I use the nylon ones. I have nothing earthed!
On another subject.....I just added 4 off Panasonic 0.33F button supercaps on the underside of the dac....and the promptly removed 2 of them as it wont fit back on the isolator. And I put tape on one remaining one as it was touching a header on the isolator.
What do others do with this situation?
On another subject.....I just added 4 off Panasonic 0.33F button supercaps on the underside of the dac....and the promptly removed 2 of them as it wont fit back on the isolator. And I put tape on one remaining one as it was touching a header on the isolator.
What do others do with this situation?
Or better still, use 3 seperate rails and caps this way.
An externally hosted image should be here but it was not working when we last tested it.
Finally have my setup making music!
So started with an Allo Sparky, but it kept stuttering, little dropouts every few seconds. I managed to get it working once at 44.1k, but then stuttering came back so I gave up on it. Maybe it has some compatibility problem with Ian's boards?
Using a Pi3 instead, running dietpi with squeezesoft and HQP/NAA and airplay.
Output state is Bisesik's iv transformer.
dual mono 9038 dac hat and fifo pi boards, with one of Ian's.
PS is two LT3042 boards generating 3.3V, each connected to two Nesscap 2.7V, 360F super ultra caps. Passively balanced with 90 ohm resistors. I know this is lower than most use, but only downside is a little more power dissipation, lower R will balance better.
So Basically two ultracaps PS's, one feeds DAC and other feeds fifo pi. Thinking about adding two more ultracap PS's for the DAC, haven't decided yet.
I have a linear walwart, 5V @ 1A to power to Pi3. 1A is working fine, I have wifi turned off, and nothing connected to USB.
Letting it break in, but I was impressed listening last night.
Ordered an adapter board and an NDK NZ2520SDA 45.1584Mhz. I'm only concerned about DSD/DOP and 44.1k, so only changing the one clock. May also see if I can get an Andrea clock setup going.
When I change the clock, I'll also bypass the fifo pi 3.3V reg, wondering if anyone has pics of how to do it?
Thanks
Randy
So started with an Allo Sparky, but it kept stuttering, little dropouts every few seconds. I managed to get it working once at 44.1k, but then stuttering came back so I gave up on it. Maybe it has some compatibility problem with Ian's boards?
Using a Pi3 instead, running dietpi with squeezesoft and HQP/NAA and airplay.
Output state is Bisesik's iv transformer.
dual mono 9038 dac hat and fifo pi boards, with one of Ian's.
PS is two LT3042 boards generating 3.3V, each connected to two Nesscap 2.7V, 360F super ultra caps. Passively balanced with 90 ohm resistors. I know this is lower than most use, but only downside is a little more power dissipation, lower R will balance better.
So Basically two ultracaps PS's, one feeds DAC and other feeds fifo pi. Thinking about adding two more ultracap PS's for the DAC, haven't decided yet.
I have a linear walwart, 5V @ 1A to power to Pi3. 1A is working fine, I have wifi turned off, and nothing connected to USB.
Letting it break in, but I was impressed listening last night.
Ordered an adapter board and an NDK NZ2520SDA 45.1584Mhz. I'm only concerned about DSD/DOP and 44.1k, so only changing the one clock. May also see if I can get an Andrea clock setup going.
When I change the clock, I'll also bypass the fifo pi 3.3V reg, wondering if anyone has pics of how to do it?
Thanks
Randy
@randytsuch,
Sorry to hear about the Sparky. Personally I was never tempted by that option.. compatibility with the main RPi distros / players is important to me.
I will say Allo's new RPi USBBridge Sig works very well under Ian's GB stack (and my Katana stack) and raised the SQ well above either a stock RPi or above one modified with linear regulation, both with UC power... see my comments in Allo's Shanti / USBBridge Sig / Revolution thread in posts #588 & #594 here:
Shanti Dual LPS 5V/3A , 5V/1.5A
Since it is at its heart an RPi, it only needs an additional driver to work with most current distro / player combos. Allo already has piCorePlayer, Moode, DietPi, RoPiee, Raspian, Volumio, and Max2Play distros with the new driver from the distro owners. Otherwise it works just like any other RPi.
Even if you don't go to a USBBridge Sig, do add another UC pair at 5V for the RPi supply... a very significant upgrade with all three of my RPi alternatives over a more stock supply.
I'm breaking in a recently populated Ian IVStd board with some filtering tweaks and Sparkos Labs discrete opamps to compare against my well-broken-in Bisesik transformer output board. At best I suspect the IVStd with Sparkos might equal the transformers, but I expect the transformers to excel. They are very good!
