ackoDAC based on ES9018

[qusp]

no problem, if youve got the coin and the knowhow, the borbely kits are probably the best choice out of those IMO, but for a balanced build you need 2 of them, so it gets steep pretty quick and the D1B1 is basically pretty similar to his I/V stage anyway. tho would require a bit more experience perhaps. erno is apparently working on a fully differential design too, but unsure of how long it will be till release. the QRV08 should be very high quality and clean with that power supply onboard, I just worry a little about the combination being a little TOO clean with not enough balls. with any luck the fact i'm taking it up a notch by making a dual mono bridged version, with mundorf MKP 3300uf mlytic in place of the stock PS elna 2200uf caps and most likely upgrade transformers from a fairly small 4va to 15va and removing the input caps altogether due to CMMR eill help out with this

 

[jkeny]

The output impedance of the Sabre is 400 ohms (200 on each phase).

Thanks for clarification Leo,
All these values running around in my head - I can't keep track of. That's why I look up datasheets when I am ready to do any mods.

 

[qusp]

195 actually but I wonder given the slightly unconventional output previously mentioned whether we can act as normal? also I guess the voltage outs will be completely different again. the 9012 is also slightly different in setup, so I wonder if they are the same. i'm going for the 9012 on mine, there is even less info getting around about this chip and although very similar to 9018, I am sure there will be some differences since the channels are already combined down to 2 in the chip

so I dont think this changes your info on 900R john, they are 2 totally separate stats and although related somehow i'm sure, I was told by brian some time ago that to add any more than about 700-750R (cant remember the exact number) a the outputs of the b32 that it would become a voltage output, and I think this was per pole, so matches well with your number of 900. that is what I was talking about. so I think it is going to be a bit of a tight squeeze with loading the XFR if it is 600R like I think (I will confirm with Bud)

yes a datasheet would indeed be nice :S

 

[jkeny]

Hi All,
Acko has told me that it's alive & undergoing sonic testing as soon as he gets back from a 2 day business trip. So we should soon be seeing the fruits of this particular labour & enjoying it's sound/experimenting with the options

 

[SR_audio]

Anyone have an idea of periodic jitter reduction vs frequency on sabre dac?

 

[qusp]

indeed, I heard from acko a couple of days ago with this news. He is taking a well earned break and will be back to continue testing and organize release tomorrow. the final BOM is done and I have been given permission to post it in case some wish to start preparing for release.

View attachment ackodac BOMs.zip

View attachment ackoDAC Setup.zip



there is also an I/V stage PCB using OPA1632 available, but the parts list is still not finalized

 

[roscoeiii]

Many thanks qusp. Very exciting. Time to price this baby out...

 

[qusp]

important ammendment

Hi guys, glad I remembered this, I spotted a typo that had slipped through the cracks for the dac board BOM OU1 decoupling caps, the farnell part number was actually for an X7R part and this wont di at all, i've just noticed that the reg BOMs I posted the other day may have this same mistake on them, the line is listed as a 1206 COG/NPO part in the BOM document, but the farnell part is incorrect. I cant edit the post now, so i'm stuck with it, for the moment, will PM one of the MODs, anyone you guys can recommend? but yeah anyway if anyone is actually at the point of ordering parts, please hold off till tomorrow and i'll confirm with acko, but i'm 99% sure that should be a 1206 COG ceramic, the one used on the dac board is fine; the murata part is COG/NPO.

 

[jkeny]

Are you sure, qusp? The Digikey part no in the xls file is correct, I think 490-1767-1-ND. However, there is a minimum order qty of 10 so an order of 20 is needed at $15.60.

Also Mouser is significantly cheaper (what's new) for the Panny SP-Caps (667-EEF-UE0J221LR) 220uf 6.3V 7mOhm @ $3.77 each so about $50 intead of $60

 

[acko]

AckoDAC Release

Hi All,

This DAC module has now passed all electrical and sonic tests and pretty much ready for release. All those who have made reservations or enquires have already been contacted. There is a bit more work on the external MCU to optimise performance and test all the listed features. At this stage I am running the DAC in its 'safe' default mode (even then it is very clear and detailed). I am also getting another build ready for 'Qusp' and 'Jkenny' to do more testing with their audio gear to check and compare.
There are lots more things to come and will keep you all posted.

