[nullstring]

I was thinking of a PCM270X based DAC, where X > 2

Something like the HagUsb or the twisted pear DAC.
Both of them don't seem to use PCB space efficiently, so a new layout would need to be designed.

HagUsb is mostly non-SMD... and twisted pear has excess spaces for the terminal blocks (I think that's what they are called)

I noticed neither of them seem to use regulators.
Were those a special requirement for the PCM2702 DAC's? or do these not include them because they are made more for SPDIF than analog.

I also know that these DAC chips have lower quality analog output, so, I'm not sure how many people would actually want to use one.

holy hell, there is a fawn in my front yard.

 

[mugdecoffee]

Quote:

Originally Posted by nullstring /img/forum/go_quote.gif I noticed neither of them seem to use regulators. Were those a special requirement for the PCM2702 DAC's? or do these not include them because they are made more for SPDIF than analog.

The 270X chips need around 5V to run. The bantam has regulators to get the cleanest power since the2702 is used for its analogue outputs though 2702 can be powered right of the usb 5v supply. The twisted pear usb receiver is uses the straight 5V from the USB probably because its being used for its spdif output where clean power isn't as critical.

 

[error401]

I finally received the rest of my parts from Mouser last week, and today I got around to building up the Carrie board. I used the posted BOM in the 2nd post and pretty much ordered what was on there. I haven't had a chance to hook it up and test it yet, but I did have a few build notes. I haven't been following the thread, so sorry if these have already been mentioned.

- Having C1F/C3F and C1/C3 be different BOM items confused me. I installed the tant caps in the wrong spot and then had to desolder them. These should be renumbered.
- C1F and C3F don't fit
- The silk for C4 doesn't show the + marking, this should be moved so it's visible
- It'd be nice to have L/R/G markings on the J1 pads; I plan on using a panel mount jack
- This is really totally unimportant, but the silkscreens all have a slight 'staircase' effect on long straight lines, like the whole board was rotated a degree or two on a low-res display. Weird.

Overall though it was pretty easy to put together. I'm reminded how much I hate TH, but aside from taking forever I didn't have any real issues with assembly. We'll see how it sounds tomorrow when I build up a BantamDAC to go with

 

[joneeboi]

Thanks a lot for that, Keenan.

That's extremely helpful. I have some time this weekend, so maybe I'll add those changes.

I noticed the horrible silkscreening too. My EAGLE files have straight lines, so I don't know what Olimex was doing. I didn't realize also just how thick the lettering would be. I will try reducing the thickness and maybe get the boards fabbed at Imagineering. I'll add the labels to the output jack on the bottom of the board, and I may just get rid of the Molex connectors altogether. I mentioned in my PM that the filter caps would be too large for the allotted board position, but that was after you already ordered your parts. I'm also reconsidering the board-mounted jack. It's nice, but it takes up so much room. I don't know what exactly I could do with that space, but I could definitely tighten up the board if I did.

 

[DKJones96]

It should be noted, I mentioned earlier running the bantam without the output caps. DO NOT do this. The 2702, when active, creates a virtual ground at about 2.3v and this makes the amp run full 5v at the phones at anything past 1/4 volume.

I only measure 11mV of dc offset at full volume otherwise.

 

[joneeboi]

Did you connect the ground of the Carrie Amp with the 2.3V and connect the DCP input V- only to USB GND?

 

[DKJones96]

Well see, when I did my initial measurements I thought it was fine because I measured no offset on the bantam(I hadn't played a song yet) but if you pause the player in the middle of a song you'll notice that it goes to 2.3 and stays there until the end of a song where it times out and the output settles back to 0 again.

It doesn't keep a constant output voltage and with the behavior of that chip I can't really see any other way to connect them except through the caps.

 

[joneeboi]

Sorry for the super delayed post. I have been working on the amp, but I've just been busy moving back to Edmonton. I went through the schematic with a co-worker who specializes in Anyway, I tweaked the schematic a little bit and made some design changes. Here they are in no particular order:

- Added capacitors to match the peak output power. With a ~10V power supply with a rail-to-rail opamp into 32 Ohms, RMS power requirements demand (7.07V^2)/32 Ohms or 1.56W per channel. Of course, no one will ever be able to listen to anything that loudly with headphones, but I've changed the design to make sure the design lines up with itself. The capacitors will supply enough power for the power capabilities.

- Removed Molex connectors.

- Moved LED, RLED and L1L so that a separate input jack can be panel-mounted. There's enough room on top of the BantamDAC for a panel-mounted DPDT switch, so you can choose between external source and the Bantam.

- Rearranged the entire power supply. Filtering section has changed, renamed all the capacitors, and moved things around

- Changed the fonts so they show up more clearly when printed

I'm getting ready to produce the boards. Any last objections?

 

[Juaquin]

Quote:

Originally Posted by joneeboi /img/forum/go_quote.gif I'm getting ready to produce the boards. Any last objections?

Yes. This is too awesome. I request you reduce the awesomeness level by at least two notches.

Definitely in for one, if there will be a run of boards.

 

[rds]

Quote:

Originally Posted by Juaquin /img/forum/go_quote.gif Yes. This is too awesome. I request you reduce the awesomeness level by at least two notches. Definitely in for one, if there will be a run of boards.

Agreed on both counts

 

[mp101]

brilliant.

If you do agroup buy, I'd be intersted.

 

[cobaltmute]

Not sure how tall all you caps are (like CF1, CF2), but will you have clearance past them to get wires to IL,IR and IG?

 

[Billyk]

Most Excellent!
Please count me in on this for whatever is needed or just to purchase a board when they are ready.

 

[joneeboi]

cobaltmute:

All the caps are 15mm high, but wiring can be done underneath the boards. There is 2.00mm clearance underneath the boards, and you can fit up to 19AWG with the holes' inner diameter.

I am preparing to order from 4PCB.com's $33 each deal. With that comes a few limitations, those of which will be restricted to the prototype board. I PMed those of you who are interested, but I realize that if you're willing to pay up, we can get more than 5 boards. Silly me. Anyway, the more the merrier, so let me know if you want in.

 

[DKJones96]

Looks good, just a couple of comments.

Where is inrush current being limited at? The DCP0105?

While 1.56w per channel is nice, the USB port can't power that. Add the extra juice from a Bantam and you are over your USB specification by about 30%. That, of course, is if you run the 2w unit instead of the 1w unit that the original build calls for and even with the 2w unit you run it over spec at anything over 1w per channel(assuming a 100% efficient amplifier). Just something to keep in mind.