The longer cables won't "cause" more noise, they will just be antennas tuned to longer wavelengths. It is fair to say that longer wires are likely to pick up more noise because they are sitting in more fields than shorter ones. But with unshielded wires anything can happen. Don't forget that the WaveIO and the Kali and the Pi will all be radiating EMI with different characteristics, creating their own fields. You could just be unlucky with wire lengths. I found shielding the I2S lines (grounded at both ends) from my Kali cleaned up the sound a good chunk.
I only use these shielded wires (bought on ebay and removed the u.fl. connectors).
Thanks for the advice with MCLK u.fl out from Kali. Wonder why it is not mentioned in Kali SPEC sheet..?? I already considered using it but was not sure about..
Nevertheless I had this hiss noises and lateley my Kali does not even boot or find a incoming signal when free wired with this wires..only when sitting directly on a pi with a piano on top it works...very strange..
Thanks for the advice with MCLK u.fl out from Kali. Wonder why it is not mentioned in Kali SPEC sheet..?? I already considered using it but was not sure about..
Nevertheless I had this hiss noises and lateley my Kali does not even boot or find a incoming signal when free wired with this wires..only when sitting directly on a pi with a piano on top it works...very strange..
Not sure, if it is a similar issue, but have you ever tried to install some in-line resistors within the i2s lines?
Have had an issue with unstable playback / crackle with kali in combination with an isolator chip. Odroid -> Isolator -> Kali -> DAC. With the Isolator installed I needed the resistor in the i2s lines to get a stable playback. Without Isolator in chain no resistors needed. And also not with setup Odroid -> DAC. And also not with setup Odroid -> Isolator -> DAC... strange but true...
For me it looked like the kali needs the resistors in line for specific i2s partners (in my case the isolator chip). As I do not use any isolator anymore I have no resistors installed at the moment ...
It's worth a try...
see: here
To be honest, I don't know the best place for them. In my test they were placed close to the source (Isolator in my case) and it worked (have only done just a quick test without thinking to much about it).
BUT in the practical examples I have found afterwards (Wave IO board - isolator chip; and DDDAC board - DAC chip input) the resistors are more placed near the destination. In the thread I have linked there is something written about termination which tells me to place them at the end of the line to prevent reflections. Would fit to the examples I have found...
My Kali works again! It was a broken LRCLK wire that did not make it boot...
I also applied 47 Ohm SMD resistors at the Kali input side. First there still was no sound. Than I removed the self made support board with connectors I created for fixing the Kali on my DAC and replaced it with free jumper wires - et voila!
Maybe a connection issue or bad wiring on the board with interference and EMI problems - no idea...
I also replaced the MCLK jumper wire with a u.fl wire from the Kali u.fl terminal to the DAC.
Hiss is still there, but less than before. I tested a direct wire connection from DAC GND to Kali GND, and it reduced the hiss significantly, but only when I touched the connection wire (isolated wire)...
At least happy my Kali still works, but SQ improves not so much compared to the direct I2S connection between hifiberry dac+pro and DAC, so maybe I will keep the set-up w/o Kali and with no hiss noise..
I also applied 47 Ohm SMD resistors at the Kali input side. First there still was no sound. Than I removed the self made support board with connectors I created for fixing the Kali on my DAC and replaced it with free jumper wires - et voila!
Maybe a connection issue or bad wiring on the board with interference and EMI problems - no idea...
I also replaced the MCLK jumper wire with a u.fl wire from the Kali u.fl terminal to the DAC.
Hiss is still there, but less than before. I tested a direct wire connection from DAC GND to Kali GND, and it reduced the hiss significantly, but only when I touched the connection wire (isolated wire)...
At least happy my Kali still works, but SQ improves not so much compared to the direct I2S connection between hifiberry dac+pro and DAC, so maybe I will keep the set-up w/o Kali and with no hiss noise..
hi nuno, didn´t try it. before, i used the wave IO and a mac-laptop. from this it was a big improvement in transparency and precision and bass.
regards, mario
Hi pistollero
I got Hifiberry digi+ pro and Ian's pi isolator. I tried with and without isolator. I powered pi and hifiberry digi+ pro with separate high quality 5V linear regulators. Eventually in my conditions I use hifiberry digi+ pro without isolator, since isolator did not bring any improvement. You need to remember that isolator brings jitter to the singnal, so it would have to be reclocked afterwards, otherwise isolator may make more bad that good to the signal.