Cheers, Acko

 

[jkeny]

Sonic testing of the DAC is proving very fruitful & according to Acko the unit is

even with my 'crude' IV stage with the burnt out bypass caps and all the wires hanging around powered by a single bench psu

Tested with SPDIF, I2S should be better.

Power Supplies are undergoing rigorous testing now - no stone being left unturned

Edit: Oops cross-post with Acko

 

[Spartacus]

so I dont think this changes your info on 900R john, they are 2 totally separate stats and although related somehow i'm sure, I was told by brian some time ago that to add any more than about 700-750R (cant remember the exact number) a the outputs of the b32 that it would become a voltage output, and I think this was per pole, so matches well with your number of 900.

The sabre chip will output both current and voltage with any finite impedance. How much voltage or current it supplies will vary linearly with output impedance - it won't suddenly turn from one into the other at a certain point.

The white paper indicates that the reduction in THD with voltage mode is down to a voltage coefficient in the internal resistance of the Sabre, so it seems reasonable that THD will linearly improve as output becomes more current and less voltage. Whether there is any benefit in below 108dB THD (the V out spec) is debatable though.

 

[ra7]

Acko,

Please count me in for one DAC, maybe IV and anything else that may be put on offer.

 

[jkeny]

Yes Leo, you are correct - it is a spectrum from low to high impedance load of current output to voltage output.

 

[glt]

It seem though it is best to use it as either current mode or voltage mode. Something in the middle seems to be worse.

Suppose you use a large impedance so that you get the full voltage swing - full voltage mode. Suppose you use a very small impedance so that your voltage swing is ~zero -full current mode.

Now suppose you use some impedance so that you get 1/2 the full voltage swing? Now, wouldn't you have thrown away 1/2 your dynamic range???

 

[qusp]

The sabre chip will output both current and voltage with any finite impedance. How much voltage or current it supplies will vary linearly with output impedance - it won't suddenly turn from one into the other at a certain point.

The white paper indicates that the reduction in THD with voltage mode is down to a voltage coefficient in the internal resistance of the Sabre, so it seems reasonable that THD will linearly improve as output becomes more current and less voltage. Whether there is any benefit in below 108dB THD (the V out spec) is debatable though.

you are correct, badly worded post there, I guess what I mean is at that point it becomes more voltage than current. now what point is too low I wonder; wish they would get back to me with NDA, lucky i'm doing primarily subjective testing

 

[qusp]

Are you sure, qusp? The Digikey part no in the xls file is correct, I think 490-1767-1-ND. However, there is a minimum order qty of 10 so an order of 20 is needed at $15.60.

Also Mouser is significantly cheaper (what's new) for the Panny SP-Caps (667-EEF-UE0J221LR) 220uf 6.3V 7mOhm @ $3.77 each so about $50 intead of $60

not the ackodac PCB, that was spotted and fixed in the next revision; the LN regs, (single and dual version) as well as the digital reg all have farnell 0U1/0.1uf 1206 ceramic part 1582786 which is X7R, though the line is listed as being a COG/NPO part, this is what it was like on the DAC BOM, but was spotted when I was doing some early ordering. so it should either be changed to X7R in the BOM document and therefor the BOM is correct, or the part number should be changed to the same COG/NPO part 490-1767-1-ND or similar. mouser isnt cheper for everything though, thats the annoying thing, I reckon they've all got some sort of deal going on . some parts are just not available at mouser though, a couple of odd resistor values etc. was just a typo that propagated through the system for the public BOMs, because when I brought this up, acko found he had the right parts both installed and in stock at home. hopefully acko will confirm this one should be changed as well so I can edit and repost them.

 

[msjjr]

Could I also place myself in line? I would want a unit fully assembled. Is this available, or are you just selling a kit? Should I email or send a PM?

Thanks.

 

[acko]

Orders

Could I also place myself in line? I would want a unit fully assembled. Is this available, or are you just selling a kit? Should I email or send a PM?

Please send PM first for reservations and queries. Available as fully assembled and tested modules or kit of parts as suited.

Regards, Acko

 

[acko]

Website

Hi All,

See attached some updates on the setup of this DAC. I have also setup a site for all the resources including ordering information. No stock at the moment but new batch of parts and modules is expected in 2-3wks.