Itz not just for dessert anymore!
Wolfsin predicts one or two Wolfsons served with a slice of Raspberry Pi will quickly move from being a tasty dessert to the main course. Whether you want i2s in or i2s out or both, kiss goodbye the time consuming special purpose digital hardware and just doit!
Continue to focus on the analog side of the DAC until production ramps up and then chow down next Fall (in the Northern hemi)
Wolfsin predicts one or two Wolfsons served with a slice of Raspberry Pi will quickly move from being a tasty dessert to the main course. Whether you want i2s in or i2s out or both, kiss goodbye the time consuming special purpose digital hardware and just doit!
Continue to focus on the analog side of the DAC until production ramps up and then chow down next Fall (in the Northern hemi)
Wolfsin predicts one or two Wolfsons served with a slice of Raspberry Pi will quickly move from being a tasty dessert to the main course. Whether you want i2s in or i2s out or both, kiss goodbye the time consuming special purpose digital hardware and just doit!
Continue to focus on the analog side of the DAC until production ramps up and then chow down next Fall (in the Northern hemi)
Is the Rasberry Pi already out? The only thing is it runs Linux and I don't know if ALSA will work with the XMOS chip...
Ciao!
Do
RPi haz a bunch of GPIO pinz. No need for XMOS. For input of i2s you could wire directly to the signal but for output you _might_ want a buffer so the DAC master could clock out the nibblez. It might be simpler to use XMOS but not sure there is any need to involve USB. There are enuff pinz so you could whisper i2c into the WM874x registers to control the DAC as well as send digital music.
Wolfsin is concerned about your reports of loss of detail when using legato and ivy iii. Have you figured out what was going on? Is your Opus analog at running @5v or 3.3?
Wolfsin is concerned about your reports of loss of detail when using legato and ivy iii. Have you figured out what was going on? Is your Opus analog at running @5v or 3.3?
I feel sure that USB Audio 2 is supported by Linux but that needs USB->i2s interface. I meant that by writing a driver to twiddle the GPIO pinz one could output i2s. The problem might be timing so a nibble wide FIFO buffer whose output was clocked by the DAC could solve lots of the issues that have been chewed over in the WaveIO thread. RPi requires only one watt so 200ma @5v for it and 19ma for each WM8740x would be a tiny digital supply.
Apprently ALSA provides direct i2s in some distros if that is what you mean "i2s pins no longer wired to header" but that is a whole 'nuther thang. Keep in mind that all this is off the top of my head and probably overly simplified. By only sending stereo i2s at, say, 96k, practically all of the messy issues could be handled by upsampling in software. Ok, Ok, 88.2 and 96k so there goez the simplicity :-(
The SoC has hardware i2s output available, but its not wired to an easily accessible pin unfortunately.
My thoughts were that the option of using existing ALSA driver as you say here:
Seems simpler than this -
How about them JC-80s ay?
My thoughts were that the option of using existing ALSA driver as you say here:
Seems simpler than this -
How about them JC-80s ay?
Wolfsin's mind goez OT too easily. Got the power connectors but am still waiting for the r-cores. When they arrive I will measure and post BOM.
My use for the JC-80 will be exclusively to drive HD-600s. The output of the WM8740 with analog supply of 5v is 2v RMS. The Max Voltage swing (Vp-p) for HD-600 is 6.7v but because they are wired balanced it is half that on each side. The JC-80 default gain of fourteen is too high. Dropping analog supply to 3.3v and/or adjusting JC-80 gain are two possibilities. Not sure what the noise implications of those might be.
Pinnocchio has reported loss of detail when driving JC-80 from TPA Legato and Ivy III. I would like to take signals directly from DAC considering the variety of digital filters available.
RPi iz for i2c ONLY. Forget software i2s due to insufficient i/o.
@pinnocchio Have you tried JC-80 directly from WM874x? Are you running the IVY / Legato as a buffer w/ filterz?
@ticknpop Tested the noise of your new regs?
Delivery latency is wearing on Wolfsin. RPi was interesting but totally inadequate to supply audio data.
@pinnocchio Have you tried JC-80 directly from WM874x? Are you running the IVY / Legato as a buffer w/ filterz?
@ticknpop Tested the noise of your new regs?
Delivery latency is wearing on Wolfsin. RPi was interesting but totally inadequate to supply audio data.
Yeah, running the IVY III as buffer in front of JC-80. Have not tried directly but my feeling is that It won't be optimal. I could however give it a try
Do
2v into 10k is the WM874x with a 5v analog supply. Are you using 7.5v before the reg so the DAC is getting 5v analog? Then IVY3 has unity gain and is just a filter.
Not sure what the DAC->JC-80 might do except get rid of two opamps on each channel.
It would really be great to leave IVY3 and its requisite power supplies out if possible. Alternatively, it would also be nice to hook BOTH to the DACs but . . .
2v into 10k is the WM874x with a 5v analog supply. Are you using 7.5v before the reg so the DAC is getting 5v analog? Then IVY3 has unity gain and is just a filter.
Not sure what the DAC->JC-80 might do except get rid of two opamps on each channel.
I'm feeding 7.5V before the onboard regs.
Let me see what I can do... A while back I've put in some really cheap 4-40 button screws and my tiny allen key has stripped the head of three of them... It is not possible to separate the two unless I break the pcb or find a way to remove the screws... I might have to be creative here...
Do
sounds screwed, dOOd!
Wolfsin braved the backwoods of Colorado in Winter to match those polystyrene caps. It is less than four miles but I thought I was in the depths of northern Canada. Wow was it cold where the sun doesn't shine
The original BOM called for IRF9510. What should one match on?
sounds screwed, dOOd!
Wolfsin braved the backwoods of Colorado in Winter to match those polystyrene caps. It is less than four miles but I thought I was in the depths of northern Canada. Wow was it cold where the sun doesn't shine
The original BOM called for IRF9510. What should one match on?
I had bought the original 2SK214 and 2SJ77 matched (used pulls) but one of the legs broke on the 77... I have not tried soldering back because it broke really close... Those are supposed to be the best in terms of linearity. Stanton told me not to use the 9510 but aim for the 9610 (more power).
If you want to give the 2SJ77 / 2SK214 a try, let me know. I'm making you a 40$ gift!
Ciao!
Do
Of course Wolfsin wants to try those MOSFETs! Thanks.
I have been thinking of how to make the output transistors 'changeable'. Using screws on the heat sinks goes part of the way but three pin devices are a mess to extract. It occurred to me to always bend one of the legs outward and then down and back so it could be clipped prior to heating the other two simultaneously to extract the device. Ideas guyz?
I have been thinking of how to make the output transistors 'changeable'. Using screws on the heat sinks goes part of the way but three pin devices are a mess to extract. It occurred to me to always bend one of the legs outward and then down and back so it could be clipped prior to heating the other two simultaneously to extract the device. Ideas guyz?
Temptation
Wolfsin iz 'drawn like a moth to a flame' toward the torch
And to think the first tone has yet to issue forth!
The irf9?10-irf?10 parts are plentiful and inexpensive (unless you want them to have been made by Fairchild). Nelson Pass has written a nice article on MOSFET matcing. Everyman should find himself at home here. Then there is the esoterica, pricey so matching pools are limited, and downright unobtaniums. Always the lure of NOS with one eye cocked, ever watchful for counterfeits. Everyman may feel uneasy here.
Postponing the decision (something at which Wolfsin excels), building easy maintenance (read upgradeability) into the component, and staying "awake at night agonizing over choices", are some strategies used by DIYerz. If only this ended once the music began to play but alas . . .
Wolfsin iz 'drawn like a moth to a flame' toward the torch
And to think the first tone has yet to issue forth!The irf9?10-irf?10 parts are plentiful and inexpensive (unless you want them to have been made by Fairchild). Nelson Pass has written a nice article on MOSFET matcing. Everyman should find himself at home here. Then there is the esoterica, pricey so matching pools are limited, and downright unobtaniums. Always the lure of NOS with one eye cocked, ever watchful for counterfeits. Everyman may feel uneasy here.
Postponing the decision (something at which Wolfsin excels), building easy maintenance (read upgradeability) into the component, and staying "awake at night agonizing over choices", are some strategies used by DIYerz. If only this ended once the music began to play but alas . . .
Request for critique !!!
Thanks to the DIYaudio search I was able to find the response below re sockets.
Do pm'd me a picture of just the socket Wolfsin had in mind. Now here's what I am thinking. 'the wire' is exceptionally quiet. If I equalize the gain of one channel of JC-80 to one channel of 'the wire' and run the +/+ into one opamp and the -/- into another (or some other arrangement) I should be able to sample the output. Smaller would be better. Various output transistors could be swapped in and out, brought to temp, and 'matched' in-situ.
Thanks to the DIYaudio search I was able to find the response below re sockets.
Do pm'd me a picture of just the socket Wolfsin had in mind. Now here's what I am thinking. 'the wire' is exceptionally quiet. If I equalize the gain of one channel of JC-80 to one channel of 'the wire' and run the +/+ into one opamp and the -/- into another (or some other arrangement) I should be able to sample the output. Smaller would be better. Various output transistors could be swapped in and out, brought to temp, and 'matched' in-situ.
I have those sockets in tin and maybe gold left. I will send to you. How many three pins sockets do you require?
Do
Wolfsin iz 'drawn like a moth to a flame' toward the torch
The irf9?10-irf?10 parts are plentiful and inexpensive (unless you want them to have been made by Fairchild). Nelson Pass has written a nice article on MOSFET matcing. Everyman should find himself at home here. Then there is the esoterica, pricey so matching pools are limited, and downright unobtaniums. Always the lure of NOS with one eye cocked, ever watchful for counterfeits. Everyman may feel uneasy here.
Postponing the decision (something at which Wolfsin excels), building easy maintenance (read upgradeability) into the component, and staying "awake at night agonizing over choices", are some strategies used by DIYerz. If only this ended once the music began to play but alas . . .
Yeah... I completely dig you on this!
The more you listen to music and the more you wanna change things... I believe it is a DIY disease...!
Ciao!
Do
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.