Uhm, I thought the I2S output of the Hifiberry digi+ already has its own isolation? I don't think adding another isolator would add much.
Hifiberry digi+ pro does not have isolation nor reclocking on I2S line. In regards of I2S it just supplies pi with external audio frequency clocks and make synchronous clocking with your destination device (dac) possible. This is already a lot!
Last edited:
I just checked here. The DIGI+ PRO does have galvanic isolation on the I2S output, and it does have its own clock(s).
It is galvanic isolation for coax with use of transformer and obviously for toslink because it is light, but not for I2S. Hifiberry I2S shares common ground with RPI and any dac connected to it.
Last edited:
Yes i think you are right... because of that i was thinking of using allo isolator because of that.
Ioan,
I for one am eagerly awaiting this. While there are other R-Pi-based DACs with pretensions to high-ish-end (I'm thinking fo the Terra-BerryDAC2 & the TauDAC as ones that appear to actually have made it to release), none combine the array of features, flexibility, & reasonable cost that you're projecting in this DAC.
I'm excited by:
1. Master/Sync mode operation with local clocks. Every ESS DAC I've run has sounded more real and alive in 'sync' mode instead of the datasheet ASRC mode.
2. Multiple-board stacking. One of the biggest issues with getting great performance out of an RPi DAC is the limited board space. Stacking gives you much more flexibility to the manufacturer do great things (such as better output stages and better power) by providing more circuit board real-estate... along with...
3. Flexibility. By separating the main board, output board and power board, you provide for a level of flexibility in options, upgrades, and meddlesome tweaking and modifying like I am wont to do.
4. Discrete output stage. The main limitation with DACs that use chips such as the ES9023 & PCM51xx series is being limited to using the on-chip output stages. My current fav RPi DACs are the Dial DAC (both the stock and my HotRod'd version) and Ian Canada's ES9018K2M (I have a prototype from Ian here... and yes, I prefer it in sync mode). IMHO, their main performance advantage comes from their separate output stages (even though the Dial DAC's PCM1793 does use on-chip I/V and Ian runs his at lower-than-I-like +3.3v only). To my ear, the separate O/P stages give these DACs a leg up in dynamics and overall realism over ones using on-chip outputs only. This new Allo.com DAC not only uses a separate output stage, but takes it a couple of levels further with a discrete outputs (and Sparkos discrete opamps have faired well in the marketplace).
5. Use of the latest ESS chips. EVEN running in sync mode, the earlier ESS chips (ES9018, ES9022, ES9023, ES9018K2M) have never given me the level of realism I get from the better R2R or R2R-hybrid setups such as the TDA1541, TDA1543, Soekris DAM, PCM1795, and PCM1793. BUT the what I'm hearing about the latest ESS chips has me very hopeful they will provide the best of both worlds, if used correctly. AND it sounds as though Allo is making a serious attempt to use them correctly!
6. Good power. Even back with their original Piano DACs, Allo did power right with industry leader LT3042 regulators in critical areas. Their Boss DAC took this a couple of steps further. While I am not a fan of DC-DC converters, I suspect that done well, they can do the job. AND if not, well, that multi-board architecture means I can more easily modify with all linear regulation (but now requiring +-18V-24V instead of the standard +5V most like to see here).
I'm hopeful that it will work well with the isolator boards available... I consider them a must for getting the absolute best out of an RPi setup. Or maybe incorporate that on-board.
AND I'm also hopeful Allo is doing local reclocking of the data stream before feeding it to the DAC chip. Even running in master/sync mode and with local clocks, the data feeds from the Pi won't be super-low jitter, especially after passing through isolators.
