I received my JLSounds board today. I installed it and it works fine, just on 5V via USB at the moment. One thing I have noticed though is that if the board is plugged in when I boot the computer it isn't recognised but if you unplug it and then replug it is fine; this won't be a problem to me when I power it with decent regulated supplies.
Ray
Ray, glad to hear that. What player are you using, are you using jl's drivers?
The PC is running Audiophile-Linux, no driver installation was necessary. I used Cantata, which is bundled with the Linux distribution, as the player, though Cantata is a front-end to mpd so it is probably more correct to say that mpd is the player.
Ray
WTF!!!! I'm a windows/JRiver guy , Linux is totally foreign to me, I guess I need to get up to speed?
Sorry, I thought you were already running AP-Linux when you responded like this to my earlier post mentioning it;
Anyway, well worth a try. I was an advocate of JRiver but have moved away from it because of their, seemingly spiteful, blocking of JPlay.
Next steps for me are to install proper power supplies for the JLSounds board and then try a simple LP filter on it's output.
Audiophile Linux/ MPD/ Cantata has provided the best sound from my DSD files, thru my USB 'no DAC'. Unfortunately, the developer confirmed that there is no way to convert PCM to DSD on the fly in Audiophile Linux and my flip-flop circuit cannot handle PCM, so I have now installed Windows 10. I am using HQPlayer (eval version) and this combination sounds good. HQ Player plays native DSD and converts PCM to DSD on the fly. I can't do an A-B comparison with AP Linux so I'm just relying on my unreliable audio memory. But i would say that Linux is still best sound. Next - I am gonna try out Ubuntu Studio - and install HQ Player for Linux. Hopefully this will sound as good as AP Linux, with the convenience of PCM-DSD conversion. Unfortunately I have spent the last 2 weeks learning all about new OS (Linux and Win10) and new media players.WHich is why I've been a bit quiet on this thread recently.
Last edited:
Hi Ray, how have you configure the output for your JL Sounds USB board? I've seen them on internet and I might get one to compare to my DIYINHK board. I think that are about 67euro, seems pretty good value.
Hmmm... yes. You'd get DC for sure.
Is that a problem? I suppose one could make a file like that, but would it occur in music? It has not been a problem for me using transformers on PCM Delta/Sigma DACs, but this is another animal.
Is the DSD stream likely to have much DC?
OK I have measured the DC on the output pins on the DIYINHK board. And not surprisingly it is 1.65v (half of the supply voltage, 3.3volts). But I think we need to stop think voltage because the output is a stream of digits - lo (0 volts) and hi (3.3volts).
My knowledge of digital is very limited, but I am guessing that a constant input of zero volts would give a DSD stream of 101010.... Max positive output would be 1111111.... Max negative amplitude would be 0000000. So for zero volts, output is switching between 0 volts to 3.3 volts at a frequency of 2.8MHz. And if we average this, it give a DC of 1.65V. Which makes sense, as the DC at zero input should be half way between the maximum and minimum. Does this make sense?
In my original implementation I took DSD signal straight off the USB board to my transformers. But this means that I was effectively putting DC into my transformer and my hypothesis is that it was saturating - hence the poor base response I reported in my OP.
Now I am using the flip flop to provide complementary outputs - so for zero signal, the data stream is 101010... and the complementary data stream is 010101..... But they both average out the same (half of the supply voltage). Indeed, I am measuring 2.94 volts DC on both outputs of the flip flop (As an aside, I have replaced my 5volts SMPS with a 6V SLA, which sounds much nicer, and has more output swing.)
So I would not recommend a 'single end' filter taking the signal straight off the USB board - maybe you would need a blocking capacitor?
Hopefully someone who knows about digital can confirm this.
I suppose it depends on the chip. I have not tried bitstream into a transformer, but have done many DAC chips into them. On differential DACs, both the plus and minus sides of the signal are at the same DC potential above ground. Since both pins are at the same potential, no current flows thru the transformer winding. No worries. The only current that does flow is when there is a difference between the pins, and that difference is AC. If the chip supplying the DSD stream has differential outputs, then both pins should be at the same DC level. No DC current will flow.
As to whether they is DC offset in the DSD stream, I don't know. Seems unlikely, but I don't enough enough about the signal to be sure.
As to whether they is DC offset in the DSD stream, I don't know. Seems unlikely, but I don't enough enough about the signal to be sure.
If the initial toe dip proves to be a positive experience I plan on taking a similar route by installing HQplayer, however, I'm pretty sure my current processor won't be up to converting FLAC to DSD256 on the fly so I'm sticking with simple playback of DSD files for now.
Ray
I haven't actually done anything with the output from the USB board yet; need to decide the approach to a LP filter, influenced by not having any line transformers to hand. I know of at least a couple of DIY project DSD filter boards under development too.
Regarding the JLSounds board, it seems like a quality item to me, very nicely but together and at a good price. It was delivered quickly and was very well packaged too.
Ray
This thread is based on the use of USBtoI2S board to provide DSD bitstream. As far as I can tell, these boards will always be single ended. WHen you talk about DACs above, remember you are taking the analog signal after it has been through the DAC. We are taking the bitstream before it hits the DAC, and its single ended. So is PCM as a matter of fact. Actually, as both DSD and PCM are digital signals I'm not sure if its right to talk about single ended and balanced signals. Anyways, this is the reason I have implemented a dual output flip flop after the USBtoI2S board. THe flip flip outputs the input (DSD) and also it's complement. So when bitstream is 0, one output is 0 and the other is 1 (which equals 6volts in my system, but can be whatever the supply voltage is). And vice versa. So is it balanced? Since its in the digital domain maybe we just say that outputs are complementary. Thats what the manufacturer's data sheet says.
Sorry, I thought for sure I saw a schematic in this thread with complimentary outputs. Was it something already passed thru a flip-flop or a buffer?
And don't fret, I know what a DAC is and the difference in taking the bitstream straight out.
Just thought I saw a complimentary output, similar to many DAC chips. Must be hallucinating.
And don't fret, I know what a DAC is and the difference in taking the bitstream straight out.
Just thought I saw a complimentary output, similar to many DAC chips. Must be hallucinating.Hazard500, et al,
First let me say I applaud your desire for and approach with simplicity!
I'm a mac guy....
I have a diyinhk XMOS USB rx board here. Yesterday, I down loaded HQPlayer but couldn't get it to work with your DSD approach. I also downloaded JRiver and was able to get things to work. There was some noise. I used the rx board directly into some nickel bifilar txs.
The sound of HQPlayer with PCM is big a forward, almost bombastic. The sound of DSD with JRiver is OK. Not as good as Audirvana (PCM) with a DAC chip. But, there's some merit. River locked for some reason, dunno why, and I ended there. I suspect that Windows with this board is a better route. I'm not going there though. I wish diyinhk would do direct DSD with a rx board instead of DOP.
Fun trying things though....
First let me say I applaud your desire for and approach with simplicity!
I'm a mac guy....
I have a diyinhk XMOS USB rx board here. Yesterday, I down loaded HQPlayer but couldn't get it to work with your DSD approach. I also downloaded JRiver and was able to get things to work. There was some noise. I used the rx board directly into some nickel bifilar txs.
The sound of HQPlayer with PCM is big a forward, almost bombastic. The sound of DSD with JRiver is OK. Not as good as Audirvana (PCM) with a DAC chip. But, there's some merit. River locked for some reason, dunno why, and I ended there. I suspect that Windows with this board is a better route. I'm not going there though. I wish diyinhk would do direct DSD with a rx board instead of DOP.
Fun trying things though....
Probably. In any case, is there DC at the outputs? Anything that would cause DC to flow in the primary of a transformer?
A little note for those of you using JRiver to stream to DSD. The DSP (volume, EQ, etc) works on files that start out as PCM. But DSD format files go out unaltered. So no DSP, volume at full, etc. Other players may do the same. Just so that you are aware.
Now I see the point of your more recent posts here, which I was reading out of context of the earlier posts.
There is DC at the outputs (of the circuit with the flip-flop to generate balanced/complementary outputs) with respect to ground, but NOT between one output and its complement. You should be able to connect a transformer across those outputs without DC going through it (other than the case demonstrated in an earlier post where all bits are 0's or all 1's, which I presume never happens with a DSD recording). You might need a buffer, depending on the transformer's impedance and the impedance of what's connected to the transformer.
Yeah, that's a 'problem' with DSD - there's no known way of digitally changing volume or EQ directly with DSD. It must be first internally converted to PCM where these things are done and then converted back to DSD, Apparently something is lost in this process for those who can tell the difference between (unaltered) DSD and PCM.