I don't know if you followed the PLL discussion in the 1021 thread but the findings there indicate that there is no stable clock available in soekris DACs that do not have an in-built USB function. In general, its not the timing of the data that is crucial but the quality of the clock signals. I suppose you are somehow trying to make the i2s source a slave? While that is a good idea, one should have a stable clock in the master - but alas, such clock does not exist due to the constant changing of the Si clock performed by logic inside the DAM.
//
//
Yes that's what I was trying to do. Thanks for clearing up that this won't with the programmable clock.
For future reference, another thing I've learned is that serial communication doesn't work when the dam's digital ground is galvanically isolated from the RPi ground. At least, not IanCanada's IsolatorPiII: connecting Tx/Rx to GPIO 15/14 (on the non-isolated header, the isolated header doesn't pass through those pins) results in gibberish on the terminal. Of course the voltage reference isn't the same.
Removing the isolator board, then connecting the I2S and serial pins works fine. Makes sense that there needs to be a common ground. Commoning the ground on the isolated and non-isolated side (defeating the purpose of the isolator board -- but hey, just tinkering here) strangely doesn't work.
Originally I had planned for digital volume control from moOde using serial communication, but here I am trading off galvanic isolation (but with software volume control) and hardware-based volume control (but with RPi's noise on ground). Since I have a stepped attenuator on my b22 headphones / pre-amp, I think I'll pick the isolation and mostly run the dam at -2 dB. Gives the best dynamic range on my unbalanced interconnection anyway.
Lastly, thanks to Søren for making this board available to us, and this community for sorting it all out and making it even better. I uploaded your filter pack TNT, set the C128dp at -2 dB to default and am thoroughly enjoying this. Also shouts out to nautibuoy for sending one of his spare dam1121 interface boards to me. You guys rock.
For future reference, another thing I've learned is that serial communication doesn't work when the dam's digital ground is galvanically isolated from the RPi ground. At least, not IanCanada's IsolatorPiII: connecting Tx/Rx to GPIO 15/14 (on the non-isolated header, the isolated header doesn't pass through those pins) results in gibberish on the terminal. Of course the voltage reference isn't the same.
Removing the isolator board, then connecting the I2S and serial pins works fine. Makes sense that there needs to be a common ground. Commoning the ground on the isolated and non-isolated side (defeating the purpose of the isolator board -- but hey, just tinkering here) strangely doesn't work.
Originally I had planned for digital volume control from moOde using serial communication, but here I am trading off galvanic isolation (but with software volume control) and hardware-based volume control (but with RPi's noise on ground). Since I have a stepped attenuator on my b22 headphones / pre-amp, I think I'll pick the isolation and mostly run the dam at -2 dB. Gives the best dynamic range on my unbalanced interconnection anyway.
Lastly, thanks to Søren for making this board available to us, and this community for sorting it all out and making it even better. I uploaded your filter pack TNT, set the C128dp at -2 dB to default and am thoroughly enjoying this. Also shouts out to nautibuoy for sending one of his spare dam1121 interface boards to me. You guys rock.
Well, I'm disappointed you didn't choose the linear 44,1 because thats what I created and all the other filters are made by others ;-) - but you being satisfied is of course the most important 
As I see it the isolation should only effect the dc levels behind the isolation but restore the difference between the 1s and a 0s and as such, should work logically - this is the whole point with the isolators - to galvanically separate the grounds. Something is strange...
//
As I see it the isolation should only effect the dc levels behind the isolation but restore the difference between the 1s and a 0s and as such, should work logically - this is the whole point with the isolators - to galvanically separate the grounds. Something is strange...
//
Well there are different power supplies. Today's configuration:
Tomorrow's configuration (there's a package waiting for me at the post office):
In either setup, the serial Tx/Rx is between the non-isolated RPi GPIO header and isolated dam1121 J2, and ground is not common so neither is the voltage reference. But any tips are more than welcome!
Code:
- RPi switching mode wall wart --(5V)--> IsolatorPiII J2 --(non-isolated 5V)--> RPi non-isolated GPIO header
- Trafo #1 both secondaries --> AMB s22 --(5,1V)--> IsolatorPiII J1 --(isolated 5V)--> dam1121 J3
\--(+/-5,1V)--> dam1121 J1Tomorrow's configuration (there's a package waiting for me at the post office):
Code:
- Trafo #1 both secondaries --> AMB s22 --(+/-5,1V)--> dam1121 J1
- Trafo #2 one secondary --> DIYINHK LT3045 --(5V)--> IsolatorPiII J2 --(non-isolated 5V)--> RPi non-isolated GPIO header
other secondary --> DIYINHK LT3045 --(5V)--> IsolatorPiII J1 --(isolated 5V)--> dam1121 J3In either setup, the serial Tx/Rx is between the non-isolated RPi GPIO header and isolated dam1121 J2, and ground is not common so neither is the voltage reference. But any tips are more than welcome!
Sure but who is to say the 0-3,3V on the RPi side is the same 0-3,3V as on the dam1121 side? It almost certainly isn't... and is the only thing I can think of!
