No dual supply is needed, just a single rail. Its only 1VRMS output after all.
I think a JG buffer might well work fine - you don't need the inductor though (from memory it has an LC filter before the JFET) as the filtering already built in is plenty fierce enough. But if you leave it in there won't be a problem.
As for the magic you should be able to find out fairly soon, once I've finished building the last couple of boards I'll then give them a final test and get them out. You'll need a single low noise supply of 8 to 9V to power it all. In total with the buffer configured to drive headphones it'll take about 160mA, for line driving we can reduce the current requirement a little bit.
I think a JG buffer might well work fine - you don't need the inductor though (from memory it has an LC filter before the JFET) as the filtering already built in is plenty fierce enough. But if you leave it in there won't be a problem.
As for the magic you should be able to find out fairly soon, once I've finished building the last couple of boards I'll then give them a final test and get them out. You'll need a single low noise supply of 8 to 9V to power it all. In total with the buffer configured to drive headphones it'll take about 160mA, for line driving we can reduce the current requirement a little bit.
Yes its I2S input though plans are in the pipe to develop a digital interface board with the same form factor. PCB dimensions are 5 * 5cm - because that's a really cost-effective option from PCB vendors here, we can panelize 4 boards into 10 * 10cm which is the largest size board which goes for a cheap price.
The height of the stack depends on which board options are chosen but typically PCBs are spaced no more than 2cm apart, the top board which sports the amp is taller (up to 3cm potentially) if higher valued output caps are chosen (for lower impedance cans).
The height of the stack depends on which board options are chosen but typically PCBs are spaced no more than 2cm apart, the top board which sports the amp is taller (up to 3cm potentially) if higher valued output caps are chosen (for lower impedance cans).
What boost converter did you use? Anything special?
That would make it pretty simple, just need to change the 18650 cell with 3.3v and connect up the boost.
Thanks
Randy
So I ended up just using one 5V reg and one pair of batteries to supply power to the allo reclocker and the dac board.
Took a headphone amp I was no longer using, and used the board for mounting the dac.
Letting it run in some, but initial impressions were good.
SW is PiCorePlayer audio version. I have LMS running on another pi.
Randy
Took a headphone amp I was no longer using, and used the board for mounting the dac.
Letting it run in some, but initial impressions were good.
SW is PiCorePlayer audio version. I have LMS running on another pi.
Randy
Attachments
a guy from a battery manufacturer told me that for series charging they need to be matched to be anyway predictable
individual cell float charging gives great piece of mind
If I had more than 2 in series, I'd worry about it more. I'll monitor them, and see how balanced they are over time. If they do go unbalanced, I can add 2 isolated PS's at that time. For now, this was a real quick and dirty way to get sound out of this card.
Listened some last night. Not sure if its broken in yet, and probably want to upgrade some caps, but it sounds good, much better than it should for the price. Didn't compare to my soekris yet, will give this some time to run in before I do that.
Randy
Can anybody explain this noise? I made a video.
YouTube
Let me explain that I am GREATLY exaggerating this DAC's noise level by pumping it through a microphone preamp into an old guitar amp line input. That was the only way to get it audible enough for the video. In normal use through speakers or headphones, the noise is pretty soft.
In the video I do four things. First I play with the proximity effect of my hand near the output caps (a la theremin
). Then I touch some metal parts with my hand. Next, I try a wire direct to ground. Finally I use the winner--the ground pin of my Mac's power brick. What???
I am quite baffled as to why this works so darn well. I have found that it only works when the other end of the power brick is connected to my Mac AND my Mac is connected to my monitor, AND the power cord of the monitor earthed via its own ground pin on the cord. If I disconnect the monitor, the Mac PSU alone only attenuates the noise a little bit. I can also work this in reverse--I can earth the Mac power brick and touch the monitor cord's ground pin to the DAC, and get the same perfect silence. In both scenarios, it does not matter if the Mac or monitor are switched on or off via the power strip...the result is the same. The way I originally discovered this effect was by accident...the same connections were being made indirectly when I hooked the new RPi DAC to a source selector box along with a Firewire DAC connected to the Mac.
BTW Randy, I am guessing you hold the Soekris in high regard? It will be interesting to read your comparison as well as any cap rolling you try. Is your new DAC more or less silent with the battery power?
YouTube
Let me explain that I am GREATLY exaggerating this DAC's noise level by pumping it through a microphone preamp into an old guitar amp line input. That was the only way to get it audible enough for the video. In normal use through speakers or headphones, the noise is pretty soft.
In the video I do four things. First I play with the proximity effect of my hand near the output caps (a la theremin
). Then I touch some metal parts with my hand. Next, I try a wire direct to ground. Finally I use the winner--the ground pin of my Mac's power brick. What???I am quite baffled as to why this works so darn well. I have found that it only works when the other end of the power brick is connected to my Mac AND my Mac is connected to my monitor, AND the power cord of the monitor earthed via its own ground pin on the cord. If I disconnect the monitor, the Mac PSU alone only attenuates the noise a little bit. I can also work this in reverse--I can earth the Mac power brick and touch the monitor cord's ground pin to the DAC, and get the same perfect silence. In both scenarios, it does not matter if the Mac or monitor are switched on or off via the power strip...the result is the same. The way I originally discovered this effect was by accident...the same connections were being made indirectly when I hooked the new RPi DAC to a source selector box along with a Firewire DAC connected to the Mac.
BTW Randy, I am guessing you hold the Soekris in high regard? It will be interesting to read your comparison as well as any cap rolling you try. Is your new DAC more or less silent with the battery power?
Last edited:
Oops. Believe it or not this is my first Youtube upload, ever. Thanks for the heads up. I had forgotten to hit the publish button. Please try the link again.
YouTube
Video works now.
I still think its some weird ground/return problem, but no idea what is causing it.
I've made many mods to my soekris dac
My soekris mod blog
and I think it sounds pretty good now. There are actually more changes I made recently, new usb to i2s converter, added potato flip flop on i2s lines to reclock them, and removed the isolators.
But I have a lot more time and money in that setup, so will see how this little guy compares.
I only tried this new dac with battery power. When I have some time, after it settles in, I can disconnect the battery and see what the difference is.
Randy
I still think its some weird ground/return problem, but no idea what is causing it.
I've made many mods to my soekris dac
My soekris mod blog
and I think it sounds pretty good now. There are actually more changes I made recently, new usb to i2s converter, added potato flip flop on i2s lines to reclock them, and removed the isolators.
But I have a lot more time and money in that setup, so will see how this little guy compares.
I only tried this new dac with battery power. When I have some time, after it settles in, I can disconnect the battery and see what the difference is.
Randy
Thanks for looking at my video. I wish I knew enough about electronics to understand why the fix works...I have a fine arts degree...this is all black magic to me!
I have been trying to guess about why the noise exists in the first place. Is it the DAC as I have assumed? Maybe I have a defective iPower instead. Or perhaps not all the GPIO pins are in full contact. I had to work hard to seat the DAC header onto the Kali. Whereas the Kali mounted flush with the Pi, I ended up with a few mm gap between Kali and DAC (I decided against using a mallet).
I have been trying to guess about why the noise exists in the first place. Is it the DAC as I have assumed? Maybe I have a defective iPower instead. Or perhaps not all the GPIO pins are in full contact. I had to work hard to seat the DAC header onto the Kali. Whereas the Kali mounted flush with the Pi, I ended up with a few mm gap between Kali and DAC (I decided against using a mallet
).I have been trying to guess about why the noise exists in the first place. Is it the DAC as I have assumed? Maybe I have a defective iPower instead. Or perhaps not all the GPIO pins are in full contact. I had to work hard to seat the DAC header onto the Kali. Whereas the Kali mounted flush with the Pi, I ended up with a few mm gap between Kali and DAC (I decided against using a mallet
I never had any mating problems on the GPIO pins. But indeed, I did have a bit of a gap between Kali and DAC. I think the Kali just has longer pins than the RPI.
In your video, when you get that perfect silence with the white three-prong: I can't remember, have you tried the iPower ground mod you mentioned before? You could whip up a ghetto one real quick with a paperclip and spare three-conductor power cord.
Also: I doubt the Kali is the problem, but take it out and see if the noise is still there. I assume it will be. But now, use something else to power the RPI + DAC. Surely you have a spare mini-USB cellphone charger. Then you can see if indeed that iFi power supply has issues. Also try powering using a USB cable from your computer to the RPI+DAC.
I'm trying to think of another common 5V power supply you might have in your house... do you have any spare PC power supplies? Those have 5V rails (google "ATX power supply pinout" for which pins to use).
In other news I moved the balanced version from the test rig to my main rig. I feel like this smooths things out a bit (could be my imagination). But I did quite a bit of listening yesterday and was really happy with the sound. Shorter XLR cables are on order.
Attachments
Thanks for looking at my video. I wish I knew enough about electronics to understand why the fix works...I have a fine arts degree...this is all black magic to me!
I have been trying to guess about why the noise exists in the first place. Is it the DAC as I have assumed? Maybe I have a defective iPower instead. Or perhaps not all the GPIO pins are in full contact. I had to work hard to seat the DAC header onto the Kali. Whereas the Kali mounted flush with the Pi, I ended up with a few mm gap between Kali and DAC (I decided against using a mallet
With floating PS mixed with mains connected PS it always gets complicated but it seems by touching the earth pin of the mains plug to the ground of the Pi DC input, you are providing a low impedance route for the noise (probably leakage current noise) back to earth which avoids it contaminating the signal ground. You can just do this directly without using the earth pin of the Mac or monitor - just use a wire between that point & mains earth.
Last edited:
StellarElephant,
It might be worthwhile to try some jumpers like:
http://a.co/fWg76Ty
to connect your Kalli away from the pi, etc. In the very least you could rule out any RFI/EMI from the Pi, etc.
Of course you could pickup noise in the now long wires between them....
Good luck tracking it down!
Gable
It might be worthwhile to try some jumpers like:
http://a.co/fWg76Ty
to connect your Kalli away from the pi, etc. In the very least you could rule out any RFI/EMI from the Pi, etc.
Of course you could pickup noise in the now long wires between them....
Good luck tracking it down!
Gable
I think @mmerrill99 probably has it right - your noise problem is most likely due to leakage currents being introduced by an offline switching power supply. The best way to deal with such an issue is to introduce just one mains earth at the source of the leakage currents. However this might be impractical as things like monitors have their own earth connection and having two mains earths might end up being worse
so looking at alternate charging methods.
This looks promising
lifepo4 battery charger
uses one tl431 to set the voltage per cell. I've found similar circuits that use LM317's instead of the tl431's.
Randy
This looks promising
lifepo4 battery charger
uses one tl431 to set the voltage per cell. I've found similar circuits that use LM317's instead of the tl431's.
Randy
So I was wrong about the batteries, I hate when I'm wrong lol
One battery is at 3.3v, and other is around 1.7v.
So, now I need two 2.5V isolated supplies
Need a little dual secondary tranny, and then I'll rig something up.
yea its the 5 v in series that worries me the most been the lower end of the range of 2.5v
wonder in a floating situation can it be biased with resistors