Sounds great.
The other day I found a RPI HAT template - A Pi HAT in ten minutes! (Perhaps
you find thew required infos in there)
Looks like a piece of cake.
Maybe you do it in 20minutes.
Local clock:
Yep I'd agree. The clock section pads should be onboard - just in case.
This project shouldn't become a Kali sales push project.
And please the PS section/input should go the left and not right as on the Mambo.
Also a jumper to be able to get powered from the bottom board (option) would be nice.
Looking forward to the project.
I guess you'll open a Subbu-HAT thread, if you guys decide to go for it.
Just let us know.
Thx,
The other day I found a RPI HAT template - A Pi HAT in ten minutes! (Perhaps
you find thew required infos in there)
Looks like a piece of cake.
Maybe you do it in 20minutes.
Local clock:
Yep I'd agree. The clock section pads should be onboard - just in case.
This project shouldn't become a Kali sales push project.
And please the PS section/input should go the left and not right as on the Mambo.
Also a jumper to be able to get powered from the bottom board (option) would be nice.
Looking forward to the project.
I guess you'll open a Subbu-HAT thread, if you guys decide to go for it.
Just let us know.
Thx,
Damned, after the sunglasses comes the Hat...*
I'm here for cape tests fi you turn a serie again,
Zorro...
* ADM-7150 LDO regs, NDK, AVX Crystal, isolator chip, FIFO (cost less and less ?).... and finally an AK chip instead the ES9023, lol ! (fed up with these ESS high frequency noise)
I'm here for cape tests fi you turn a serie again,
Zorro...
* ADM-7150 LDO regs, NDK, AVX Crystal, isolator chip, FIFO (cost less and less ?).... and finally an AK chip instead the ES9023, lol ! (fed up with these ESS high frequency noise)
Last edited:
If anyone's thinking seriously about a Sabre-based PI-HAT, why not use the ES9018K2M instead of ES9023? Main thing that interests me about the 9018 is the I2C interface, which allows changing digital filters and also digital-domain volume control. It seems typically people implement a microcontroller for interfacing with the ES9018 I2C, but since you're talking about interfacing with an RPi, you implicitly already have something that can function as a microcontroller (see the wiringPi C library, or RPi.GPIO in Python).
I am actually working on this using the DIYINHK ES9018K2m board. Not an RPi HAT, but similar. When done it'll be Rpi -> Allo Kali -> DIYINHK ES9018K2M. I'll have a rotary encoder connected to the RPi as well, so I can physically control volume, or control it via software.
The Asahi Kasei chips (e.g. AK4490) also offer I2C control... at least of the various digital filters (can't remember if they do volume control or not).
I am actually working on this using the DIYINHK ES9018K2m board. Not an RPi HAT, but similar. When done it'll be Rpi -> Allo Kali -> DIYINHK ES9018K2M. I'll have a rotary encoder connected to the RPi as well, so I can physically control volume, or control it via software.
The Asahi Kasei chips (e.g. AK4490) also offer I2C control... at least of the various digital filters (can't remember if they do volume control or not).
My two cents about -- why ES9023
1. The 9023 can sound really good if well implemented. I think it's not just me thinking that way. Especially after my latest mods.
2. many 9018 sound even worse because they are poorly implemented and/or
3. they are much more complex (SW and HW) implemented. Many times less is more.
4. The whole filter swapping is more or less nonsense and makes stuff just more
complex again. If a DAC manufacturer can not decide what's the best filter
he hasn't done his homework - IMO. Or he just wants to build/offer a costly playground for some audiophile fools. Look at AKM. If you look at their communicated filter responses it's pretty obvious what to go for - without even listening to it. No reason to have a choice of 5.
5. Budget - I think cost/performance ratio is the best out there.
Enjoy.
1. The 9023 can sound really good if well implemented. I think it's not just me thinking that way. Especially after my latest mods.
2. many 9018 sound even worse because they are poorly implemented and/or
3. they are much more complex (SW and HW) implemented. Many times less is more.
4. The whole filter swapping is more or less nonsense and makes stuff just more
complex again. If a DAC manufacturer can not decide what's the best filter
he hasn't done his homework - IMO. Or he just wants to build/offer a costly playground for some audiophile fools. Look at AKM. If you look at their communicated filter responses it's pretty obvious what to go for - without even listening to it. No reason to have a choice of 5.
5. Budget - I think cost/performance ratio is the best out there.
Enjoy.
Yep.
You'd need to keep the I2S pins where they are, just in case you're coming up with an updated Kali version.
####
One more.
The other day a Kali related thought crossed my mind.
If the Kali would offer a Toslink input with automatic source detection, we wouldn't need such an input on either a PI or the DAC.
On top we'd get a perfectly reclocked Toslink signal on our HATs.
Background: I finally would have a solution at hand to connect my TV to my stereo.
And with Toslink it's also isolated.
Cheers
Yep. Give us an inch, we'll take a mile.
You'll grow with the challenge.
As I said earlier that's a "nice-to-have" option and just reflects my personal wishlist. Nothing to worry about.
Me? Not now.
From an overall product management perspective:
The 3.3V is a must. Otherwise you're incompatible with other HAT DACs.
That's IMO been a design-flaw of your REV1 board that needs to be fixed.
You just heavily limit your target market if you don't do that.
And the more I think about it:
A very minimalistic ES9023 approach could even use that "clean" Kali supplied 3.3V (I hope you read and/or know what LDO to go for).
That'll allow for an ever lower initial BOM. Then somebody could use the MCLK and just a few parts and we might see a rather
nice $15-20 (XXL Salami-Cheese sandwich price-range to stay with Vasilis terminology) no frills ES9023 HAT coming.
Cheers
Last edited:
Dont tell me the LT3042...it adds more than 2$ to the BOM...how about a nice (and cheap) LP5907
Yep. That'll do. That's the one used on the MamboBerry too.
But. Good that you bring that LT3042 up. Noise performance is outstanding.
Folks. Wouldn't that be the right choice for a highest quality Sabre HAT implementation ?
Last edited:
Klaus, Greg, @Dimdim, Ioan:
Loosly related, quick vid setting that cheepie Sabre Dac to sync mode:
https://www.youtube.com/watch?v=fJ1Uq6ZyHLc
Greg, even tried short i2s wires, forcing diff bit rates, 32bit etc, that kali192 just will not work compared to the 384, interesting...
Loosly related, quick vid setting that cheepie Sabre Dac to sync mode:
https://www.youtube.com/watch?v=fJ1Uq6ZyHLc
Greg, even tried short i2s wires, forcing diff bit rates, 32bit etc, that kali192 just will not work compared to the 384, interesting...
Last edited:
We also got some weird behavior with the Kali 22/24Mhz.. we will investigate but it will take a while
So.
I also did the sync-MCLK-mod surgery.
1. relocated the 51R ( I cut the trace that connected to pin 27 - this way viases 29/27 make some nice pads for the 51R )
2. removed the clock LDO
3. used a 3-wire ribbon (gnd-clock-gnd) for the time being
4. Against earlier plans I'm using the old 51R pad to inject the clock.
5. I bent the ribbon away from the other I2S lines. (routing is not as shown on the image)
Now it's heating up.
PS1: Sorry Vasilis, Not much left of the MamboBerry.
PS2: What you see on top of the caps are NOT my old chewing gums.
I also did the sync-MCLK-mod surgery.
1. relocated the 51R ( I cut the trace that connected to pin 27 - this way viases 29/27 make some nice pads for the 51R )
2. removed the clock LDO
3. used a 3-wire ribbon (gnd-clock-gnd) for the time being
4. Against earlier plans I'm using the old 51R pad to inject the clock.
5. I bent the ribbon away from the other I2S lines. (routing is not as shown on the image)
Now it's heating up.
PS1: Sorry Vasilis, Not much left of the MamboBerry.
PS2: What you see on top of the caps are NOT my old chewing gums.
Attachments
First impressions:
My ears are bleeding. I'm serious. I'm still waiting that the ear-ringing disappears.
The relaxed and very nice presentation from last night is gone.
First it seemed cleaner on one side. But then it is also been more aggressive.
Further I'd say partially things get kind of bloated. Instruments getting big in the front row.
Timbres are lowered.
The sound really turned into the wrong direction.
Overall it seems that the potentially good points that come with the sync mode get
eaten up by certain negative aspects of that (my) compromised MCLK mod-setup.
But that's just a wild guess.
Of course there's also an option that flaws that existed earlier are becoming obvious now.
I can't measure anything of relevance. My 100MHz scope shows me nice MCLK sine waves.
I'll give it some time again. That proved to be the right strategy several times during the project.
However. I actually don't think that this mod needs much of a break-in.
I might try a coax tomorrow instead of the ribbon. That's one idea.
If that doesn't work, I'll roll back.
Cheers
My ears are bleeding. I'm serious. I'm still waiting that the ear-ringing disappears.
The relaxed and very nice presentation from last night is gone.
First it seemed cleaner on one side. But then it is also been more aggressive.
Further I'd say partially things get kind of bloated. Instruments getting big in the front row.
Timbres are lowered.
The sound really turned into the wrong direction.
Overall it seems that the potentially good points that come with the sync mode get
eaten up by certain negative aspects of that (my) compromised MCLK mod-setup.
But that's just a wild guess.
Of course there's also an option that flaws that existed earlier are becoming obvious now.
I can't measure anything of relevance. My 100MHz scope shows me nice MCLK sine waves.
I'll give it some time again. That proved to be the right strategy several times during the project.
However. I actually don't think that this mod needs much of a break-in.
I might try a coax tomorrow instead of the ribbon. That's one idea.
If that doesn't work, I'll roll back.
Cheers
Last edited:
You're welcome!
I found it lacking either way, it's bass really annoyed me so could not get past that, not sure what it is-- really using this board as an i2s tester, not to listen to.
But did make it selectable with a jumper, and as a side benefit can power it's on-board clock separately. I'll try to identify what sounds different between the 2 modes sometime, maybe
Ext MCLK / Onboard MCLK:
Attachments
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.