Also, I'm hopeful that the control options for the ESS DAC have been carefully selected and can be easily field-updated (hopefully via the driver). There are some good comments by Charles Hansen of Ayre on this Computer Audiophile thread on getting the best sound out of the ESS chips.. and careful selection (read that turning off many of the options) of the chip setup is key:
Ayre Acoustics QX-5 Twenty – The Digital Hub - DAC - Digital to Analog Conversion - Computer Audiophile
AND finally, I'm hopeful that Allo has provided for optional balanced output mode, even if just points to solder in leads to separate XLRs.
I'm expecting this will be something even more special than the Piano and Boss DACs and their Kali and Isolator boards.
Greg in Mississippi
P.S. Ioan, any other features added to make me more excited?
I for one am eagerly awaiting this. While there are other R-Pi-based DACs with pretensions to high-ish-end (I'm thinking fo the Terra-BerryDAC2 & the TauDAC as ones that appear to actually have made it to release), none combine the array of features, flexibility, & reasonable cost that you're projecting in this DAC.
I'm excited by:
1. Master/Sync mode operation with local clocks. Every ESS DAC I've run has sounded more real and alive in 'sync' mode instead of the datasheet ASRC mode.
2. Multiple-board stacking. One of the biggest issues with getting great performance out of an RPi DAC is the limited board space. Stacking gives you much more flexibility to the manufacturer do great things (such as better output stages and better power) by providing more circuit board real-estate... along with...
3. Flexibility. By separating the main board, output board and power board, you provide for a level of flexibility in options, upgrades, and meddlesome tweaking and modifying like I am wont to do.
4. Discrete output stage. The main limitation with DACs that use chips such as the ES9023 & PCM51xx series is being limited to using the on-chip output stages. My current fav RPi DACs are the Dial DAC (both the stock and my HotRod'd version) and Ian Canada's ES9018K2M (I have a prototype from Ian here... and yes, I prefer it in sync mode). IMHO, their main performance advantage comes from their separate output stages (even though the Dial DAC's PCM1793 does use on-chip I/V and Ian runs his at lower-than-I-like +3.3v only). To my ear, the separate O/P stages give these DACs a leg up in dynamics and overall realism over ones using on-chip outputs only. This new Allo.com DAC not only uses a separate output stage, but takes it a couple of levels further with a discrete outputs (and Sparkos discrete opamps have faired well in the marketplace).
5. Use of the latest ESS chips. EVEN running in sync mode, the earlier ESS chips (ES9018, ES9022, ES9023, ES9018K2M) have never given me the level of realism I get from the better R2R or R2R-hybrid setups such as the TDA1541, TDA1543, Soekris DAM, PCM1795, and PCM1793. BUT the what I'm hearing about the latest ESS chips has me very hopeful they will provide the best of both worlds, if used correctly. AND it sounds as though Allo is making a serious attempt to use them correctly!
6. Good power. Even back with their original Piano DACs, Allo did power right with industry leader LT3042 regulators in critical areas. Their Boss DAC took this a couple of steps further. While I am not a fan of DC-DC converters, I suspect that done well, they can do the job. AND if not, well, that multi-board architecture means I can more easily modify with all linear regulation (but now requiring +-18V-24V instead of the standard +5V most like to see here).
I'm hopeful that it will work well with the isolator boards available... I consider them a must for getting the absolute best out of an RPi setup. Or maybe incorporate that on-board.
AND I'm also hopeful Allo is doing local reclocking of the data stream before feeding it to the DAC chip. Even running in master/sync mode and with local clocks, the data feeds from the Pi won't be super-low jitter, especially after passing through isolators.
Also, I'm hopeful that the control options for the ESS DAC have been carefully selected and can be easily field-updated (hopefully via the driver). There are some good comments by Charles Hansen of Ayre on this Computer Audiophile thread on getting the best sound out of the ESS chips.. and careful selection (read that turning off many of the options) of the chip setup is key:
Ayre Acoustics QX-5 Twenty – The Digital Hub - DAC - Digital to Analog Conversion - Computer Audiophile
AND finally, I'm hopeful that Allo has provided for optional balanced output mode, even if just points to solder in leads to separate XLRs.
I'm expecting this will be something even more special than the Piano and Boss DACs and their Kali and Isolator boards.
Greg in Mississippi
P.S. Ioan, any other features added to make me more excited?
Last edited: