Anyone know a good audio DAC breakout board?
- Thread starter bigpilot
- Start date
- Status
PaulStoffregen
Well-known member
Constantin
Well-known member
Well, they helpfully posted the schematic. If the design is is what you need, simply recreate it on your own. A bit more money than $14 (just the stencil will set you back more than that, never mind the chips + the board) but a good learning experience.
PaulStoffregen
Well-known member
Ah, I see. You didn't mention audiophile grade! The WM8711 is most certainly NOT an audiophile level part. Yes, it's made by Wolfson, but they make many different parts ranging in performance. The WM8711 is a low power part with modest performance. In fact, its specs are pretty much the same as the SGTL5000 we use. Both are meant for low power portable devices.
I actually bought a couple Wolfson WM8740 chips some time ago, with the intention to eventually make a top quality DAC breakout board. My understanding is the WM8740 is among the very best chips DACs they make. PJRC probably won't ever turn that into a product, but if I ever do get around to making one, I'll certainly publish the PCB and construction details so you can make it yourself. Then again, you could just get a WM8740 chip and lay out your own PCB. The WM8740 datasheet has very detailed recommended circuits (using premium/expensive opamps).
Wolfson does make a breakout / eval board for the WM8740 chip. It has S/PDIF input to a CS8427 chip that converts to I2S. The I2S signals go to a header with jumper blocks. I believe you could just remove those jumper blocks and feed the I2S signals from Teensy 3.1 (or anything else that creates digital audio with standard I2S output) to the right hand side of that header.
Of course, Wolfson's board isn't cheap. Mouser has 2 of them in stock. I hope you're sitting down before you enter this part number on Mouser's site and see the price. It's part number 238-WM87406054DS28EV.
PaulStoffregen
Well-known member
My main curiosity to build a WM8740-based board is doing some double blind listening test to find out if anyone can really tell the difference. It may be years until I ever get around to doing such a test, but I really want to do it someday.
If you ever make such a board and do a comparison, I'd be curious about the results. But $281.30 might be considered a bargain in the "audiophile" category, when you can pay 10x that for speaker cables... http://www.pearcable.com/sub_products_anjou_sc.htm
Fair question. The Audio board product page has no information on the line-in and line-out voltage levels or impedances. (I was going to point you to the info on that page, but it isn't there).
From the data sheet, at 3.3V supply voltage the Line-in is a max of 2.83 V peak-to-peak with 100kΩ input impedance (Table 4, although Table 6 says 10kΩ) and the Line-out at full digital output level (0dBFS) is 2.53 (min) 2.61 (typ) 3.11 (max) V pp with 320Ω output impedance. Load on Line-out must not be lower than 10kΩ. That seems to imply extremely limited output current (0.3mA).
If a suitable bipolar power supply is available, a simple op-amp unity gain buffer would allow connecting to a wider range of typical loads. The voltage is sufficient for consumer line-level input.
Last edited:
I recall Paul saying the board follows the datasheet example circuit, which as I recall used 1 uF caps to couple the output. If your output impedance is lower than 10k, and you don't want DC current flowing through your load, you would need to either (1) use larger output caps, which might be physically large as well and/or (2) use a ground reference different from the on-board ground. Assuming you don't want to deal with a bipolar power supply or transformer coupling.
Jp3141
Well-known member
Freescale has updated their datasheet for this part -- LINEIN input impedance is now typ. at 29 kohm (Table 4, also Table 6) based on bench measurements.
MIN input impedance is now lower also -- 2.9 kohm
http://cache.freescale.com/files/analog/doc/data_sheet/SGTL5000.pdf?pspll=1
MIN input impedance is now lower also -- 2.9 kohm
http://cache.freescale.com/files/analog/doc/data_sheet/SGTL5000.pdf?pspll=1
PaulStoffregen
Well-known member
This is near the top of my to-do list.
The levels are actually adjustable over a significant range. Right now, I believe the defaults in the beta test code are (unintentionally) much more than consumer line level. My intention is to review these in the next few days and get the defaults for all in and out, except the PWM, as close to consumer line levels as reasonably possible.
A function will probably also get added to the SGTL5000 control object, so you can choose any level the hardware is capable of providing, and eventually that function will be well documented.
- Status