BTW, I compared the Sparkos against OPA1612 and LKS & Burson discretes. All were lifted with the filtering tweaks, but the Sparkos stayed on top for me. Other ears and preferences might chose differently.
@Misterdog's setup with a single supply and 3 local Ultracaps may be nearly as good as 3 separate UC supplies when supplied via your 360F UC pair. Perhaps a next Q&D step?
Pix anywhere?
Greg in Mississippi
P.S. How does it compare to your other setup(s)?
Sorry to hear about the Sparky. Personally I was never tempted by that option.. compatibility with the main RPi distros / players is important to me.
I will say Allo's new RPi USBBridge Sig works very well under Ian's GB stack (and my Katana stack) and raised the SQ well above either a stock RPi or above one modified with linear regulation, both with UC power... see my comments in Allo's Shanti / USBBridge Sig / Revolution thread in posts #588 & #594 here:
Shanti Dual LPS 5V/3A , 5V/1.5A
Since it is at its heart an RPi, it only needs an additional driver to work with most current distro / player combos. Allo already has piCorePlayer, Moode, DietPi, RoPiee, Raspian, Volumio, and Max2Play distros with the new driver from the distro owners. Otherwise it works just like any other RPi.
Even if you don't go to a USBBridge Sig, do add another UC pair at 5V for the RPi supply... a very significant upgrade with all three of my RPi alternatives over a more stock supply.
I'm breaking in a recently populated Ian IVStd board with some filtering tweaks and Sparkos Labs discrete opamps to compare against my well-broken-in Bisesik transformer output board. At best I suspect the IVStd with Sparkos might equal the transformers, but I expect the transformers to excel. They are very good!
BTW, I compared the Sparkos against OPA1612 and LKS & Burson discretes. All were lifted with the filtering tweaks, but the Sparkos stayed on top for me. Other ears and preferences might chose differently.
@Misterdog's setup with a single supply and 3 local Ultracaps may be nearly as good as 3 separate UC supplies when supplied via your 360F UC pair. Perhaps a next Q&D step?
Pix anywhere?
Greg in Mississippi
P.S. How does it compare to your other setup(s)?
Sparky was cheap, I picked it up used here so no big deal that it didn't work out.
I'll save up my pennies to get the USBBridge, sounds like a big upgrade. I was thinking about adding a couple of ultracaps for the pi PS.
The thing I worry about for ultracaps is charging them up, and how hot the regs get. I stuck little heatsinks I had on the LT3042's, and had a fan blowing on them when I was charging those ultracaps.
Using a transformer makes the PS much simpler, you just need 3.3V and 5V. Plus I had heard good things about his transformers. Still early, but I like what I'm hearing so far.
I'll take some pics tonight if I remember.
Also still hoping someone posts instructions to bypass the fifopi 3.3V regs. Otherwise I'll have to break down and try to figure it out. Are there fifo pi schematics anywhere?
EDIT: NVM, guess I should read the manual more carefully. Need to short out R58 and R59.
One more question though, anyone try disabling DoP with S1 switch, and see if DoP sounds better or worse?
Randy
I'll save up my pennies to get the USBBridge, sounds like a big upgrade. I was thinking about adding a couple of ultracaps for the pi PS.
The thing I worry about for ultracaps is charging them up, and how hot the regs get. I stuck little heatsinks I had on the LT3042's, and had a fan blowing on them when I was charging those ultracaps.
Using a transformer makes the PS much simpler, you just need 3.3V and 5V. Plus I had heard good things about his transformers. Still early, but I like what I'm hearing so far.
I'll take some pics tonight if I remember.
Also still hoping someone posts instructions to bypass the fifopi 3.3V regs. Otherwise I'll have to break down and try to figure it out. Are there fifo pi schematics anywhere?
EDIT: NVM, guess I should read the manual more carefully. Need to short out R58 and R59.
One more question though, anyone try disabling DoP with S1 switch, and see if DoP sounds better or worse?
Randy
Last edited:
@randytsuch,
I expect you'll hear a good bump in SQ when you bypass the FiFoPi LDOs. The FiFoPi sounded good with UCs feeding it at 5V through the LDOs. BUT even better with the UCs at 3.3V and the LDOs bypassed. One of the worse setups I tried was anything feeding the FiFoPi 3.3V with the LDOs not bypassed. That just didn't sound very good no matter what 3.3V source.
Greg in Mississippi
I expect you'll hear a good bump in SQ when you bypass the FiFoPi LDOs. The FiFoPi sounded good with UCs feeding it at 5V through the LDOs. BUT even better with the UCs at 3.3V and the LDOs bypassed. One of the worse setups I tried was anything feeding the FiFoPi 3.3V with the LDOs not bypassed. That just didn't sound very good no matter what 3.3V source.
Greg in Mississippi
Received a PM with questions about my setup, but I'll answer here, seems like others may want this info.
My Power supply is a 5V, 1A linear regulated walwart from Jameco. It feeds two LT3042 regulator boards. I'm using the free boards that Ian gave out if you asked when placing the DAC fifopi order.
The LT3042 has a built in current limit, I think it defaults to 250 ma if you don't install a resistor, that's how mine are setup. Ian's pl sets it to 200 ma.
The difference is how much heat they generate. To combat the heat, I had a little stick on heatsink which I cut in half with a hacksaw, and stuck 1/2 of each 3042. Be careful to get the heatsink as square as possible to the 3042. Without the heatsink, they get hot fast when charging the caps.
Each 3042 output goes to two ultracaps, wired in series. A 90 ohm resistor is installed from neg to pos terminal on each ultracap, so need a resistor for each cap.
Ultracap "output" (3.3V terminal) to a switch, bottom term. I used a double pole double throw. Switch mid term to boards, one to DAC and one to fifopi.
To be cont, with pics
Randy
My Power supply is a 5V, 1A linear regulated walwart from Jameco. It feeds two LT3042 regulator boards. I'm using the free boards that Ian gave out if you asked when placing the DAC fifopi order.
The LT3042 has a built in current limit, I think it defaults to 250 ma if you don't install a resistor, that's how mine are setup. Ian's pl sets it to 200 ma.
The difference is how much heat they generate. To combat the heat, I had a little stick on heatsink which I cut in half with a hacksaw, and stuck 1/2 of each 3042. Be careful to get the heatsink as square as possible to the 3042. Without the heatsink, they get hot fast when charging the caps.
Each 3042 output goes to two ultracaps, wired in series. A 90 ohm resistor is installed from neg to pos terminal on each ultracap, so need a resistor for each cap.
Ultracap "output" (3.3V terminal) to a switch, bottom term. I used a double pole double throw. Switch mid term to boards, one to DAC and one to fifopi.
To be cont, with pics
Randy
So I connected it all up, with caps not charged (0 VDC). Per Greg's recommendation, used lots of kapton tape at the ends to keep the caps from shorting to anything. Based on lifepo4 batteries I've used, I'm sure a charged ultracap will dump a done of current if shorted to something. But I don't think the mounting posts connect to anything, so I didn't worry about covering them.
I wired the caps to the DAC and fifopi like the ultracaps were the PS, since they are the current source. So I made the ground wires from caps to boards as short as practical, same with wires from caps to switch to board. Switches add wire length, but I didn't want to wire directly from caps to boards.
I solder power to the DAC and fifopi for all three power sources (including 5V for the pi3). I don't like using the screw down type connectors for clean power, trust solder a more.
So with switch off, I plugged in the 5v walwart to start charging the caps. With no heatsink, it gets hot fast. Heatsink slowed down temp rise quite a bit, but still gets pretty toasty and caps had a long way to go.
So I pulled out a PC fan from when I was trying to make a really quiet PC a long time ago, and set it up to blow air over the heatsinks. Fan wasn't blowing very hard, but this helped a lot, with a fan, temps never got over 100F. Without fan, probably hit 150F after maybe 5 mins and I pulled power.
I accidentally installed a 5V reg in one position, so it started to exceed 3.3V. At 3.4V I knew something was wrong and pulled the plug. Replaced regs with caps charged, and tried again. It worked for a while, but then it went wonky on my and started to discharge, even with power applied. Something may have latched up. I thought I blew the reg, but left it powered on. After it reached 0, it started to work and charged the caps back up. I came back and was happy to find both rails at 3.3V.
At this time, with caps charged and dac running overnight, reg heatsinks were at 90F, so regs probably a little hotter but under 100F.
BTW, I use a little IR thermometer to measure temps. Just point and pull the trigger, they work great if you're working on power supplies or regs and need to know chip temps.
BTW, to power up, I plug in the walwart that powers the Pi, then I turn on my dac/fifopi switch. Not sure if it matters, but I thought Pi should be turned on first. Pi power also powers up the display.
EDIT:
Wanted to add that I bypassed the 5v fifopi reg last night. I had to pull off my output/tx board, then the dac board, then the fifopi board. Flip over the fifopi, and found R58 and R59. There are no parts installed, just the pads, so they kind of stand out. Soldered very short wire across the pads and reassembled.
Nice improvement, it does sound much better with this simple change.
I also removed one oscillator, as I'm only using DSD or 44.1k so I didn't need the 48k clocks.
Now really looking forward to getting a better 45.xxxx MHz clock, but its coming from China so will need to wait a bit. Hate it when I could have ordered something a while ago but didn't think about it lol
One last thing. I'm no where near as tweaky as Greg. So I won't spend a lot of time (or maybe none) trying different things, and comparing. I'm sure if I tried harder, I could get maybe 5% improvement (made up number, please don't argue about it), but at some point I just want to listen to music. So I do what I think should sound good, much of it has been based on this thread or previous projects or other projects I know of, and go with it.
Randy
I wired the caps to the DAC and fifopi like the ultracaps were the PS, since they are the current source. So I made the ground wires from caps to boards as short as practical, same with wires from caps to switch to board. Switches add wire length, but I didn't want to wire directly from caps to boards.
I solder power to the DAC and fifopi for all three power sources (including 5V for the pi3). I don't like using the screw down type connectors for clean power, trust solder a more.
So with switch off, I plugged in the 5v walwart to start charging the caps. With no heatsink, it gets hot fast. Heatsink slowed down temp rise quite a bit, but still gets pretty toasty and caps had a long way to go.
So I pulled out a PC fan from when I was trying to make a really quiet PC a long time ago, and set it up to blow air over the heatsinks. Fan wasn't blowing very hard, but this helped a lot, with a fan, temps never got over 100F. Without fan, probably hit 150F after maybe 5 mins and I pulled power.
I accidentally installed a 5V reg in one position, so it started to exceed 3.3V. At 3.4V I knew something was wrong and pulled the plug. Replaced regs with caps charged, and tried again. It worked for a while, but then it went wonky on my and started to discharge, even with power applied. Something may have latched up. I thought I blew the reg, but left it powered on. After it reached 0, it started to work and charged the caps back up. I came back and was happy to find both rails at 3.3V.
At this time, with caps charged and dac running overnight, reg heatsinks were at 90F, so regs probably a little hotter but under 100F.
BTW, I use a little IR thermometer to measure temps. Just point and pull the trigger, they work great if you're working on power supplies or regs and need to know chip temps.
BTW, to power up, I plug in the walwart that powers the Pi, then I turn on my dac/fifopi switch. Not sure if it matters, but I thought Pi should be turned on first. Pi power also powers up the display.
EDIT:
Wanted to add that I bypassed the 5v fifopi reg last night. I had to pull off my output/tx board, then the dac board, then the fifopi board. Flip over the fifopi, and found R58 and R59. There are no parts installed, just the pads, so they kind of stand out. Soldered very short wire across the pads and reassembled.
Nice improvement, it does sound much better with this simple change.
I also removed one oscillator, as I'm only using DSD or 44.1k so I didn't need the 48k clocks.
Now really looking forward to getting a better 45.xxxx MHz clock, but its coming from China so will need to wait a bit. Hate it when I could have ordered something a while ago but didn't think about it lol
One last thing. I'm no where near as tweaky as Greg. So I won't spend a lot of time (or maybe none) trying different things, and comparing. I'm sure if I tried harder, I could get maybe 5% improvement (made up number, please don't argue about it), but at some point I just want to listen to music. So I do what I think should sound good, much of it has been based on this thread or previous projects or other projects I know of, and go with it.
Randy
Last edited:
Another idea I had was to put multiple ultracap pairs, connected to one 3.3V reg providing power to them. It would take longer to charge up the caps to 3.3V, but then each pair becomes its own power supply.
My premise is that the cap pairs have so much capacitance/storage that they will act as a buffer so the power supply used to charge them is less important. I think this is true for lifepo4 batteries, at least at the limits of my system and hearing.
I'm wondering also if the regs I'm using are overkill for this application, so I just bought some simple 3.3V to220 regs to try.
Randy
Hi to all. I am using two raspberry to play music. One as a sever (Pi4 with jriver) and another as a player (pi3b + fifopi + ess dual mono + opamp 1612 i/V) connected together via network {dedicated switch with cat8 Ethernet cables}
May I use hdmi connection {transmitter and receiver) instead of network connection entering into fifopi with i2s signal? Is it better or it is the same? And why?
May I use hdmi connection {transmitter and receiver) instead of network connection entering into fifopi with i2s signal? Is it better or it is the same? And why?
Inspiring write-up, Randy... a welcomed addition to Greg's excellent reports.