@jimk04,
My understanding is that most RPi distros and software are designed so that you can just pull the plug and no SD card corruption happens. Volumio and Moode I believe are exceptions to that (I don't use them, so can't be sure).
I totally understand on the LED's. I use a permanent black marker to color over all of the LEDs in my setups. One of my setups is in my bedroom, it is too bright without that!
@sq225917, I suspect one could write a script that would trigger a shutdown via a signal on a GPIO pin. I also suspect there is one (or many) out there... if you get adventuresome and search, let us know. I use a distro/player that doesn't write to the SD card (PiCorePlayer) and so is immune to hard-power-down corruption, so I haven't worried about this issue.
Later y'all!
Greg in Mississippi
My understanding is that most RPi distros and software are designed so that you can just pull the plug and no SD card corruption happens. Volumio and Moode I believe are exceptions to that (I don't use them, so can't be sure).
I totally understand on the LED's. I use a permanent black marker to color over all of the LEDs in my setups. One of my setups is in my bedroom, it is too bright without that!
@sq225917, I suspect one could write a script that would trigger a shutdown via a signal on a GPIO pin. I also suspect there is one (or many) out there... if you get adventuresome and search, let us know. I use a distro/player that doesn't write to the SD card (PiCorePlayer) and so is immune to hard-power-down corruption, so I haven't worried about this issue.
Later y'all!
Greg in Mississippi
Search around, there is a way to disable every single LED on the RPI and with a register change you can disable the lock LEDs on the DAC board, then you can cover the rest of the LEDs with some electrical tape (I've gotten best results with black tape).
FWIW, I think most Pi distros really want a graceful shutdown where you issue a software shutdown or poweroff command before you pull the plug.
But I've pull the plug on Pi's many times without issuing the sw command, usually because the pi crashed and I can't log in to issue the command, and so far it hasn't caused a problem.
There is a small chance you can corrupt the OS/SD card if you power down the pi without performing the software shutdown.
Picoreplayer is the exception I know of, it doesn't write to the sd card at all unless you force it to, so you can't corrupt it by shutting off power.
But I've pull the plug on Pi's many times without issuing the sw command, usually because the pi crashed and I can't log in to issue the command, and so far it hasn't caused a problem.
There is a small chance you can corrupt the OS/SD card if you power down the pi without performing the software shutdown.
Picoreplayer is the exception I know of, it doesn't write to the sd card at all unless you force it to, so you can't corrupt it by shutting off power.
With regards to piCorePlayer, if using piCorePlayer as just a "player" then you can just pull the plug.
Sometimes the dirty bit gets set on partition 1 but that doesn't really affect anything. We have a script to fix the dirty bit. This script is years old and we haven't really had the need to include it but I'll add to the beta area just for those the like things tidy.
If you are using LMS on top of piCorePlayer, we "recommend" shutting down properly. With database applications it's better to be safe. Having said that, I have never had pCP/LMS go bad after a power failure. I have recently added this message on the LMS page for the next release (if it passes scrutiny)
The reason piCorePlayer is more resilient than other Linux distributions is due to the fact that it re-installs Linux on each boot. Effectively a fresh system each boot from readonly extensions.
Sometimes the dirty bit gets set on partition 1 but that doesn't really affect anything. We have a script to fix the dirty bit. This script is years old and we haven't really had the need to include it but I'll add to the beta area just for those the like things tidy.
If you are using LMS on top of piCorePlayer, we "recommend" shutting down properly. With database applications it's better to be safe. Having said that, I have never had pCP/LMS go bad after a power failure. I have recently added this message on the LMS page for the next release (if it passes scrutiny)
The reason piCorePlayer is more resilient than other Linux distributions is due to the fact that it re-installs Linux on each boot. Effectively a fresh system each boot from readonly extensions.
D
Deleted member 537459
With moode you use roon in airplay mode and is limite to 24/192. No dsd is supported
Thanks Do,
That's great. Does RoPieee support native DSD over GPIO or still DoP? Which driver did you pick up for me ESS DAC HAT? What's the maximal Fs it can go? Did you use a TFT screen as GUI? If yes, which one did you use?
Do you have pictures of your system?
