Hifiberry DAC+ Pro - HW mods anybody?

[soundcheck]

Hi.

Refering to "Ultrafi" (Interesting alias btw - hope you enjoy the "hifiberry" )

My two cents:

1. The option of using SMPSs have been discussed and brought up more than once.
Welcome to the club.
I still prefer the iFi iPower , IMO one of the best SMPSs out there.
I did shorten the cable and added buffer capacity to get more meat on the bones.
2. To not use any onboard gadgets (wifi etc) or HDDs attached to the PI is more than well
known and discussed. That'll also take the load (and noise) down btw.
3. I don't see his load variations and high loads. ( see ckramer)
4. I don't think the Hifiberry DAC+ Pro is worth it. There'll be some IMO just minor improvements to the DAC+.
E.g. the Sabre with the Kali supplied master clock plays in a different league.
even without Kali i do prefer the Sabre.
However. To put the jitter discussion into perspective.
On an 1-10 impact scale, where 10=max impact. I'd rate the DAC+ Pro impact "2-3"
Even the Kali I wouldn't rate higher the "3-4" applied to a Sabre.
5. The PCM51xx on-dac filters and volume control is all but OK to me.
6. I do not agree with his conclusion that you can fix from the outside what's messed
up on the inside.
To me the "right" choice of capacitors and parts around the DAC chip gets an "8-9" on
my impact scale. ( And I did not follow Gregs advise btw.)
7. Combing this or that LDO with this or that buffer/decoupling scenario might end up in
gambling for better sound. Drawing the right conclusions without doing measurements and simulations can easily drive you into the wrong direction.
8. The 3.3V supply situation on the HifiBerry disqualifies these DACs to me.
9. Moode is nice. ( Otherwise I wouldn't contribute to it - ) . But. Tim has to accept certain compromises to meet all the different demands.
Most people prefer features over soundquality. You can't have em both.

Ah. One more.

10. On my Sabre ( former Mamboberry DAC) I swapped LiFePO4s against a well buffered iFi iPower supply. What more do I want!?!? I'm more than happy that I got rid of batteries and linear supplies. ( Greg most probably prefers his overexpensive supercap supply though. At least, I'm using the same LDO - with my caps of choice. )

Enjoy.

 

[Bas Horneman]

Getting the wifi working with the HiFi Berry DAC + Pro is really a struggle, even with a wifi dongle; but, once the network was set to use channel 10, it works well - even with the on-board wifi. This is because the local oscillators on the HiFi Berry DAC + Pro interfere with the wifi signaling on channels 1, 6, 11 and 14, see: https://www.tablix.org/~avian/blog..._interference/ Thus, it is best to avoid those channels by avoiding automatic channel selection by the router and selecting another channel, for example, I choose channel 10.

I could not get the wifi to work on the rpi3. In the end I just built one around the rpi2 to give to my mate.

This issue you describe must have been the problem!

 

[Greg Stewart]

<SNIP>
1 - ... running Moode 3.1... Other software has been tried but my opinion is that Moode Audio sounds best. My local group of audio friends also agree<SNIP>

Interesting. I'd settled on PiCorePlayer version 1.22, which lacks a digital edge I found in more recent versions. 3.02 seemed better than any 2 version I tried, but still had some of the edge... though I LOVED how fast V3 booted in comparison to 1.22.

I've been interested in Moode after Soundcheck worked with Tim to encorporate his and Clive's patches. I'll give it a try.

<SNIP>
1 - ... with on-board wifi and configured with as static IP address, my router being set to channel 10.
<SNIP>

I'll watch out for that if I go to Pi3s. Right now I'm happy with my old 2Bs, but I guess I should get a 3 and a Zero and compare.

<SNIP>
2 - Power for the Raspberry Pi3 is feed into GPIO header pins 2 and 6, see https://pinout.xyz, bypassing the protection circuitry after the micro usb connection comprised of a current mirror, etc. shown in the upper left hand corner here: https://www.raspberrypi.org/docum.../schematics/RPI-3B-V1_2-SCHEMATIC-REDUCED.pdf Somewhat surprisingly, this is a pretty nice upgrade over powering via the micro usb connector. Others have found the same, albeit in a little different configuration.

That is very useful. I've never fed power via the micro USB, thinking it too small and flimsy. Now I never need to try.

Also thx for the schematic link. I guess I had not explored enough.

<SNIP>
3 - Power to the Raspberry Pi3 is from a 5 volt, 2.5 Amp, micro usb, switch mode wallwart, see:

QVS 2.5Amp 5.1v Switching Power

Supply for Raspberry Pi B with Built-in 4ft Micro-USB Cable ARUSB-2.5A - Micro Center

Before you laugh and dismiss what I’m about to say, just because it is a SMPS, here’s what has been tried in terms of linear supplies. Both of the later mentioned DIYINHK regulator boards, a garden variety LM317 implementation, a discrete high speed wide bandwidth linear regulator of my own design loosely based on some work I did with John Swenson, a huge HP lab supply weighing more than 35 pounds and another more generic adjustable linear supply. In every case, the little wallwart won out sonically - and not by a small margin.

Why might that be? Well, and this is true of all computers, it turns out that the Raspberry Pi3 load fluctuates wildly. For instance, on one of the aforementioned linear supplies that has a digital read out for current, the current varied any where from a couple tenths of an amp to over 2 amps. I know this isn’t exactly accurate but, it give you some idea of what is happening. Further, two of these linear supplies have the ability to set an upper current limit, It wasn’t until the current limit was set to over 2.5 amps that the supplies worked for the Raspberry Pi3.

I am using a bus power USB Hard Drive.

Still, computer loads vary wildly and are very dynamic in nature. Practically, SMPS are better suited than their linear counterparts in dealing with this.

Again, the sonic differences here were not subtle.

I remember you mentioning this. I was intrigued then and am now. As I mentioned in that conversation, the results I've gotten powering either any of my R-Pis or my SD-Trans384 SD Card Player with an Uptone Audio LPS-1 have been mixed. Using up to 3 of them to power the rails of either a HFBD+P or a Mamboberry and a Kali have been VERY impressive indeed. Using an LPS-1 to power a Pi, it was definitely cleaner than my hopped up K&K Low Voltage Power Supply (12W version with larger filter caps), but the K&K supply was more dynamic. I wondered if the 1.1A max current of the LPS-1s were just too low for the Pi. Also I saw this by John Swenson describing how he recommends setting up a good computer digital processing supply:

Sonore microRendu - Page 31

It supports your results.

I recently got 2 more LPS-1, current-matched units that will parallel well. I plan to try them for Pi power in parallel, set to 7v, with a fairly larger output
capacitor, and a good moderately-high-current 5v regulator (2A Belleson or OPC 2A dual paralleled LT3042) mounted right at the plug to the Pi.

In the meantime, I'll order one of your MicroCenter adapters AND also try one of the Meanwell 7.5v 2A SMPS that came with the LPS-1, again with a 5v regulator as above right at the Pi. The price is right on both options!

Also, might be interesting for you to provide separate power to the USB HD and re-try your power options.

<SNIP>
4 - Pins 1 and 17 in the GPIO header are bent over so that 3.3v does not pass from the Raspberry Pi3 to the HiFi Berry DAC + Pro.

And I did this by cutting traces on the HFBD+P board.

<SNIP>
I did run a PCM5102A in PLL mode in comparison to the HiFi Barry DAC + Pro, and if you know what lower jitter sounds like - this is it. The sonic advantage is clear.

I did this running an Allo Piano without a Kali. Not a good experience in comparison to the other options I have. It is MUCH better with a Kali, but still doesn't come up to the modified HFBD+P in my setup.

<SNIP>
Before we get into what I did to the HiFi Berry DAC + Pro let me say, I’m a TI/BB DAC chip fan, have been for a long time. I’ve tried many others, one of which is the ESS brand. I know there are HAT DACs using the ESS chip(s) with their own oscillators but I always hear the ASRC in those designs.

The secret to getting good sound from the PCM51XX series lies in the implementation and, more so, to the supplies. Here’s what I did.

Totally agree, besides my HFBD+P, I have TI/BB chips in my Sony HAP Z1-ES and Panasonic S-47 DVD player, both highly modified and in the upper tier of my digital setups. Main mods to each were power supply upgrades and they all responded VERY well to those. The stock units in each case had nothing special for DAC power and sounded like that... but with better power they ALL moved into my top tier.

EXCEPT I found religion running the Mamboberry in sync mode with the clock from a Kali... it was very good too... and the ASRC sound was almost entirely gone (I don't think it goes completely away on the ES902x chips until you upsample to 384, if it even works.

I've heard from others whose ears and technical abilities I respect that the higher-end ESS chips in sync-mode setup well can be VERY, VERY good and a different world from them running their ASRC. So I plan to play with them more, especially with the new and heralded ES9028 chips.

Though I have to say that my current BEST setup is my somewhat modified Soekris 1021 fed via from a Pi with a Kali powered by an LPS-1. There is a sense of space and life that I get with no other setup I own. In it's honor, I got one of the AudioPhonics TDA1387 R-Pi DAC HATS to play with AND am gathering the parts for a TDA1541 setup fed via Ian's I2S-PCM converters (and either his FIFO or a Kali).

<SNIP>
1 - I used WBT Nextgen RCA jacks. This is a no brainer in my opinion and makes a noticeable difference. I caution anyone against removing the RCA jacks on the HiFi Berry DAC + Pro. The traces from the PCM5122 output filter to the through hole solder connections provided for adding your own RCA jacks is quite thin and goes under the on-board jacks. Frankly, I think HiFi Berry could have done a bit better job here with this trace work. You’ve been warned.

Ok, Ok, OK, I'll try this!

<SNIP>
2 - I changed the caps (C7, C8) in the LPF after the PCM5122 to PPS caps, Digikey P/N PCF1457CT-ND as suggested by Greg Stewart. Generally, I have found over more than a decade of messing with DACs that more often than not, you’re better off without these LPFs. But, not in this case. Changing to this type of cap (PPS) is a definite improvement.

Thanks for the confirmation.

These caps are also part of the secret sauce on my Sony HAP Z1-ES, Panasonic S-47, and Mamboberry mods... even the Soekris, where have and will I use a crazy number of these caps along with a wildly crazy number of Black Gates.

<SNIP>
3 - I also did the other capacitor changes suggested by Greg; but, once you’ve done the supply changes below, their efficacy is reduced.

Also thanks for this confirmation.

<SNIP>
4 - For the PCM5122 analog supply, I removed the 0 ohm resistor (R14) and the on-board 5 to 3.3V regulator (U2), bridging across the input and output pin land areas,

and feed 3.3V into P3. The regulator in the supply is the TI TPS7A4700, see: 4.17uV Ultralow noise DAC power supply regulator 3.3V/5V 1Ax2 - DIYINHK I’ve found over time that I prefer TI regulators with the analog section of Ti DAC chips - go figure… If you want to leave the on-board regulator and feed 5 volts in, you can do that; but, it is a little better without U2.

Kind of a trade off between the regulator being proximate the load and having a better regulator.

For completeness, a garden variety LM317 based, a discrete current sourced, shunt regulated, and several other of the latest generation series type regulators were

compared here. The TPS7A4700 won out.

ESPECIALLY THANKS FOR YOUR TIME AND EFFORT COMPARING DIFFERENT OPTIONS. THIS IS VERY VALUABLE!!

I'll have to try TPS7A47000 regs on my TI/BB DAC chip setups. I used ADM7151-based units in the Sony HAP Z1-ES, small discret shunts that are similar to simplified Salas in the Panasonic S-47, and OPC's parallel LT3042 in my HFBD+P. All worked well, but now I wonder if I can get a bit more with the TI regs.

And I won't bother to remove the on-board reg.

<SNIP>
5 - For the PCM5122 digital section and the clocks, having bent pins 1 and 17 over in the GPIO header, I put a white wire jumper between GPIO extension header through

holes 1 and 17 and feed 3.3V into vias 17 (+3.3v) and 20 (GND). The supply is based on the LT3042 with a pass element for additional current handling and has a common-mode choke on the input, see: 0.8uV Ultralow noise DAC power supply regulator 3.3/5/7V 1.5A*x2 - DIYINHK The notion here was to provide the wildly varying peak peak current demands associated with processing while prevent any noise from making its way in or out of what was being supplied.

I did try separate supplies for the digital section of the PCM5122 and the clocks; but, frankly, there wasn’t enough difference to make it worth the trouble.

Again, for completeness, a number of supplies as mentioned herein above were compared; but, this one won out.

And again, thanks for the comparison. The OPC paralleled LT3042 only goes up to 1A or so, but has a high amount of capacitance storage both before and after the regs, which I suspect provides some, but not all of the same benefits as the pass element and choke.

And again, I'll skip the separate clock power trial based on your results.

<SNIP>
6 - For the AC feed to the two above supples, I use a 50VA 6V+6V toroid, having a 4.17 amp rating per winding. I tried lesser rated toroids and E/I based transfomers but the 50vA toroid really made the sound stage much, much better. I was surprised.

This doesn't surprise me, ever since Paul and Stan of PS Audio got very good results with their first HCPS using a wildly oversized transformer in the early 80s, I've made it a point to do the same whenever I did totally DIY'd supplies. Nowadays I consider it a best practice.

<SNIP>
Finally, all of the forgoing was mounted to a 7” by 12”, 1/8” thick 6061 aluminum plate. Front and rear panels were fabricated from 1” by 1/2” aluminum channel stock.

A lid of correspondingly sized 6061 (not shown) was also made. The difference from mounting everything to the plate may likewise come as a surprise to some; but, I’ve had this happen before. Metal standoffs and rigid often times helps clocks and, in this case, you could sure hear it.

I have a similar experience. I have mine mounted with long brass bolts and appropriate spacers tying the whole stack down to a wooden base. It sits on Herbie's damping feet and I put one damper puck on the regulator/DAC stack and top it off with a 5 pound lead diving weight. It is a significant improvement to add the feet, damper and weight.

I do need to experiment with additional part damping, as Soundcheck did on his Mamboberry mods. I do this pretty regularly, but normally I wait until I am sure I am finished as it is a pain to remove and change parts.

<SNIP>
At this point, there’s not a lot else to try. Maybe some galvanic isolation; but, that’s about it.

I'm eagerly awaiting both Allo's and Ian's R-Pi isolators. I do suspect they might not be the best with a bare HFBD+P due to the jitter they add... but based on the information CDSGames posted on the various configurations for clocking on the PCM5122, we really need to understand how the HFBD+P is configured to know how to best optimize the clocking environment. My experiments putting a Kali under my HFBD+P were promising, I need to spend more time with that setup.

<SNIP>
Hopefully, this inspires others to try something. You just might find that having a USB interface between you computer and your DAC isn’t all it is cracked up to be...

I have done most of this already and you still have inspired me to try more. MANY, MANY THANKS!

Greg in Mississippi

 

[soundcheck]

"...I wondered if the 1.1A max current of the LPS-1s were just too low..."

You recall what I've been saying all the time.

And the or one of the problem(s) IMO is that the LPS-1 comes with a LDO at the output -- if I'm not mistaken.

It's never been a good idea to run multistage regulation.

My guess: Just a set of huge supercaps ( I use a large Rifa instead) feeding your on-dac LDO (or the PI) might be the much better choice.
There wouldn't be any current limitations. And....
Yep. Things even get much cheaper.


I also tried other switchers (e.g. Meanwell) beside my preferred iPower. Non of them bettered the iPower.

############

Folks.


Just a hint. Latest Moode and PiCorePlayer kernels support 384kHz samplerates for HAT DACs.

Especially people running PCM51xx DACs can make use of it.
ON-DAC filters get passed by at these rates. I think it's worth it.
Give it a try!

 

[ultrafi]

I did try up sampling to 384k to avoid the filters - I didn't feel it was an improvement...BTW how many bits you select also has an effect?!

I do use the IIR filter though....

 

[soundcheck]

I use sox for resampling.

Below my old custom_convert.conf (Logitechmediaserver) entry that I used to use during
PCM51xx times.

flc pcm * *
# F
[sox] -D -q -t flac $FILE$ -t wavpcm -e signed -b 24 - rate -v -s -L 384000

I actually wrote a script that did synchronous upsampling. It converted 44.1*x to 352.8
and 48*x to 384

Note: Heavy realtime resampling on the PI might eat up the potential benefits from not
having a filter in the loop.

You also need to enable the feature inside your config.txt.

 

[jonners]

Folks.


Just a hint. Latest Moode and PiCorePlayer kernels support 384kHz samplerates for HAT DACs.

Especially people running PCM51xx DACs can make use of it.
ON-DAC filters get passed by at these rates. I think it's worth it.
Give it a try!

Is the filter also bypassed at 352.8kHz I wonder? I see no mention of that freq in the datasheet, but of course it gives a few % lower demand on the CPU if upsampling from 44.1kHz.

 

[sckramer]

Hi @soundcheck

Can you please set up and listen to this chain? I'd be interested in what you think.

— piZERO (or piA+) / 384K_kernel / LMS upsampled 384K (no PCM51xx oversampling) / kali384 / piano 2.1 —

Any pi's probably fine, those are just the ones I use for audio.

At least a good supply on the kali, up to you on powering things separately (do your thing)

note: *not* the piano 2.0, the 2.1 has an onboard LT3042 for analog out.

Thanks!

 

[pistollero]

I did try up sampling to 384k to avoid the filters - I didn't feel it was an improvement...BTW how many bits you select also has an effect?!

I do use the IIR filter though....

I tested 384khz from hbd pro with i2s out to another dac and the sound is ok but i don t feel improvement too...
On normal settings i feel diference beetween moode mpd and squezelite.
Mpd more cleaner highs but thinner sound.
Squezelite more bass and powerfull sound but not so cleaner highs.
They should play equal but they don t. Can someone explain me why??

 

[ultrafi]

Over time, I've always preferred playing at native resolution. Adding additional padding of bits hasn't caused problems.

Up sampling always sounds like this: imagine a balloon that represents the soundstage. Now, squish the ballon between two sheets of glass one behind and one in front of the ballon. Things get bigger in what I would call the x and y plane; but, the z axis suffers.

If you don't care about soundstage then up sampling is great. Likewise, if your system lacks in sound staging ability.

 

[soundcheck]

@sckramer

I'm not that motivated to setup an entire "chain" with a DAC that I took off my
wishlist long ago to run some listening test.

Actually I'm more than pleased with what I have right now - after
fiddling around with all the HAT stuff for almost a year.

I did a clean install from scratch before Christmas. Now the whole setup
is wrapped and sealed (more or less).

None of the HAT DACs I tested (HifiBerry DAc+ Pro and MamboBerry) proved to be acceptable on the long run.
They are quite OK. But if you start the HW tweaking journey you can't stop
until nothing is left from the original.
Happened to me with the HifiBerry and also with the MamboBerry - Thx Greg.

Now. Allo announced to come up with a new (quality) DAC sooner or later. Soekris also announced the DAM1231. There's movement all over the place.

For now I'll enjoy some music.

######################################
To the others:

I do agree that resampling is not necessarily the way to go. I don't do it myself.
And in my opinion Sabre DACs are the DACs of choice if you want to skip
resampling and other filtering stuff.



######################################

Ah. I did also some underclocking and undervolting on the PI3. I liked what I heard.
@scramer what were your recommendations again?

 

[redjr]

@sckramer

I'm not that motivated to setup an entire "chain" with a DAC that I took off my
wishlist long ago to run some listening test.

Actually I'm more than pleased with what I have right now - after
fiddling around with all the HAT stuff for almost a year.

I did a clean install from scratch before Christmas. Now the whole setup
is wrapped and sealed (more or less).

None of the HAT DACs I tested (HifiBerry DAc+ Pro and MamboBerry) proved to be acceptable on the long run.
They are quite OK. But if you start the HW tweaking journey you can't stop
until nothing is left from the original.
Happened to me with the HifiBerry and also with the MamboBerry - Thx Greg.

Now. Allo announced to come up with a new (quality) DAC sooner or later. Soekris also announced the DAM1231. There's movement all over the place.

For now I'll enjoy some music.

######################################
To the others:

I do agree that resampling is not necessarily the way to go. I don't do it myself.
And in my opinion Sabre DACs are the DACs of choice if you want to skip
resampling and other filtering stuff.



######################################

Ah. I did also some underclocking and undervolting on the PI3. I liked what I heard.
@scramer what were your recommendations again?

I'm glad I found this thread as I'm about to embark on a similar project based around the HiFi Berry DAC+ Pro. I haven't read the entirety of it yet.

I'm not likely to fiddle until I find that ultimate audio nivana. I'm 64 and have long lost a lot of my top-end hearing capability. Life changes!

I have read that a good linear PSU for the DAC is a requirement to get the best sound from the DAC. What did you finally use?

redjr

 

[soundcheck]

I use 2 * iFi iPower (1*PI, 1*Kali&DAC)

I did shorten the iPower cable and I added a reservoir/buffer cap (better dynamics).
I have no intention to use any of the linear supplies I have around here.

To me that "linear supply is better" story is nothing but a (old school audio -) myth.

I've also been running 2*3 LiFePO4s regulated with a TPS7A4700 to feed the DAC only. That gives a little extra
soundquality. Is it worth it!?!? Swapping the caps around the DAC chips has ten times more impact.
As I said earlier - You can't fix from the outside what's messed up on the inside.


If you start the journey now and if you don't plan to apply this or that internal mod I'd rather go for the MamboBerry DAC.
Don't let you fool with the advanced clocks of the DAC+ Pro or the HifiBerry I2S Master mode. These are IMO overrated - when compared to e.g. the Mamboberry (Sabre) in Slave mode.


Good luck.

 

[redjr]

I use 2 * iFi iPower (1*PI, 1*Kali&DAC)

I did shorten the iPower cable and I added a reservoir/buffer cap (better dynamics).
I have no intention to use any of the linear supplies I have around here.

To me that "linear supply is better" story is nothing but a (old school audio -) myth.

I've also been running 2*3 LiFePO4s regulated with a TPS7A4700 to feed the DAC only. That gives a little extra
soundquality. Is it worth it!?!? Swapping the caps around the DAC chips has ten times more impact.
As I said earlier - You can't fix from the outside what's messed up on the inside.


If you start the journey now and if you don't plan to apply this or that internal mod I'd rather go for the MamboBerry DAC.
Don't let you fool with the advanced clocks of the DAC+ Pro or the HifiBerry I2S Master mode. These are IMO overrated - when compared to e.g. the Mamboberry (Sabre) in Slave mode.


Good luck.

Hey soundcheck-

Thanks for your insight and comments. I haven't reached any conclusions yet, and will be researching this any threads for more info and user expereince. I'll likely be standing up 2 different pi-based streamers using different hardware/software and comparing the results. Lot to learn.

redjr

 

[ultrafi]

Some new news in the form of what to do - see my blog: Ultra Fi Cables

 

[Abartels1965]

Some new news in the form of what to do - see my blog: Ultra Fi Cables

seems to me like commercial activities?

 

[Greg Stewart]

seems to me like commercial activities?

And why would you say that Abartels? ultrafi provided an update to a previous post with more information, along with additional experience. He has not offered the modification for sale nor did he say he sold it to anyone.

Here we have a knowledgeable person providing useful information that can be easily duplicated and tried by anyone who can solder. What is the problem with that?

Responses like these are why people such as John Swenson, Alex Crespi, and other very knowledgable individuals from industry who in many cases got their start in DIY audio don't post here anymore.

Abartels, I appreciate your contributions in this and many other forums, especially in the Head-Fi LKS DAC thread. Please don't poke at people freely providing information, even if they are in the industry.

Greg in Mississippi

 

[Abartels1965]

And why would you say that Abartels? ultrafi provided an update to a previous post with more information, along with additional experience. He has not offered the modification for sale nor did he say he sold it to anyone.

Here we have a knowledgeable person providing useful information that can be easily duplicated and tried by anyone who can solder. What is the problem with that?

Responses like these are why people such as John Swenson, Alex Crespi, and other very knowledgable individuals from industry who in many cases got their start in DIY audio don't post here anymore.

Abartels, I appreciate your contributions in this and many other forums, especially in the Head-Fi LKS DAC thread. Please don't poke at people freely providing information, even if they are in the industry.

Greg in Mississippi

Hi Greg,

It wasn't in my intention to poke around, but whenever a link shows up in a forum like this, which points to a website where the poster himself sells a lot of stuff,
it gives me the feeling the link (maybe?) was posted for other reasons than purely and solely inform others in relation to the forum thread.

In this case the provided info for modding the RPI/HFB DAC+Pro wasn't that comprehensive that he couldn't provide the same info directly to the thread.
If it was pages long, with lot's of pictures to explain, then I would say it would justify to post the link, but not this one.

I don't doubt that the poster did this with his best intentions, but if I was moderator on the thread I would like to keep the thread clean and straight forward.

My apologies if I offended anyone, but it's just my 2 cents.

Btw, waiting for my Kali to arrive and will compare to Ian's Fifo II kit real soon

Regards,
Alex

 

[Greg Stewart]

Hi Greg,

It wasn't in my intention to poke around, but whenever a link shows up in a forum like this, which points to a website where the poster himself sells a lot of stuff,
it gives me the feeling the link (maybe?) was posted for other reasons than purely and solely inform others in relation to the forum thread.

In this case the provided info for modding the RPI/HFB DAC+Pro wasn't that comprehensive that he couldn't provide the same info directly to the thread.
If it was pages long, with lot's of pictures to explain, then I would say it would justify to post the link, but not this one.

I don't doubt that the poster did this with his best intentions, but if I was moderator on the thread I would like to keep the thread clean and straight forward.

My apologies if I offended anyone, but it's just my 2 cents.

Btw, waiting for my Kali to arrive and will compare to Ian's Fifo II kit real soon

Regards,
Alex

Alex, I understand your concerns, yet I've read a number of Larry's posts and interacted with him outside of the forum and have seen NO indication that he has tried to push any of his products in this forum or our interactions. Instead he had just been informative on what he's tried and provided experiences that anyone could copy. If he wants to post an update relevant to his earlier comments along with a useful picture of the work on his commercial website, I say more power to him

Others (not in this thread, thankfully) are not so circumspect.

Very curious to hear what you think of the Kali compared to Ian's FIFO. I have a setup with the FIFO as the interface to an R-Pi front-end ALMOST ready to run, but it keeps getting bumped for other projects. Partly because of the improvement I found from using the Kali->R-Pi as a source for several DAC projects. I now would not run the Pi as a source without the Kali, even for something like the Soekris DAM DAC which already does on-board reclocking.

Over the last week, I've been running in and tweaking up a copy of my networked LMS Server->SqueezeLite setup from my big system in my smaller system. I've been very happy with it, an all-linear supplied front end (I do need to try Larry's recommended R-Pi SMPS power brick) from the server computer to the router and WiFi router to the FMC's to provide a fully-galvanically isolated network link just before the Pi. Now that I'm getting it dialed in, it is meeting and mostly beating a Sony HAP Z1-ES I had done significant mods to, which justifies me taking it out of the system for further work.

Back to the topic of this thread, the Player-DAC combo is an R-Pi feeding the HiFiBerry DAC+ Pro I detailed my mods to earlier in this thread with a well-powered Kali between them. I'm very happy that the treble is clearer, cleaner, and yet sweeter and smoother than that of the Sony from this $45 DAC (of course modified beyond recognition AND with 4 pretty stout and massive linear supplies powering the elements of the front-end, all of which would put it as a near-cost equivalent to the Sony if the Sonys casework were excluded).

I'm definitely not saying this is the end-all setup. My Soekris setups both beat it nicely. AND I was reading that Head-Fi LKS thread to gather info I can use on that Ian FIFO-fronted setup, which uses a tried and true Buffalo-II DAC as the D-A conversion (and compared to current conventions, will need some updates)..

But this tweaked-out HFBD+P is sounding MUCH better than I ever expected and than it probably has any right to.

Greg in Mississippi

 

[Abartels1965]

Alex, I understand your concerns, yet I've read a number of Larry's posts and interacted with him outside of the forum and have seen NO indication that he has tried to push any of his products in this forum or our interactions. Instead he had just been informative on what he's tried and provided experiences that anyone could copy. If he wants to post an update relevant to his earlier comments along with a useful picture of the work on his commercial website, I say more power to him

Others (not in this thread, thankfully) are not so circumspect.

