I have sine solved the problem with a few changes
1) the laptop power supply I was using was extremely noisy the worse I have seen. They are not all equal. over 200mv of complex ripples a Toshiba one I had was a night and day different 5mv. I have since switched to a low noise linear power supply off ebay to be on the safe side.
Even if your system is running on batteries a really bad switching power supply running in the Lifepo4 could send waves in your system even your entire house.
2)I installed a dedicated line for cleaner ground with no switching supplies on that line.
3) used a star grounding system to avoid ground loops
I grouped the ground in the streamer I am building in 2, Chassis ground that is grounded to house power line earth, and power ground. Now it will be easier to AB test (connect power ground to earth ground or not) so far it is close I will do some more sound tests later and decide may connect them via a parallel (resistor and x2 capacitor)
After the positive comments from Greg about the huge 3000F ultracapacitors I tried them myself.
I got me some AVX samples and tapped M4 thread in the poles (after measuring the thickness with a metal thickness measuring instrument from work) to connect the wires.
I also tried aluminium welding some tabs but the aluminium is not a weldable type.
The UcConditioner is attached and charging needed about 4 hours.
They are now feeding my USBridge Sig.
This is real overkill but I must say that I hear a positive effect in comparison to the previous setup with 2 UcConditioners in parallel (to the right in the picture)
The sound seems a bit more open and transparant, low registers are a tad more detailed.
Are there others that have experimented with these big caps?
Regards,
I got me some AVX samples and tapped M4 thread in the poles (after measuring the thickness with a metal thickness measuring instrument from work) to connect the wires.
I also tried aluminium welding some tabs but the aluminium is not a weldable type.
The UcConditioner is attached and charging needed about 4 hours.
They are now feeding my USBridge Sig.
This is real overkill but I must say that I hear a positive effect in comparison to the previous setup with 2 UcConditioners in parallel (to the right in the picture)
The sound seems a bit more open and transparant, low registers are a tad more detailed.
Are there others that have experimented with these big caps?
Regards,
Attachments
If the 19V DC power input is too noisy, you need to use nylon standoffs to isolate the negative terminal to the case. And also, disable the 5V rail. You can use a LinearPi as RPi 5V power supply. Even though, RPi ground still need to be isolated from the case ground which normally be connected to the battery ground.
Ian
Hi all,
I'm running Pi Core Player on a Pi 3B+, using Ian's Fifo Ultimate Q2 for a couple of years. This initially initially fed an Allo Boss v1 with Fifo mclk input, currently Ian's older 9038 Dual Mono and Bisesik's internationally-traveled transformer output (subjectively, audio quality was Dual Mono/Transformer IV > Boss with mclk input > Boss as master). The Pi is powered by a generic wallwart, Ian's boards take 3.3v and 5v from a toroidal transformer through DIYINHK's dual 0.56uV linear regulator board.
I'd like to upgrade my power supplies to a LiFePO4 board, replace the Pi with an Allo USBridge Signature, and possibly include the ultra-capacitor boards (if my old ears are continuing to hear improvements); so I have a couple of questions regarding use of the power supply that I haven't found answers to in many hours of (re)reading through these incredible forum threads:
1. Is anyone powering a USBridge Sig via the non-isolated 5V 2A output? The only reference I've seen on the Allo site is for a 3A supply, however I'll be powering other boards independently rather than through the USBridge which may reduce potential maximum current draw;
2. Reading suggests I should use 6 separate 3.3v rails for the remainder of the stack: main power input (with onboard ldo mod's) and 2 clocks for Fifo Q2; avcc, dvcc, and vcca for the DAC. Given we're feeding clean battery power, are others readily noticing an improvement with separate supplies compared with just the single input for each board? I'm trying to decide if I go straight to the 6 separate rails, or start with just 2 and experiment from there...
Regards, Simon
I'm running Pi Core Player on a Pi 3B+, using Ian's Fifo Ultimate Q2 for a couple of years. This initially initially fed an Allo Boss v1 with Fifo mclk input, currently Ian's older 9038 Dual Mono and Bisesik's internationally-traveled transformer output (subjectively, audio quality was Dual Mono/Transformer IV > Boss with mclk input > Boss as master). The Pi is powered by a generic wallwart, Ian's boards take 3.3v and 5v from a toroidal transformer through DIYINHK's dual 0.56uV linear regulator board.
I'd like to upgrade my power supplies to a LiFePO4 board, replace the Pi with an Allo USBridge Signature, and possibly include the ultra-capacitor boards (if my old ears are continuing to hear improvements); so I have a couple of questions regarding use of the power supply that I haven't found answers to in many hours of (re)reading through these incredible forum threads:
1. Is anyone powering a USBridge Sig via the non-isolated 5V 2A output? The only reference I've seen on the Allo site is for a 3A supply, however I'll be powering other boards independently rather than through the USBridge which may reduce potential maximum current draw;
2. Reading suggests I should use 6 separate 3.3v rails for the remainder of the stack: main power input (with onboard ldo mod's) and 2 clocks for Fifo Q2; avcc, dvcc, and vcca for the DAC. Given we're feeding clean battery power, are others readily noticing an improvement with separate supplies compared with just the single input for each board? I'm trying to decide if I go straight to the 6 separate rails, or start with just 2 and experiment from there...
Regards, Simon
I keep thinking I'm in over my head on this LifePO4 power supply situation. I'm moderately good at soldering and have a good iron and all the requisite materials, but having just bought a mk III from audiophonics - I keep seeing references to soldering in battery trays and holders; The unit I have appears to already have all that done.
Am I wrong to think that if I just need to use:
1 x 3.3V (FIFOPi clean side)
1 x 5V (FIFOPi dirty side to power Pi)
possibly 1 x 5V to power HDMIPi
That all I really need to do with the board I have is solder in the knob button per the instructions, install 6 LifePO4 batteries in proper polarity (I have 10 coming, I just don't know that I need that many for the outputs I'll be using).
I'm using a Dell 19V laptop PSU for now because it's on hand. Once I am comfortable with how it's all operating, I plan to change that out for something like a linear PSU. The Dell PSU I have is recommended by HDPlex for powering their LPSUs.
I certainly plan to get everything installed and measured on a bench before putting it into service, but am I oversimplifying this a ton?
Am I wrong to think that if I just need to use:
1 x 3.3V (FIFOPi clean side)
1 x 5V (FIFOPi dirty side to power Pi)
possibly 1 x 5V to power HDMIPi
That all I really need to do with the board I have is solder in the knob button per the instructions, install 6 LifePO4 batteries in proper polarity (I have 10 coming, I just don't know that I need that many for the outputs I'll be using).
I'm using a Dell 19V laptop PSU for now because it's on hand. Once I am comfortable with how it's all operating, I plan to change that out for something like a linear PSU. The Dell PSU I have is recommended by HDPlex for powering their LPSUs.
I certainly plan to get everything installed and measured on a bench before putting it into service, but am I oversimplifying this a ton?
Did you purchase both the UcConditioner 5V & UcConditioner 3.3V?
IMO, you would want the cleanest and most robust solution directed at the clocks.
The RPi is considered "dirty", so you don't need to be as strict with power as it is already somewhat compromised. While with the clocks, you can give it complete uncompromised clean power.
For the most optimal solution, you may want to consider a StationPi which further separates the RPi power from the FiFoPi Q3 while also providing some RFI/EMI shielding.
But please be careful not to mix the UcConditioner if you only have one. The 5V and 3.3V serve different purposes.
If you only have UcConditioner 5V, use FiFoPi J3.
If you only have UcConditioner 3.3V, use FiFoPi J5.
If you have both UcCondtioner 3.3V + UcConditioner 5V, use FiFoPi J3 with UcConditioner 5V and FiFoPi J5 with UcConditioner 3.3V.
Try avoiding powering the RPi directly via microUSB or USB-C. Power via the FiFoPi GPIO if possible.
No worries, we all 'trial and error' here and sometimes overlook our requirements. I know I have.
Depending of your current system, this is a rough roadmap of possible improvements:
BYOPS = Bring your own Power Supply
So if you are currently at BYOPS + UC5V, the next optimal steps IMO are:
- BYOPS (UC5V) + UC3.3V for Q3 J5
- LinearPi 5V (UC5V)
- LinearPi Dual (UC5V) + UC3.3V
- StationPi + BYOPS (UC3.3V + UC5V) + BYOPS UC3.3V
- StationPi + BYOPS (UC3.3V + UC5V) + LinearPi 3.3V UC3.3V for J5
- LinearPi Dual (UC5V + UC3.3V) + UC3.3V + LinearPi 3.3V
- LiFePO4 MKIII (UC5V + UC3.3V)
- LiFePO4 MKIII (UC5V + UC3.3V) + StationPi + UC3.3V
Here's Ian's LinearPi video:
Asynchronous I2S FIFO project, an ultimate weapon to fight the jitter
Here's the optimal StationPi config:
Asynchronous I2S FIFO project, an ultimate weapon to fight the jitter
Asynchronous I2S FIFO project, an ultimate weapon to fight the jitter
Here's a discussion if you can use 5V on Q3 J5:
Asynchronous I2S FIFO project, an ultimate weapon to fight the jitter
I don't think you can without compromises and don't recommend on Q3, but it's good reading.
The StationPi launch:
Asynchronous I2S FIFO project, an ultimate weapon to fight the jitter
Last edited:
I think i have the Q2 version. I bought from Ian when TransportPi was released. It wil accept 5v, so Q2. DIY PS + Accusilicon.
I did try the UC5v on the J3 (Pi) vs J5 (clean) Strangly i do preffer the UC5v on the J3. When connected on J5 something is off. Sounds gets al little thin/technical, the pace of the bass is slower vs the rest. The soul/live of the music is gone. On J3 it is back + more space and details. Gonne listen some more today.
I did try the UC5v on the J3 (Pi) vs J5 (clean) Strangly i do preffer the UC5v on the J3. When connected on J5 something is off. Sounds gets al little thin/technical, the pace of the bass is slower vs the rest. The soul/live of the music is gone. On J3 it is back + more space and details. Gonne listen some more today.