Regards,
Ian
Hi Ian,
I’m off travelling at the moment but will be able to take picture later on and give you the infos as I don’t remember which DAC I had selected. I know for my USB DAC it does native DSD but can’t remember which DAC I had selected for your dual ES9038K2M. RoPieee is only for the Pi, unlike DietPi which supports most SBC.
I’m not using external LCD at the moment, just the iOS remote.
Do
I’m off travelling at the moment but will be able to take picture later on and give you the infos as I don’t remember which DAC I had selected. I know for my USB DAC it does native DSD but can’t remember which DAC I had selected for your dual ES9038K2M. RoPieee is only for the Pi, unlike DietPi which supports most SBC.
I’m not using external LCD at the moment, just the iOS remote.
Do
Last edited:
Thanks guys for your sharing.
Just flashed a RoPieee image into a RPi3. My Roon core is running on a QNAP.
I use my phone as remote controller. With my ESS controller in analyzer mode, things look good. Both PCM and DSD (DoP) format and sample rate are monitored at real time exactly as same as what displayed on the Roon controller.
I'll do listening test on weekend for the sound quality.
Have a good night.
Ian
Clock impressions
I have posted on the Well Tempered Master Clock thread my full review of FIFOPi equiped with the WTMC. If interested, have a look at that thread.
The Well Tempered Master Clock - Building a low phase noise/jitter crystal oscillator
I recall there were a few posts suggesting buying a high end clock was nuts and there were better ways to extract the best from FIFO. I can only say that FIFO is a great platform, and its true potential is only governed by the quality of the clock signal it receives.
I have posted on the Well Tempered Master Clock thread my full review of FIFOPi equiped with the WTMC. If interested, have a look at that thread.
The Well Tempered Master Clock - Building a low phase noise/jitter crystal oscillator
I recall there were a few posts suggesting buying a high end clock was nuts and there were better ways to extract the best from FIFO. I can only say that FIFO is a great platform, and its true potential is only governed by the quality of the clock signal it receives.
Hi Wlowes,
Looks very nice. You did good job. Thank you so much for sharing.
I'll take time to go through all the post. I'll try it by myself later.
Regards,
Ian
@Wlowes,
GREAT work, GREAT writeup! Thanks for envisioning, developing, and building what is an example of DIY overkill done right. AND then sharing what you did, why, and the sonic impacts of all of that for us to follow.
Since I missed getting any Pulsar clocks, this looks like the next best thing available. SO I'm eagerly waiting for the next GB.
A couple of related questions for you...
First, I know you are using a series pair of 1.5F/5.5V Ultracaps in your FiFoPi. Are you still using that? AND do you know what the charging current and time to charge with your supplies? I suspect it puts the LDOs into current limit for a short while, but curious if you measured it. I'd like to use similar parts in my FiFoPi and Ian GB DAC boards, but I power them with Ian's LiFePO4 board and am concerned about the instantaneous current draw at power-on. Likely I'll have to do some tests with various sizes of limiting resistors to get a feel for that, but hoped you might have some data that will help.
Second, I know you've been making incremental improvements in your setup over time in a measured and deliberate fashion (good for you!). One thing you MIGHT want to look at is the RPi itself. I don't know if you've been following this, but Allo is coming out with what they are calling the USBBridge Signature, an RPi-compatible (mostly) board using the same architecture as an RPi, but with a low-noise linear regulated carrier board containing the IO devices and using an RPi compute module for the processing.
I expect this will be a game-changer for RPi audio just as Ian's FIFOs have been for I2S audio. I have a couple of RPi 2B I've modified to replace the onboard noisy DC-DC converters with linear regulators. This is like a baby step in the direction of Allo's USBBridge Signature. Those are my best RPi sources and begin to get into the same sound quality realm as the very tweaky SDTrans384 SD Card Player (similar level as a stock one with good, but not the best power).
You along with a few others (like @Jacklee and @Simon Dart, but not just them) currently have some of the most advanced implementations of Ian's RPi gear around. Either a USBBridge Sig or modifying an RPi 2B for linear regulation such as I did could take your setup up another level. If you are interested in going the modified RPi route, let me know and I can provide info and resources.
OF course, while the USBBridge Sig has RPi-compatible mounting post holes and GPIO header, the base board is a good bit larger for all of the separate regulation circuits and filtering Allo added. SO fitting it may be an issue.
My 2 cents.
Greg in Mississippi
P.S. Because of changes to the power architecture of later RPis, the 2B is the last of the RPis that will work with the linear-regulation mod. ALSO, Allo's USBBridge signature will not have on-board WiFi to keep RF noise down. IF you use WiFi (I don't recommend it that close to audio processing, but many use on-board WiFi anyway), then you have to use a seperate USB Wifi dongle or a WiFi access point connected via Ethernet.
GREAT work, GREAT writeup! Thanks for envisioning, developing, and building what is an example of DIY overkill done right. AND then sharing what you did, why, and the sonic impacts of all of that for us to follow.
Since I missed getting any Pulsar clocks, this looks like the next best thing available. SO I'm eagerly waiting for the next GB.
A couple of related questions for you...
First, I know you are using a series pair of 1.5F/5.5V Ultracaps in your FiFoPi. Are you still using that? AND do you know what the charging current and time to charge with your supplies? I suspect it puts the LDOs into current limit for a short while, but curious if you measured it. I'd like to use similar parts in my FiFoPi and Ian GB DAC boards, but I power them with Ian's LiFePO4 board and am concerned about the instantaneous current draw at power-on. Likely I'll have to do some tests with various sizes of limiting resistors to get a feel for that, but hoped you might have some data that will help.
Second, I know you've been making incremental improvements in your setup over time in a measured and deliberate fashion (good for you!). One thing you MIGHT want to look at is the RPi itself. I don't know if you've been following this, but Allo is coming out with what they are calling the USBBridge Signature, an RPi-compatible (mostly) board using the same architecture as an RPi, but with a low-noise linear regulated carrier board containing the IO devices and using an RPi compute module for the processing.
I expect this will be a game-changer for RPi audio just as Ian's FIFOs have been for I2S audio. I have a couple of RPi 2B I've modified to replace the onboard noisy DC-DC converters with linear regulators. This is like a baby step in the direction of Allo's USBBridge Signature. Those are my best RPi sources and begin to get into the same sound quality realm as the very tweaky SDTrans384 SD Card Player (similar level as a stock one with good, but not the best power).
You along with a few others (like @Jacklee and @Simon Dart, but not just them) currently have some of the most advanced implementations of Ian's RPi gear around. Either a USBBridge Sig or modifying an RPi 2B for linear regulation such as I did could take your setup up another level. If you are interested in going the modified RPi route, let me know and I can provide info and resources.
OF course, while the USBBridge Sig has RPi-compatible mounting post holes and GPIO header, the base board is a good bit larger for all of the separate regulation circuits and filtering Allo added. SO fitting it may be an issue.
My 2 cents.
Greg in Mississippi
P.S. Because of changes to the power architecture of later RPis, the 2B is the last of the RPis that will work with the linear-regulation mod. ALSO, Allo's USBBridge signature will not have on-board WiFi to keep RF noise down. IF you use WiFi (I don't recommend it that close to audio processing, but many use on-board WiFi anyway), then you have to use a seperate USB Wifi dongle or a WiFi access point connected via Ethernet.
Last edited:
I've been needing to post an update on my Ian GB setups.
First, I don't have any exotic clocks yet, so am still using either Crystek CCHD-957 or NDK SDA... both have some good points, I'm using the NDKs more often than the Crysteks.
I had posted how I'd been modifying the IVStd I/V stage filtering along with adding a Rasmussen-type pre-I/V filter, based on my personal preferences. My latest setup uses 1500pF SMD PPS film caps in the I/V stage feedback filter (stock is 270pF) and a 1uF/2R Rasmussen filter built into the DAC board connector. I've been going back and forth between LKS, Sparkos, and Burson discrete opamps. They all seem to be working a bit better with less high-frequency noise getting into them and the gap between them is closing. I'm currently using the latest Burson and think I might like them a bit better than the Sparkos and definitely better than the LKS to my ears.