Very curious to hear what you think of the Kali compared to Ian's FIFO. I have a setup with the FIFO as the interface to an R-Pi front-end ALMOST ready to run, but it keeps getting bumped for other projects. Partly because of the improvement I found from using the Kali->R-Pi as a source for several DAC projects. I now would not run the Pi as a source without the Kali, even for something like the Soekris DAM DAC which already does on-board reclocking.

Over the last week, I've been running in and tweaking up a copy of my networked LMS Server->SqueezeLite setup from my big system in my smaller system. I've been very happy with it, an all-linear supplied front end (I do need to try Larry's recommended R-Pi SMPS power brick) from the server computer to the router and WiFi router to the FMC's to provide a fully-galvanically isolated network link just before the Pi. Now that I'm getting it dialed in, it is meeting and mostly beating a Sony HAP Z1-ES I had done significant mods to, which justifies me taking it out of the system for further work.

Back to the topic of this thread, the Player-DAC combo is an R-Pi feeding the HiFiBerry DAC+ Pro I detailed my mods to earlier in this thread with a well-powered Kali between them. I'm very happy that the treble is clearer, cleaner, and yet sweeter and smoother than that of the Sony from this $45 DAC (of course modified beyond recognition AND with 4 pretty stout and massive linear supplies powering the elements of the front-end, all of which would put it as a near-cost equivalent to the Sony if the Sonys casework were excluded).

I'm definitely not saying this is the end-all setup. My Soekris setups both beat it nicely. AND I was reading that Head-Fi LKS thread to gather info I can use on that Ian FIFO-fronted setup, which uses a tried and true Buffalo-II DAC as the D-A conversion (and compared to current conventions, will need some updates)..

But this tweaked-out HFBD+P is sounding MUCH better than I ever expected and than it probably has any right to.

Greg in Mississippi

Howdy Greg,

I suppose I was a little too fast with my preconceptions, won't happen again

It's only 2 years ago or so when I discovered that USB DDC's could bring much higher SQ from my dedicated AudioPC to my spdif based dac instead of the onboard spdif output.

Taking the route from building an expensive DualPC setup, both with AudioPhile Optimizer and Jplay, ControlPC in gui and AudioPC in cmd only, I went from modifying a Gustard U12 (see headfi forum) to a complete overhauled Melodious MX-8 (see headfi) which competed with and outperformed AudioByte Hydra-Z.

I then tried isolated XMOS interfaces with I2S output, which outperformed all previous connected spdif based DDC's. I do prefer Amanero based interfaces above Xmos btw.

Since all USB strugglings, cabling, regenerating and so on, influencing SQ, I decided to ditch USB based DDC's en finally went the low computing route with RPI.

First test was with modified HFB DAC+Pro (modified with NDK NZ2520SD's and and powered 3.3V and 5V with 0.8uV low noise psu's listened to analog output. Not bad at all. Finally I only used it for I2S output to feed my MonsterDAC. Combination of RPI and Ian's FIFO II kit brought me a very big increase in SQ, and I didn't need HFB DAC+Pro's I2S anymore, straight I2S into FIFO II did the trick.

Still listen to this setup on a daily bases, but I experience some troubles with Ian Canada's FIFO II kit, which probably is defective right now.
It sometimes skips fractions of music sections ahead, 1/4 second or so, and it also accepts 176,4 kHz only for the moment, not good at all,,,,,
Ian is very busy lately, but I am very confident we can solve this in one or another way,,,

About Soekris, I will definitely try one in the near future, or maybe an Audio-GD R2R board. Soekris, at least the current version, is too difficult to mod to power it the way I prefer, and, also the programmable clock isn't an option I would choose for, would love to see possibility to power ALL the psu lines independently (external) AND have the possibility to have two exchangeable audio clocks on board. Not sure if the upcoming version does provide all this features. I'm very confident it would be a REAL killer if we could lay hands on such a version and power it the way I did within my MonsterDAC.

Many people think I am crazy when looking at my MonsterDAC, but, the ones who listened to it lately all are convinced of the monstrous design

Hope Allo releases their Isolator HAT real soon, waiting for Ian's isolator too.

Have a good weekend!
Alex