Ian,
i have connected as per the instructions above to the PCM board, but I cannot go higher than 88 (or 96) kHz sampling; above this the signal breaks up in two halves.
I tried with J08; and with J08 plus J06: neither works above the sampling frequency.
Source is WaveIO; I feed the isolator through a K-multiplier board and this is tapped from the PCM board. so it gets a clean 3,8V. No FIFO. Player: iTunes from Mac.
The TDA1541 +power supply worked when I had it connected to a SAA7220A
- Do you or anyone else have suggestions?
Hi Albert,
PCM board was designed and tested up to 384KHz, so should be no any problem @88.2/96KHz.
Since you don't have a FIFO in your system, I'm a bit worry about your MCLK, where it came from? whats the frequency?
Are your using TDA1541A DAC? Did you configure it at simultaneous mode?
BTW, you will need a oscilloscope with higher bandwidth, at least 100Mhz. Otherwise it will be limited from seeing real waveform.
regards,
Ian
I first used the german TDA1541 board, similar problem. Then changed from shunt to discrete and to my own board (that was made for NOS and proven to work with SAA7220). Same problem.
I have TDA1541 non-A inserted right now (is a pre-A one with separate connection, a pull from a player, with a capacitor); and it works at 88 kHz.ian said:
My 88kHz and 96 kHz sampling frequency in simultaneous mode:
check with real TDA1541A : same problem
I have it from the bottom of the WaveIO board. Soldered thin shielded cable.Ian said:
Goes to receptacle at top of board, J4, called MCLK. Thought it was a no-brainer to connect it there. :roll eyes:
Checked frequencies on input of DAC board (pin 2, CLK) and the output of WaveIO (pin going to MCLK):
Sampling freq....CLK output....... MCLK input
88 kHz .............1.588 MHz........22.578 MHz
96 kHz .............1.728 MHz .......24.578 MHz
176 kHz ...........3.176 MHz .... ..22.580 MHz
352 kHz ...........zero ....... ......22.578 MHz
I have jumper I05-IO6 (HALF) open initially; closed it, and this give no difference on lack of output, the zero CLK and is like the table above.
hope this helps in pinpointing the issue.
At highest speed there is no output from dac. (no CLK either)
Last edited:
Hi Albert,
Did you get the AYA II board as I advised you in our last conversation once year ago (more or less) ? It is proven to work at 192 K hz with the first Wave I/O which was limited to 192 K Hz (the Wave I/O). This pcb should work at 380 K Hz (Ask for instance to Supra member which had the wave I/O, Ian's front end line, AYA II 2014, and Crystek 44/45 XOs ?). Since Lorien with Ian has solved the issues if I remember on the Wave I/O... at least you may use the non isolated side with uf-l plugs for a perfect connection with Fifo/isolator/clock/pcm combos . Or even with the Wave I/O and the PCM board only : uf-l connection only at these speeds).
Ian said me even the PCM is ok with 380 upsampled material on his TDA1541 board...).
I had also myself some problem because the connection of the clocks adaptators on the tulipe DIP12 of the clock boards had a non enough solid connection (if you force a little the pins to improve the coupling, it's ok with and without the decoupling at the bottom of the adaptators clock boards !)
I have also a tda1541 non A but it is marqued Y and I have a doubt about the signification of it (I even read some non A worked only with an external DEM timing cap
).
You may have a layout problem somewhere, not sure it is about your own pcb...
Ask also to Lorien if your firmaware version is 352/380 K Hz enable (you may need to upgrade the eprom with a Windows pcb)
With Simultaneous board and your 22 M Hz Crystal you need iirc to strap your PCM in half speed mode as well.
Did you always heard your MAC usb outputt plays 352 upsampling material ?
Did you get the AYA II board as I advised you in our last conversation once year ago (more or less) ? It is proven to work at 192 K hz with the first Wave I/O which was limited to 192 K Hz (the Wave I/O). This pcb should work at 380 K Hz (Ask for instance to Supra member which had the wave I/O, Ian's front end line, AYA II 2014, and Crystek 44/45 XOs ?). Since Lorien with Ian has solved the issues if I remember on the Wave I/O... at least you may use the non isolated side with uf-l plugs for a perfect connection with Fifo/isolator/clock/pcm combos . Or even with the Wave I/O and the PCM board only : uf-l connection only at these speeds).
Ian said me even the PCM is ok with 380 upsampled material on his TDA1541 board...).
I had also myself some problem because the connection of the clocks adaptators on the tulipe DIP12 of the clock boards had a non enough solid connection (if you force a little the pins to improve the coupling, it's ok with and without the decoupling at the bottom of the adaptators clock boards !)
I have also a tda1541 non A but it is marqued Y and I have a doubt about the signification of it (I even read some non A worked only with an external DEM timing cap
).You may have a layout problem somewhere, not sure it is about your own pcb...
Ask also to Lorien if your firmaware version is 352/380 K Hz enable (you may need to upgrade the eprom with a Windows pcb)
With Simultaneous board and your 22 M Hz Crystal you need iirc to strap your PCM in half speed mode as well.
Did you always heard your MAC usb outputt plays 352 upsampling material ?
Last edited:
No, I missed the AYA board rerun. Have not done a lot for a long time
Well, thats DIY.because the clock signal disappears, this can well be the issue. I will contact Lorien.
Correct, I collected Philips players (Piekfijn), and these often contained the non-A with a capacitor.I have also a tda1541 non A but it is marqued Y and I have a doubt about the signification of it (I even read some non A worked only with an external DEM timing cap
I tried to solder the plugs but they are a real disaster for a DIY.
Hi Albert,
Can you post a picture of your PCM board /1541DAC hookup to see if I have any suggestion?
Regards,
Ian
Ian, I shame myself for what I am going to post, but I have some consolation knowing many out there work like this :roll eyes:
One Ufl has a plug (i was only able to get one soldered, all others failed on me) while two have a uff cable only and one (the latch enable) has no shield.
Shields are connected both sides.
I thought the shielded cables were better than the one with the plug which is not shielded (black if I remember) , but I have been thinking of reverting to the plug you supplied with the attached cable.
Well, it gets to a mess once one starts fooling around
Back in December 2013, I connected an I2S-PCM board between a WaveIO and my diy tda1541a dac and also could not get it to work properly with music with a sampling rate higher than 96kHz. I puzzled over it for a while and wrote about it on this forum (post #603 to #622) but there was no resolution. Then I installed the FIFO, isolator, and clock boards and 192kHz music played successfully. I did not do anything other than install the FIFO, isolator, and clock boards.
Ian thought the problem may have been caused by a ground loop and EMC noise, and that was cut by the addition of the isolator.
I eventually upgraded the crystal oscillators to the higher frequency Crysteks and even though with the higher frequency clocks I should have been able to play 352.8kHz and 384kHz material, all I could get was music masked by a lot of noise. Since the ability to play material of these frequencies was not critical to me, I didn't pursue the matter further.
Perhaps my issues with I2S-PCM board and tda1541a dac and higher than 96kHz material may shed some light on the matter.
My layout and wiring:
Ian thought the problem may have been caused by a ground loop and EMC noise, and that was cut by the addition of the isolator.
I eventually upgraded the crystal oscillators to the higher frequency Crysteks and even though with the higher frequency clocks I should have been able to play 352.8kHz and 384kHz material, all I could get was music masked by a lot of noise. Since the ability to play material of these frequencies was not critical to me, I didn't pursue the matter further.
Perhaps my issues with I2S-PCM board and tda1541a dac and higher than 96kHz material may shed some light on the matter.
My layout and wiring:
Mac based with WaveIO... I think Eldam is onto something.. if you have an early version of WaveIO and haven't updated the firmware, there will be timing errors which will become apparent at higher sampling frequencies. Fortunately addressed by Ian and implemented by Lucian in the PC (platform only unfortunately) downloadable update.
But still 192k was no problem with the earliest WaveIO to 1541A via Ians PCM board so perhaps something else (if you're limited to 96k).
Update firmware and be sure that the jumper selection is correct as per Ian PCM.pdf for use with the clocks on the WaveIO, Ian is correct about their frequencies.
More info about the I2S timing errors and impact on higher sampling rates to be found around ie; http://www.diyaudio.com/forums/digi...project-ultimate-weapon-fight-jitter-335.html
LH/S
Last edited:
Ludwig,
I also thought about the USB cable. The standard cables are not shielded, and even not twisted like CAT6. So I made a cable, a short one, based on Cat6. But no improvement.
What i wil test today is some beads.
I have a separate chassis for the transformers. So the 1 meter cable gives all the chance for coupling of the earth wires . . .
192 kHz is OK for me, I have a filter for that. The 88 kHz simultaneous mode is so incredibly clear (clear steps without any latency between the bit levels), I will post an example.
I have a PM for Lucian. I use a 2015 board if I am correct.
I read now that the WaveIO is enabled to 192 kHz, thus I have to accept that limit.. But for now I am stuck with 88 kHz.
Condo: Thanks for remarking this. I will revisit this Kondo filter again and show my results; I have it in my DAC-player.
First prio is getting this USB stuff into 'production' (the living room).
albert
Anyway I will not be worry about USB as far your Wave I/O is powered by a clean 5V and you use the isolated side !
Surely the layout of the dac chip pcb.
If I understood your photograph, the yellow wire is TDA I2S pin 1. Is it shielded ?
In anaother way, I2S wires are very shorts which could not be bad even if the yellow wire LRCK is not a 50 ohms (?) at this length and not so high speeds (?). I say that because the yellow one looks like a 100 ohms network cable (?)....
maybe a Gnd pb of the TDA1541 pcb ? Mainly the MCLK connections between Wave I/O and PCM board and quality of BcK connection (and gnd layout) between PCM and dac chip board.
If I remember the 2015 Wave I/O is ok for 354/380 speeds. Just need maybe the half-speed strap on the PCM board as crystals are 21/24xxx on the Wave I/O !
Surely the layout of the dac chip pcb.
If I understood your photograph, the yellow wire is TDA I2S pin 1. Is it shielded ?
In anaother way, I2S wires are very shorts which could not be bad even if the yellow wire LRCK is not a 50 ohms (?) at this length and not so high speeds (?). I say that because the yellow one looks like a 100 ohms network cable (?)....
maybe a Gnd pb of the TDA1541 pcb ? Mainly the MCLK connections between Wave I/O and PCM board and quality of BcK connection (and gnd layout) between PCM and dac chip board.
If I remember the 2015 Wave I/O is ok for 354/380 speeds. Just need maybe the half-speed strap on the PCM board as crystals are 21/24xxx on the Wave I/O !
Thanks, Eldam,
the yellow wire is single strand wire, for LE, not isolated at all, and I thought for this low speed it would be OK. Thinking about it, the rising edge is a timing edge and is fast of course. . will make a shielded one.
The ground is a plane and give the cleanest output I have seen
I willstop pursuing 352 etc.
I suspect a problem here: everyone uses the isolated outs from WaveIO and then connects a non-isolated MCLK
How did you do it?
albert
I was wrong in this wild idea: the Mclock on WaveIO after isolator goes to Ians's card.
+ I now used an isolated cable for the LE.
+ I inserted inductors in all entering power lines (not just the earths).
+ and I added a parallel choke on the WaveIO power feed.
And my Macbook is always running from batteries.
Nothing helped.