I'm finding I like both the transformer output and this iteration of the IVStd about equally. The transformers using either the Lundahl LL1545A or now unavailable Onetics are very relaxed with good definition and good dynamics. BUT the discrete and further filtered IVStd setups while still lagging a bit in relaxation, lead in definition, dynamics, and frequency extension. I'm really looking foward to getting Bisesik's output transformers in, my pair is on the way.
Power is still a single LiFePO4 supply, so I haven't added a 2nd to provide split power to the three rails on the DAC board and separately power the clocks. AND I haven't recently experimented with Supercaps / Ultracaps on either the FiFoPi or DAC boards. SO I have room for improvement in the power realm AND in the clock realm. ONE of my current projects is to build up some additional UCPi point-to-point (as I don't know that Ian has anymore boards) so I can use up to 8 on Ian's GB stack (3 for the FiFoPi, 3 for the DAC board, and 2 on his prototype OP861 IV stage, which while it shows promise, lags behind the two output setups above in detailing and definition. Ian suggested using UCPi instead of the onboard TL431-based regulation, so that's why I need more of them!)
Still this setup is sounding VERY good and edging ahead of my modified Allo Katana. My favorite power scheme for the Katana uses 2 of Ian's UCPi at 5V for the DAC and Microprocessor boards and using a good +-supply with Sparkos +-15V regulators for the Opamp board. Replacing the UCPi (which are being float charged by a pair of Uptone Audio LPS-1.2's) with an Allo Shanti is close, I expect largely because they also use Ultracaps in that supply.
I know I don't reach the rarified performance levels Ian does using his Pulsar clocks or WLowes with his really tweaked Well-Tempered Clock, but I have to say I'm really happy with what I'm getting currently.
Greg in Mississippi
First, I don't have any exotic clocks yet, so am still using either Crystek CCHD-957 or NDK SDA... both have some good points, I'm using the NDKs more often than the Crysteks.
I had posted how I'd been modifying the IVStd I/V stage filtering along with adding a Rasmussen-type pre-I/V filter, based on my personal preferences. My latest setup uses 1500pF SMD PPS film caps in the I/V stage feedback filter (stock is 270pF) and a 1uF/2R Rasmussen filter built into the DAC board connector. I've been going back and forth between LKS, Sparkos, and Burson discrete opamps. They all seem to be working a bit better with less high-frequency noise getting into them and the gap between them is closing. I'm currently using the latest Burson and think I might like them a bit better than the Sparkos and definitely better than the LKS to my ears.
I'm finding I like both the transformer output and this iteration of the IVStd about equally. The transformers using either the Lundahl LL1545A or now unavailable Onetics are very relaxed with good definition and good dynamics. BUT the discrete and further filtered IVStd setups while still lagging a bit in relaxation, lead in definition, dynamics, and frequency extension. I'm really looking foward to getting Bisesik's output transformers in, my pair is on the way.
Power is still a single LiFePO4 supply, so I haven't added a 2nd to provide split power to the three rails on the DAC board and separately power the clocks. AND I haven't recently experimented with Supercaps / Ultracaps on either the FiFoPi or DAC boards. SO I have room for improvement in the power realm AND in the clock realm. ONE of my current projects is to build up some additional UCPi point-to-point (as I don't know that Ian has anymore boards) so I can use up to 8 on Ian's GB stack (3 for the FiFoPi, 3 for the DAC board, and 2 on his prototype OP861 IV stage, which while it shows promise, lags behind the two output setups above in detailing and definition. Ian suggested using UCPi instead of the onboard TL431-based regulation, so that's why I need more of them!)
Still this setup is sounding VERY good and edging ahead of my modified Allo Katana. My favorite power scheme for the Katana uses 2 of Ian's UCPi at 5V for the DAC and Microprocessor boards and using a good +-supply with Sparkos +-15V regulators for the Opamp board. Replacing the UCPi (which are being float charged by a pair of Uptone Audio LPS-1.2's) with an Allo Shanti is close, I expect largely because they also use Ultracaps in that supply.
I know I don't reach the rarified performance levels Ian does using his Pulsar clocks or WLowes with his really tweaked Well-Tempered Clock, but I have to say I'm really happy with what I'm getting currently.
Greg in Mississippi
Last edited: