Asynchronous I2S FIFO project, an ultimate weapon to fight the jitter


I searched the forum and didn’t find an answer to this question (apologies if I missed it).

My current setup is simple:
RPI 4b + ProtoDAC (TDA1387 x8)
(Running moOde + Roon bridge)
(Power: iFi Power 5V -> RPI -> ProtoDAC via GPIO)

I plan to improve it with:
1-Cleaner power (several options), and
2-FifoPi Q7 reclocker.

Inserting the Q7 between RPI and ProtoDAC requires an adapter with GPIO pins 2-4 removed in order to provide 5V to the DAC, as mentioned earlier on this forum.

So I need 3 clean power sources:
1- 5V to RPI
2- 5V to ProtoDAC
3- 3.3V to Q7

If money (and space) was no object, I could use 3 UcPure packs.

If not, perhaps one UcPure MkIII 5V w parallel DC cables to the RPI and DAC, and one UcPure MkIII 3.3V for the Q7. (Any comments on using parallel - or Y cables?)

If not, then I’m wondering: Where is it the most critical to have clean (no ripple) and highly available (low ESR, high Amperage) power? My guess would be at the DAC, because when the sound signal changes abruptly, the OpAmp needs to be able to feed the I/V stage without being limited by a PS that cannot cope with the sudden current demand.

On the other hand, the RPI’s job is to feed the bit stream to the I2S interface, which is not particularly demanding in terms of Amps, but could still benefit from a really quiet DC source.

If my reasoning is right, I could use my iFi Power 5V on the RPI, and use UcPure packs for the DAC and Q7.

I thank you in advance for your comments!
  • Like
Reactions: 1 user
I think you best Value for this project is a PurePi also if using a Fifo Pi and Protodac I discovered a slightly better way see My Video coming this Friday you may find it helpful as I talk about how to combine a Fifo and Protodac in a far better way and how 3 different configurations Sound.
Last edited:
  • Like
Reactions: 3 users
@Trebtek @Gabster 2000 I also have a pretty high tech version of Ian Canada's streamer, including PureStationPi Pro, two LinearPi, three UCconditioners, a FifoPi Q7 and TransportPi AES. It works extremely well with my R2R NOS Lab12 DAC. However, I was curious about the ProtoDAC and I am going to test it against my Lab12 DAC. I am not yet sure where I can place the ProtoDAC inside the streamer box, because it's full with components, but I will sort something out. In terms of power supply, I am thinking of using my 'middle' UCconditioner 5V (see attached photo), which is currently powering the clean side of my StationPi Pro. So, that UCconditioner will hopefully power both the clean side of the StationPi Pro and the ProtoDAC. I hope that should work. Any comments, @hifiamps @Gabster 2000 ?

I also happened to watch Gabi's recent video about using three UFL cables to connect the ProtoDAC with the FifoPi Q7.
That's something I am considering. And if can have the three UFL connections of the FifoPi Q7 connected to the ProtoDAC and the MasterClock of the FifoPi connected to the TransportPi AES, I guess I will be able to run a direct (simultaneous) A-B comparison test between my ProtoDAC and my Lab12 DAC. That would be interesting...


  • IMG_9998.jpeg
    451.8 KB · Views: 111
  • IMG_0151 (1).jpeg
    IMG_0151 (1).jpeg
    259.8 KB · Views: 106
  • Like
Reactions: 5 users
That should work. Love those Jantzen superior caps, also would love to hear you feedback on sound comparison.
Also please Note I did a correction on the order of the I2S signal thanks to Michael as the resistors and the Chip are crossed so
at 13:16 Video should say WS DATA BCK (NOT WS BCK DATA)
If connecting on the Resistor end, like in the video the Correct order is WS DATA BCK as resistors are crosslinked not as per schematic. If you are connecting on the Chip end then the order as per video is correct.
  • Like
Reactions: 2 users
Ok So I checked the Grounding on the Pure PI / FifoPi (Ref Protodac Fifo PurePi combination) @pkonstantinidis
You will need to Connect the I2S wires to Ground as the ground 5v on the Pure Pi is not connected to the Fifo UFL ground because that ground comes from the 3.3v side. (Also remember we are no longer connect to any GPIO pins)
The 2 power supplies are galvanically isolated and do not share ground I think Mark's concerns are valid but only if you are sharing the ground from the same PSU, but not in this case.
I wonder if that also contributed to the better sound.
It is nice to see concerns as that helps us further understand and double check our steps.
hope that helps demystify this always complicated ground subject :)

Also please Note I did a correction on the order of the I2S signal thanks to Michael as the resistors and the Chip are crossed so
at 13:16 Video should say WS DATA BCK (NOT WS BCK DATA)
If connecting on the Resistor end, like in the video the Correct order is WS DATA BCK as resistors are crosslinked not as per schematic. If you are connecting on the Chip end then the order as per video is correct.
Last edited:
  • Like
Reactions: 2 users
Thanks for the clarification about the soldering sequence of the i2S cables, @Gabster 2000 . Duly noted. And based on your comments and also those of Mark (@hifiamps ), who is the designer of the ProtoDAC, if I have not misinterpreted anything, I should solder the sheath of the three i2S cables on pin 6 of the ProtoDAC and then connect that pin with the respective pin on the GPIO of the FifoPi Q7. I can use a Dupont cable for that. Can you confirm, please?

Here is a recent photo of my streamer, having added a MonitorPi Pro (with an extension flat band) and a push-on switch on the semi-transparent front face.

I will probably put the ProtoDAC in a separate, external small metal box, to keep the ProtoDAC shielded. Unfortuantely it cannot fit in my streamer box. At the same time, I suppose I should keep both the I2S cables and the power cables from the (middle) UCconditioner 5V as short as possible.


  • IMG_0173.jpeg
    599.9 KB · Views: 93
  • Like
Reactions: 2 users
After waiting a long time to get the desired enlclosure I could finally start building.

I use a piece of bamboo kitchen cutting board as a base and install some dividers to make 3 compartments.
I ordered a Modu Galaxy enclosure of 170 mm deep and as I had an unused 230 mm deep enclosure I could use this to cut to 170 so have 4 side panels instead of 2.

Cutting to length:





Test fitting with left a linear PS, middle the PurPi with Raspberry and rigth the FifoPi Q7 and TransportPi AES


To connect the FifoPi to the RPi I had to make a custom length flat cable.
Started with 20 cm and measure where to put another connector an cut the rest.


The flat cable goes through a ferrite core to reduce EMC even more.
I had to cut a divider but unfortunate broke it but can still use the broken piece.






Unfortunate the connector at the FifoPi side blocks the holes for the studs so had to cut the ends to make it fit and clear the holes.





The UcConditioner has a small issue that when using the Maxwell ultra capacitors the treaded stud interferes with the capacitor.



Putting everything together it looks like this ... the power supply board is not finished yet but wanted to share my results so far.
I guess another 1-2 weeks I should be able to get is up and running.

It will get a front and rear panel and a slotted top cover for ventilation.
I realise I also better drill some holes in the bamboo at the power supply and PurePi/RPi section to get some airflow.






  • Like
Reactions: 8 users
I just got a Raspberry Pi5 to use as a streamer with my DAP. My DAP only has a USB input. Would there be any sonic benefit to using a PurePi II to power the Raspberry since I am using the USB output of the Raspberry Pi?
I guess I should mention that I am using a USB drive to store my music.
Eventually I will get either a standalone dac or a dac hat.