• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

New FIFO buffer for RPI/SBCs

Member
Joined 2008
Paid Member
I have tried this with promising, but so far mixed results. Especially if you have a Rev 1 or 2 Soekris before he added HW reclocking after the FPGA, it should be useful.

Are you sure about that? I'm under the impression that where the reclocking is done hasn't changed in any dam1021 revision; it's only changed in the completely different board, the dam1121 OEM module.
 
Member
Joined 2003
Paid Member
I have got the latest soekris, rev 4.

&

Are you sure about that? I'm under the impression that where the reclocking is done hasn't changed in any dam1021 revision; it's only changed in the completely different board, the dam1121 OEM module.

My bad, it was only on the other models that Soren added post-FPGA reclocking

See my post here where I reference several other posts referring to the DAM reclocking including Soren's post saying that he added them to the 1121 & 1101:

http://www.diyaudio.com/forums/vend...magnitude-24-bit-384-khz-108.html#post4695897

Sorry for the confusion.

There as some other good posts by Nige2000 where he discusses his experiences with various reclocking schemes, well worth reading.

Greg in Mississippi
 
Last edited:
Question... I've not been able to get my modified HiFiBerry DAC+ Pro to work with a Kali. I have updated the config.txt to add the dtoverlay slave parameter (I've seen it listed as either ',slave' or ',slave=on', tried both).

This is on PCP 1.22 (which I prefer sonically, however I've not tried anything newer than 2.05) and PCP 2.05.

Do I need to try a newer version of PCP or am I missing something else?

TIA!

Greg in Mississippi

P.S. I just posted some comments on the Uptone Audio LPS-1 Listening thread over on CA, the R-Pi->Kali->various DAC HATs are quite improved using 1 or 2 LPS-1 supplies in place of my DIY linear supplies, my modified HFBD+P even sans Kali was stunning, need to hear it on a Kali!

Hi Greg,

Your dtoverlays look correct. It's probably the old version of PCP.

I've only done this on a fresh compile of Clive's 4.4.x simple & a squeezelite 80x compile.

I can try it on piCorePlayer 3.02 for you--

--Scott
 
I'm not aware of any 9023 HATs that could support that, but I've tested a 9023-based DAC clocked by a local high quality oscillator (Crystek) vs. MCLK from Kali with no significant changes in SQ.

However, if the 9023's local oscillator is low-end, running it synchronously with the Kali will probably improve things.
 
Member
Joined 2003
Paid Member
The Mamboberry DACs will require a mod to run in sync mod, but it is a minor one (at least in my world!).

If you look at the great picture Soundcheck posted of the MBLS in his thread on the DAC, you'll see just to the left of the clock a 51R1 resistor. The top pad of that resistor goes to the MCLK pin on the DAC (pin 13). You remove that resistor, then connect from the MCLK pin from the Kali to either the top pad of the resistor (ok) or directly to the MCLK pin 13 (better, but requires even bettery SMD soldering skills). Best to use a section of the coax from a u.fl cable, making sure to connect the ground close to where you solder each end. You might even be able to fit in a u.fl socket at that pad and the pad for the MCLK pin from the Kali on the expansion area of the MBLS.

If you like it, un-power the clock by cutting the 5v power trace to the input of the clock's regulator. See the rendering of the clock section of the circuit board that VASILIST posted. Do that right at the regulator and you can easily re-engage the on-board clock by scrapping off some of the mask at the regulator input lead, doing a small solder bridge to re-power the regulator and clock, removing your MCLK connecting cable, and replacing that 51R1 resistor.

Greg in Mississippi

P.S. This works about the same way with the original Mamboberry, the resistor is a 100R instead of 51R1 and you are cutting the clock power trace right at the local reservoir capacitor, but otherwise it is functionally the same. My mod list for the Mamboberry's will include this.
 
Last edited:
Member
Joined 2003
Paid Member
Hi Greg,

Your dtoverlays look correct. It's probably the old version of PCP.

I've only done this on a fresh compile of Clive's 4.4.x simple & a squeezelite 80x compile.

I can try it on piCorePlayer 3.02 for you--

--Scott

Thanks Scott. I'll load up an SD card with 3.02 this weekend and give it a try. If it doesn't work for me again, then maybe you can try it and help me figure out what I'm doing wrong.

I'll let you know how it comes out!

Greg in Mississippi
 
Member
Joined 2003
Paid Member
I'm not aware of any 9023 HATs that could support that, but I've tested a 9023-based DAC clocked by a local high quality oscillator (Crystek) vs. MCLK from Kali with no significant changes in SQ.

However, if the 9023's local oscillator is low-end, running it synchronously with the Kali will probably improve things.

I have heard comments by others who tried running the ESS DACs in sync mode that the presentation was different and they preferred it to the typical asynch configuration on those DAC chips. This included several people commenting in Ian's FIFO thread and Barrows talking about running the Sonore USB-I2S interface in sync mode.

But like you said, you have to have no worse of a clock signal going into the DAC in sync mode for it to work better. I'm not convinced we can do that by feeding the clock from the Kali, but it is cheap, easy, and worth a try.

Heck, if one wanted to burn money, I saw that the Pulsar clocks are available again... put a couple of those on your Kali (with appropriate mods to handle the increased power requirements) and I BET one would be in hog heaven... and there sync mode would be a no-brainer!

Greg in Mississippi <- being silly there at the end!
 
I'm not aware of any 9023 HATs that could support that, but I've tested a 9023-based DAC clocked by a local high quality oscillator (Crystek) vs. MCLK from Kali with no significant changes in SQ.

Further to what Greg said about the Mambo, with the Audiophonics.... you just move the 100R resistor from R2 (enable on-board OSC) to R1 (enable external MCLK) position if you want to use external MCLK.
 

Attachments

  • audiophonics-i-sabre-dac-es9023-tcxo-raspberry-pi-a-b-20-i2s_ext_osc.jpg
    audiophonics-i-sabre-dac-es9023-tcxo-raspberry-pi-a-b-20-i2s_ext_osc.jpg
    123.9 KB · Views: 499
Last edited:
Hello i have a problem powering my Kali. Every powerjack i've tried does not fit. The pin inside Kali's powersocket seems to be too thick.
Can i use the pins beside the powersocket? I mean the pins of wich the documentation states "Do not use. For internal testing only" ?

Best regards,
Bart
 
Hello i have a problem powering my Kali. Every powerjack i've tried does not fit. The pin inside Kali's powersocket seems to be too thick.
Can i use the pins beside the powersocket? I mean the pins of wich the documentation states "Do not use. For internal testing only" ?

Best regards,
Bart

Hi Bart, I use the 2 pin header next to the power socket (I didn't have anything to fit either, and continuity buzzed out as connected) It's dead easy to get the polarity reversed so be careful!