[Clutz]

Haggerman suggested that instead of going with multiple reg101's, we use one reg101 for a 3.3 volt psu, one reg101 for a 5 volt psu and do a good job of local bypassing at all of the power supply connections.

The only real downside I can see of this is it may make the routing of voltage traces more difficult. I'm going to try and start laying out the schematic.

 

[ble0t]

I already have a schematic almost fully done as well as a start on a layout if you'd like me to post it. I've been designing it to fit in a Hammond 1455J1201BK enclosure (approx 4" x 3" x 1").

 

[00940]

Bleot : please post the schematic (in .sch and .jpg), so we can start discussing it.

Power : I prefer to have different regulators for all chips. It makes it easier to keep the sections independent and separate. Btw, what are we using for supply ? One single supply and buffered ground ? Or dual supply ? AC or DC ?

Layout : how small do we go for the smd parts ? 1206, 805 ?

 

[SnoopyRocks]

Things seem have progressed a bit since I last checked in.

I take it no one else uses Pads2005. What PCB software is being used?

I prefer 0805 (or even smaller) caps for chip supply decoupling. Otherwise 1206 is fine. Is the plan to make this entirely SMT?

For the power supplies, I'm in favor of requiring a separate dual DC supply (e.g. +-12V) from which all other voltages can be derived on board. The alternative would be to have a single DC supply to generate the other voltages, but would require AC output coupling and low voltages opamps.

 

[Clutz]

I'm with Snoopy on this. 0805 for decoupling caps. 1206 is just too large when you've got to have a bunch of them in close proxmity to each other - just too hard to work with. I also agree with requiring a +/- 12 volt DC input PSU required.

 

[MegaMeee]

sounds good to me.

 

[ble0t]

OK...just let me clean them up a bit

EDIT: OK...I have a .pdf made and I have the .sch (I use EAGLE) zipped up, but both are too large to post on the forum

Anyone related to the project have some web space to post on?

 

[Clutz]

I have crappy webspace I can put it on, on my own server - yes. It's nothing fancy, but it'll work.

I'll PM you ble0t.

 

[ble0t]

YGM

 

[Clutz]

Come and get it.

http://twtch.zoology.ubc.ca/

 

[00940]

Ok, got the schematics.

- good idea for the usb esd protector.

- it'd be best to have resistors on the I2S lines between each chips

- I wonder if it is good to have the ferrite before the electrolytic caps. Considering the electrolytics will be further away from the chips, we allow high frequencies to wander a bit. Wouldn't it be better to have the ferrite in between the electrolytic and the ceramic ?

- On the pcm1798, I would connect vcomr and vcoml to +5V and not to VCC1, through its own ferrite.

- We need an hardware reset of the DIR1703.

- you cannot use the REG101 straight from the supply. Its max input is 10V. We need preregulation. I would use one LM317 at 7V for all digital chips and a pair of LM317/337 for +/-9V for the I/V and the analog supply of the PCM1798.

Looks fine otherwise.

 

[ble0t]

Quote:

- good idea for the usb esd protector.

Thanks, it works nicely on my PCM2704 DAC I just got finished.

Quote:

- it'd be best to have resistors on the I2S lines between each chips

Yep...I've been working on getting the PCM2707 and DIR1703 (now that we've decided on it) sections up and going before anything else.

Quote:

- I wonder if it is good to have the ferrite before the electrolytic caps. Considering the electrolytics will be further away from the chips, we allow high frequencies to wander a bit. Wouldn't it be better to have the ferrite in between the electrolytic and the ceramic ?

It's not a big deal either way. I'll move them inside as per your suggestion.

Quote:

- On the pcm1798, I would connect vcomr and vcoml to +5V and not to VCC1, through its own ferrite.

Isn't that what I did?

Quote:

- We need an hardware reset of the DIR1703.

No problem...I'll add an RC section to that.

Quote:

- you cannot use the REG101 straight from the supply. Its max input is 10V. We need preregulation. I would use one LM317 at 7V for all digital chips and a pair of LM317/337 for +/-9V for the I/V and the analog supply of the PCM1798.

Ahh...thanks for noting that. I'll put an LM317 section to drop it down to 7V before feeding it to the REG101's.

Work has been a bit hectic lately, but hopefully I'll get some time to improve upon it over the next couple days. Thanks for the suggestions!

 

[00940]

Bad news gentlemen : http://www.diyhifi.org/forums/viewto...=288&start=120

 

[ble0t]

Yep...saw that this morning as well.

But, even given that information, what choice do we have if we would like to use USB as an option? We're including SPDIF and implementing an ASRC (running from a different master clock) to help the situation as much as we can.

 

[Clutz]

This is definitely disappointing news. Does this leave us with any other options? Are there better USB to SPDIF chips available? I can't imagine another TI chip performing substantially better if they're all using SpACt. Is it time to go and search the other semiconductor companys for a different USB receiver?

I suppose alternately we could then think about it as an SPDIF input DAC that also has the benefit of having a USB receiver with it?