The voltage levels that the XMOS is outputting for the I2S signals versus what voltage level I2S signals the 1305 needs in order to work, which depend on the 1305's working voltage.
I've gone through your experiments reported in this thread. It would be a bit too much work for me to go through the other thread too.
What I am seeing is an XMOS device not working with a 1305. This, according to the datasheets, is to be expected, due to the mismatch in voltage levels.
I would start troubleshooting by first fixing the problems I am aware of, namely the voltage levels. Look for a fast level shifter IC. In fact, pretty much any level shifter (74AHCT125 ?) should be fast enough for 44.1K I2S signals.
Or you could kill two birds with one stone by using an isolator IC such as the Si8641 which will do both galvanic isolation and level converting in one go.
I've gone through your experiments reported in this thread. It would be a bit too much work for me to go through the other thread too.
What I am seeing is an XMOS device not working with a 1305. This, according to the datasheets, is to be expected, due to the mismatch in voltage levels.
I would start troubleshooting by first fixing the problems I am aware of, namely the voltage levels. Look for a fast level shifter IC. In fact, pretty much any level shifter (74AHCT125 ?) should be fast enough for 44.1K I2S signals.
Or you could kill two birds with one stone by using an isolator IC such as the Si8641 which will do both galvanic isolation and level converting in one go.
The OWON benchtop DMM automatically displays what it detects on the probes. In post #116, the display was always presented in Vdc.
I repeated measurements, in various configs, using my trusty and reliable Fluke 87. This produced similar results for same experiment on the Fluke 87 -- no surprise.
I also performed additional experiments in other configs. Photographed most of them, but too many to post -- gets confusing.
Something addiional I did discover is that Vdc at various digital pins increases by about 1vdc when the USB cable -- coming from PC to Chinese pcb -- is disconnected. Only the DATA pin drops (to about 0vdc). Both the OWON and Fluke DMMs show same behavior.
Volt levels of four digital pins (OEM board, no adapter used. Molex tap used to probe with bench-top OWON XDM2041 DMM or Fluke 87. See prev photos)
BCK 1.565 v (audio playing/ not playing indifferent); 2.6 v (USB unplugged)
MCK 1.36 v (audio playing/ not playing indifferent); 2.4 v (USB unplugged)
WS 1.565 v (audio playing/ not playing indifferent); 2.6 v (USB unplugged)
DATA 0.4 (no audio playing) - 0.8 (audio playing); 0 v (USB unplugged)
============
No problem, I'll save you some time/effort
The same XMOS adapter works fine with other Philips TDA dacs like 1543 and 1387. They are both 5v devices. But neither requires MCK.
Actually, yes it does work ... and very well. See photo in Naim 1305 thread.
Apologies: my bad earlier in the that thread in that I referred to it as a 1543 (the OEM CDP in that post is a weird Philips NOS player which indeed used the 1543). I swapped it out with an SOIC 1387 (on a modded DIP adapter), put an electro Jamicon cap for V_ref, cut the I2S lines on the DIP adapter, and soldered in wires for a Molex. And all that connects to the XMOS (shown in that photo).
The 1387 is completely powered by the OEM Philips CDP. Of course, GND is shared. But that's it.
And that earlier success with the same XMOS adapter is what led me down the rabbit hole ... would it work for another 5v Philips TDA dac?
The 1305 and other Bitstream (noise-shaping) DACs are overlooked. Much attention goes to earlier Philips 154x DEM and CC series. But, FWIW, Bitstream (noise-shaping) is what convinced Chord's Rob Watts into pursuing his own projects; before, as he claims, he was only a vinyl guy.
Last edited:
Now Playing:
Bob Catley
Album: Spirit of Man
Track: End of the Story
Same Xmos adapter as above, with diy /2 added on . Not using Chinese kit (see photos). Single TDA1305T built per datasheet on protoboard and breadboard. Proof of concept. Have not taken any measurements except DC offset (1.0 mv both chs). But: Good sound quality on output.
Bob Catley
Album: Spirit of Man
Track: End of the Story
Same Xmos adapter as above, with diy /2 added on . Not using Chinese kit (see photos). Single TDA1305T built per datasheet on protoboard and breadboard. Proof of concept. Have not taken any measurements except DC offset (1.0 mv both chs). But: Good sound quality on output.
Which is what I did in my last post with the photos of the updated experiment.
The noise problem may have been, as I noted earlier, the strange dual-parallel arrangement of two 1305Ts' of the Chinese board. That may have "messed up" input impedance enough to make the OEM CS8414 arrangement work -- which the designers tweaked for perhaps -- but not an after-market XMOS adapter.
Given that: Those clearly-labeled and INVITING i2s taps on the oem PCB are just plain weird!
The TDA1305 can work just fine, even in the simple Philips 1305T datasheet ckt (p. 4, Fig. 1; all voltage rails at 5v., see photo of Tekpower with 5.00vdc feeding the 1305) , despite the XMOS adapter (with the diy /2 module tacked on) running on ONLY 3.3v.
It can be this simple ....