopefully) my final revision before attempting a proto. I've tried to clean things up more and I also mitered all the corners and made sure the silkscreen is good to go. As always, comments/suggestions appreciated

Schematic:

Layout:

 

[guzzler]

Your C15 should be going to ground, not in series with the Vcc trace. Won't work as it stands! Inductors on the USB data lines aren't a terribly good idea as they will blunt the pulses, possibly leading to errors. Resistors are sufficient to isolate the PCM. I would recommend not having the Dout trace crossing the crystal trace as well; the crystal trace is the most sensitive on the board

 

[ble0t]

Thanks for the reply!

The C15 label actually applies to the large electrolytic cap, not the bottom layer. Sorry for the confusion there...that bottom layer is actually a ferrite. As for the ferrites in the data lines, I guess I could try it with them jumpered and with them in and see what happens for now. Finally, I think you're right about the DOUT line...I never liked crossing the clock signal with that for fear of contamination on either of them. I'll see what I can do to rework that section. Thanks!

EDIT: I found these for possible use on the data lines. Sourcing could be difficult if they don't sell in small quantities (they only sell direct). I also found these which would not only provide me with the 22R resistors, but also with the pullup for D+ and some filtering.

 

[ble0t]

OK...here's the final edit before I submit today for a proto. I added the small USB termination/filter IC that I linked to in my previous post and also tried to improve some of the routing.

Schematic:

Layout:

 

[DaKi][er]

There is plenty of room to have the digital output go up the left hand side of the chip, then put the resistor across the crystal there on the bottom layer and move the crystal in as close as you can

 

[halman]

You might also want to tie the ´negative´ end of the transformers output to ground via a capacitor, to help reduce emi.

 

[ble0t]

Ahh...both good ideas. I put in the order for the proto before implementing either of them, but I will most likely make the changes for a future version.

Oh and DaKi][er...thanks for that link (not sure when/where it was posted) for BatchPCB. It is hands down the best place for this type of thing

 

[ble0t]

Just a quick update/FYI for anyone interested...

I contacted TI tech support concerning the output impedance of the DOUT pin on the PCM2704 and also asked about the PCM290* as well. Here's what they had to say...

Quote:

We do not have maximum value for this spec, but the estimated maximum values are +/-30% of 50ohms. This is the case for the PCM27* and PCM29*.

Doesn't narrow it down too much, but it's better than nothing I suppose...

 

[halman]

You should buffer it and use something with a known impedance anyhow...

imho

 

[ble0t]

Good point. I guess I was wrong to think that TI would have some more concrete specs than that

 

[halman]

It might be so that the impedance is not equal for transitions from 0->1 vs 1->0. In wich case you _really_ should use a buffer for it.

 

[ble0t]

Got the board today along with the MAX4410's I ordered from Maxim. I'll try and get some shots posted later tonight. Hopefully, the rest of the parts will be here tomorrow and I can get it built by the weekend. Woohoo

 

[nleahcim]

Quote:

Originally Posted by ble0t Ahh...both good ideas. I put in the order for the proto before implementing either of them, but I will most likely make the changes for a future version. Oh and DaKi][er...thanks for that link (not sure when/where it was posted) for BatchPCB. It is hands down the best place for this type of thing

batchpcb is very very slow - often taking more than a month. 4pcb.com does much better boards of pretty much any size in a week for $33 if you're a student. (and they actually know what customer service is too)

Anyways - one thing that bothers me about your design: where are the thermals? I expect you'll have some trouble soldering some of those parts that are connected to the groundplane - it'll just act like one big heatsink.

 

[chaostic]

Bit off topic, but did you get asked by ti to verify that you are american when you asked about the 290*? I did when I asked about the test circuit they had.

 

[ble0t]

nleahcim: I ordered from BatchPCB on Jan 15th and got my board yesterday. Two weeks for double sided soldermask and silkscreen isn't bad turn around if you ask me

As for the thermals, I didn't have any issue soldering on Alf's board and I wouldn't expect much here. It's a question of 'local' heat, and the groundplane isn't thick enough and/or big enough to dissapate enough heat to prevent soldering.

chaostic: No, they never asked me anything like that. When I submitted my request, they asked me (in a web form) for my name, address, etc.

Lastly, here are links to some scans of my board if anyone is interested...I hope to get some soldering done tonight

Top
Bottom