-
Chip change
08/23/2021 at 02:02 • 0 commentsIt turns out that the receiver chip I chose for the first versions - the STA-120 may be nearing obsolescence, and thus may be difficult to find going forward. Fortunately, there's another option - the DIR9001. It's not at all compatible, but it does have the virtue of coming in a TSSOP-28 instead of a SOIC-28, so it's smaller. Of course, that doesn't benefit us a whole lot seeing as how the board is now the 38x59 mm size for the chassis. But if it turns out that the STA-120 becomes unavailable, at least there is a second choice.
Of course I can't test that design right now because they're backordered until next year.
-
Future options
08/23/2021 at 01:56 • 0 commentsThe idea of perfecting this project is rapidly consuming me.
Making the board slightly bigger (38x59 mm) makes it fit very nicely into a Hammond extruded aluminum case, just like my GPSDO. And like that project, I can probably design a pair of PCBs to fit into the ends with the correct perforations for the jacks and the error LED light pipe. I still can't match the Amazon price point selling them on Tindie, but I pretty much put that down to volume. I can get them down to a retail-plus-free-US-shipping price of about $20 for the board, $20 for the case, and $10 for a power supply. That's at least not Fred Falvy levels of price foolishness.
-
Making it a little cheaper
08/22/2021 at 23:17 • 0 commentsThe TORX1355s are super expensive. Turns out that there is a much cheaper alternative. The PLR237/T10BK is more like $1.50 at Q:10. They also have 25 MHz of bandwidth (overkill for the CS4334 DAC and the STA120, which can only do up to 96 kHz audio - even at 64 bits per sample that's only 6 MHz). It works at both 3.3 and 5 volts, but they suggest adding output capacitance at 5 volts, so running it at 3.3v saves us a part. I've also taken a shine to USB C of late, so swapping out the barrel connector for USB C isn't a bad idea. Since we only want 5 volts at way less than 5 watts we can just use 5.1 kΩ CCx pull-down resistors.
Lastly, it may not be a bad idea to add an LED for the VERF pin and a light pipe to aim it out the front.
-
One more change
08/18/2021 at 00:09 • 0 commentsIt turns out that correcting the mode of the STA120 was not enough. The audio sounded way distorted. The music was identifiable, but it was very, very fuzzy. I tried playing with the !DEM/SCLK pin, but that didn't change anything. The steady state of the de-emphasis output from the STA120 was off (high) anyway, so on a whim I hooked !DEM/SCLK up to the SCLK pin of the STA120 and that made the audio perfect.
So 3 total changes: adding a series inductor to the supply pin of the optical module, fixing the mode pins on the STA120, and hooking the SCLK pin between the two.
I'm going to order one more set of boards from OSHPark and call it a day.
-
Good news, bad news
08/17/2021 at 22:39 • 0 commentsThe boards came back and there are two errors (so far).
- The optical receiver I sourced is only good to 500 kb/sec. The optical audio stream I’m trying to decode is way faster than that. I’ve ordered some of the 10 MHz variant. The datasheet for those says that they recommend a series inductor on the supply pin, so that’ll be a change to the board I need to make.
- I made a mistake wiring the mode pins up to the decoder. I’ve asked for mode 4, but I need mode 2. So that too will need to be a board revision.
I was able to test this by poaching the optical output stream from my Amazon TOSLINK decoder box. It sorta worked, but the L/!R signal was way wrong, so I will need to lift and rewire M1 and M2 and try again. If I see correct audio on the scope, then I will feel good about my chances.
EDIT: Upgrading the optical receiver to the TORX1355 fixed the first issue, and lifting the two pins and using bodge wires fixed the second.