Yes the LT3045 is overkill but the DIYINHK kits are very well priced. They don't stock LT3042's anymore either *and* the LT3045 version can burst to 2A (while it sustains 500 mA) whereas previous versions did so up to 1-1,5A. Don't know of many other kits that can source such current (sure the AMB s11, TPA kits or LinearPi... speaking of overkill!) Also the DIYINHK kits have a nice Murata ground choke that further stops the dirty side from contaminating the mains power.
This specific board has an optically-isolated on/off control, but the I2S isolator seems to be an IC. Manual including block diagram here: DocumentDownload/IsolatorPiIIUsersManualRev2.0.pdf at master * iancanada/DocumentDownload * GitHub
Yes the LT3045 is overkill
but the DIYINHK kits are very well priced. They don't stock LT3042's anymore either *and* the LT3045 version can burst to 2A (while it sustains 500 mA) whereas previous versions did so up to 1-1,5A. Don't know of many other kits that can source such current (sure the AMB s11, TPA kits or LinearPi... speaking of overkill!) Also the DIYINHK kits have a nice Murata ground choke that further stops the dirty side from contaminating the mains power.This specific board has an optically-isolated on/off control, but the I2S isolator seems to be an IC. Manual including block diagram here: DocumentDownload/IsolatorPiIIUsersManualRev2.0.pdf at master * iancanada/DocumentDownload * GitHub
Last edited:
No - you did B, didn't you?
And there not the same reg as in 1021. And there seem to be a filter here which is not in the 1021... so improved in 1121... maybe not much to gain here.
I'm hesitating....
//
In a state of covid bordom I did put Sanyo 330uF polymer cap on the clock power feed. Lived with it for a week and now just took it away. I found that this alteration lowered the hight of the sound stage and made the higher frequency a bit aggressive. Off they go.
//
Myself I’m a DIYINHK LT3045 regulator (sold as a pair, hooked the other one up to my RPi) for the digital section and a single AMB s22 tracking dual-rail PSU for the analog section. The first is a full kit, the second requires some shopping on Mouser or sorts, but a full parts list is provided so is easy enough to do.
Yep, the LDO offerings are excellent. If you wan't to keep the installation simple they have modules available that only need the addition of a suitable transformer.
Power Supply
I suggest this one for the 5V digital section;
Single Rail TPS7A4700 Ultra Low Noise</br>Power Supply
and this for the analogue output section;
LT3045/LT3094 Ultra Low Noise Dual Polarity Power Supply
You might choose to use additional power supplies for the two sections of the JLSounds USB board I think you plan to use?
Those boards do look good. I think this is the path for me. However, as far as i can tell, I can't directly connect a transformer to them. They still need a DC input. Any recommendation on a power supply to add before them?
As for the USB side, I plan on using the JLSounds board with its own dedicated power supply (2 separate boards).
Did anyone try to add Ian's FIFO board to a DAM 1121?
The modules I suggested have rectifiers and filter caps onboard and only require a transformer - read the specifications. I use them in some of my projects so have first-hand experience - they work well.
Just get an extra couple of the +5V units.
No, but if you plan to use the JLSounds USB board why would you need to? The JLSounds board has isolation and reclocking. Also, IIRC, the DAM1121 has onboard reclocking after the FPGA.
Last edited:
Looking beyond powering your DAM1121, you could do worse than getting an AudioZen SK Lite to control it;
AZ OLED setup kits for Soekris dam1021/1121 control
I know, right! So you're still holding onto your dam1121? I thought I read in some other thread you had sold it but maybe I was mistaken.
As for the external reclocker, yeah, the difference is not subtle. I was pretty skeptical about it. I mean, the dam already has a FIFO buffer on it, right? I'm something of a "rational audiophile", in there between objectivism and subjectivism, and felt I was taking a gamble. But sure enough the difference is very audible, especially in stuff like brass sections and distorted guitars.
So you're still holding onto your dam1121? I thought I read in some other thread you had sold it but maybe I was mistaken.As for the external reclocker, yeah, the difference is not subtle. I was pretty skeptical about it. I mean, the dam already has a FIFO buffer on it, right? I'm something of a "rational audiophile", in there between objectivism and subjectivism, and felt I was taking a gamble. But sure enough the difference is very audible, especially in stuff like brass sections and distorted guitars.
I still have the DAM1121. I'm actually in the process of repackaging it into a new chassis - building it into the 300B headphone amp was a mistake in limiting use of both the DAC and the HPA so each will be its own chassis before too long (quicker if I don't spend too much time reading stuff that is a distraction...).
I know, right!
As for the external reclocker, yeah, the difference is not subtle. I was pretty skeptical about it. I mean, the dam already has a FIFO buffer on it, right? I'm something of a "rational audiophile", in there between objectivism and subjectivism, and felt I was taking a gamble. But sure enough the difference is very audible, especially in stuff like brass sections and distorted guitars.
Hi Roderickvd,
Can you share some info on how you implemented Ian's FIFO board with the DAM 1121?
I also think I don't truly understand the use of RPi in these builds. Is it there to create a steamer-DAC combo, or does it serve another purpose? If I plan on using my computer as its source while feeding my DAC through USB - should i still consider using an RPi in my build?
Tank.
Last edited: