I understand but have you tried REAL current measurements while playing hires files with NDK oscillators?
I assume Salas works better with more power headroom but if REAL current is about 350mA wouldnt it be better to just set it at 400-450mA than 600mA? Even for heat dissipation purposes? Have you tried how it sounds with different settings?
I assume Salas works better with more power headroom but if REAL current is about 350mA wouldnt it be better to just set it at 400-450mA than 600mA? Even for heat dissipation purposes? Have you tried how it sounds with different settings?
Hi Zibra
Yes Merlin that’s my take also.
I have my Salas BJT set up for a CCS of 670MA which seems to work well. From what I’ve read the Salas likes plenty on the CCS so just use a nice big heat sink.
I have my board and heat sink as follows
Franz Gysi style Salas
http://www.diyaudio.com/forums/power-supplies/192625-sslv1-1-builds-fairytales-10.html
Runs just warm and sounds sublime.
I used a linear reg previously and swapping in the Salas made a noticeable difference to the SQ.
I’ve had my WAVEIO for just over a year now and I’m still loving it.
I’m using my board with TP LVDS sender and receiver and that works very well indeed. Looking forward to WAVEIO’s Daughter.
Great work Lucien
For 5Vout I recommend you use Mosfet instead BJT in Salas BiB reg.
I know that measuring current by multimeter could have some "errors" but whats general idea of setting Salas shunt regulator for powering WaveIO.
If it takes real current for lets say 350mA with NDK oscillators on hires files and Salas works better with more power headroom, why dont you set it at 1A with very good heatsink? Does that mean it will work even better??
Im just curious whats the purpose of having such big power headroom in that case.
Sorry for beginner questions but I would really like to set it at its best.
Jakub
If it takes real current for lets say 350mA with NDK oscillators on hires files and Salas works better with more power headroom, why dont you set it at 1A with very good heatsink? Does that mean it will work even better??
Im just curious whats the purpose of having such big power headroom in that case.
Sorry for beginner questions but I would really like to set it at its best.
Jakub
That was the point: to work
Lucian,
I mainly user the player that comes with the Ipad, but also tested other apps like "Golden Ear" (which plays almost any file format) and Spotify. I guess that you can virtually listen to any app that outputs sound ....
Nevertheless don´t forget that if you want to connect an USB cable to the Ipad you´ll need a "camera adapter" like this one:
Apple iPad Camera Connection Kit - Apple Store (U.S.)
Cheers
Pepe
Hi Jakub,
I am using WaveIO(NDK) with Placid HD shunt. Playing 24/192 res file the device consumes 320 mA as measured with Fluke 287. I've got the Placid adjusted to have 100 mA shunt headroom meaning the total current through the Placid is 420 mA. I hope this helps.
Cornelius
Hi Merlin
Yes I did consider using the mosfet Salas but I only had BJT to hand. I’ll use the BJT for 3.3V on another project later.
A (slight) compromise in performance
Err yup.
But
Still loads better than the linear that I was using before.
I’m going to build up a 5V mosfet Salas when time allows.
Just as an aside
My board has the FOX oscilators
Hi Lorien, I have a board on order with your good self, yet to be received.
To this end I am a little unsure if the daughter board would be required, my new board may include the modification?
Sorry for the dumb question but I have not quite grasped how this works, maybe you need more money for the board I have on order to include the modifications.
Regards
Tony
To this end I am a little unsure if the daughter board would be required, my new board may include the modification?
Sorry for the dumb question but I have not quite grasped how this works, maybe you need more money for the board I have on order to include the modifications.
Regards
Tony
my adum usb isolator only works in windows recognized as a class 1.0 usb device. even then i have to fiddle with waveio every time i turn on the computer to get it working properly. what i do is put the waveio back in usb power mode, switch it to ext. power, then plug in the isolator inbetween. in linux it's not even recognized. any input anyone?
Well the point is to provide isolation between (USB + XMOS + Surrounding logic) vs (oscillators + Flip-Flops). In addition I hope the jitter will be lowered this way. I do not know if the SQ will be improved but I really do hope so since this is the main goal
As for PSUs, there will be some in addition those already used but I'm still in designing process and I do not now for sure how many. For the moment two are involved which have to be completely different from what WaveIO uses (no ground connection).https://www.sellarz.co.kr:127/shopping/shopping-detail.php?pi=1856
im trying to use this tda1543 board with the direct out of waveio with no success. ive connected data, lrck, bclk accordingly and left mclk untouched. the grounds of the tda1543 goes to usb ground on waveio. and... there is no sound!
I saw that you solved your issue so I hope it will be a good SQ. As for isolators, there were others that have tried ADUM USB isolators with WaveIO but without success probably because is capable of USB 1.x speed and WaveIO enumerates itself as a High speed device.
Noted, thank you!
There's no daughter card integration as for now I only have done the DFU feature (which is released "by default" with all new cards shipped after 10 Sept), free of any charges.
Kind regards,
L
Forgot about a question to all of you guys: I would need help in searching for a really good power supply for the oscillators which can be squeezed in a relative small PCB space (3.3V @ 100-200mA max)? I opted for Salas but he does not like the idea to have his design implemented in a commercial product so I respect his decision and move on. There's LP5900 approach (already used in WaveIO) but I feel that is something better out there...
Thank you,
L
Thank you,
L
if its just for clocks, which are pretty low current and honestly i'm not even sure why you would bother with a regulator that is focused on high dynamic performance and low impedance, like the salas shunts. great regs, but for clocks the requirements are pretty different, they just want the lowest noise, they are about as benign a load as can be, just need constant current, constant voltage, hardly any dynamic regulation is needed.. for something to fit in a small space, I would do something akin to the flea, very similar to what acko uses for his clock regs, just a low noise reference like LTC6655, MAX6126 in series with, or wrapped in the feedback loop of an amplifier
basically something like this
for compact high performance clock and analogue reference circuits, lately i'm using the wire SE-SE ultra low noise headphone amp with the front end of the above but with the LTC6655 in place of the MAX to make for an adjustable, wide bandwidth very low noise dual regulator capable of up to 12-13vdc @250ma. the problem with the linear chip is it kinda wants a large film cap bypass for best noise performance, you've seen the result of that on the latest gen of ackos regs with the huge blue caps. if you dont need that much current, like if you only need 20-25ma you dont need the buffer, just the opamp or an opamp and a small mosfet like above. works out pretty compact and the PCB is the only heatsink needed for up to 250ma at low voltages, but my version with the buffers and 'the wire' isnt exactly low dropout (but then neither is the salas, its better than that)
for clocks i've also played with using the same circuit but instead of the reference, just connect a LiFePO4 battery (conveniently 3.3-3.45vdc), ive not had to charge it yet, it draws µW from the battery due to the massive input impedance of the opamp, so should last almost indefinitely
basically something like this
for compact high performance clock and analogue reference circuits, lately i'm using the wire SE-SE ultra low noise headphone amp with the front end of the above but with the LTC6655 in place of the MAX to make for an adjustable, wide bandwidth very low noise dual regulator capable of up to 12-13vdc @250ma. the problem with the linear chip is it kinda wants a large film cap bypass for best noise performance, you've seen the result of that on the latest gen of ackos regs with the huge blue caps. if you dont need that much current, like if you only need 20-25ma you dont need the buffer, just the opamp or an opamp and a small mosfet like above. works out pretty compact and the PCB is the only heatsink needed for up to 250ma at low voltages, but my version with the buffers and 'the wire' isnt exactly low dropout (but then neither is the salas, its better than that)
for clocks i've also played with using the same circuit but instead of the reference, just connect a LiFePO4 battery (conveniently 3.3-3.45vdc), ive not had to charge it yet, it draws µW from the battery due to the massive input impedance of the opamp, so should last almost indefinitely
Last edited:
