Soekris' DAC implementations

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
E

Eldam

Banned
Joined 2012
what you need is always a gnd pin near a signal one for the closest return path while this gnd return path (G) allow some isolement from the next I2S signal path (S).

So you need a flat cable with 6 wires : SGSGSG or 3 coax cable where the shielding is the gnd return path. verify the impedance of the cable !

So i believe the Wave I/O (not have yet, waiting mine) have a non isolated output with uf-l conectors and isolate output with a plug wich allow GSGSGS...

Choose what is the most useful for the Soekris's input side... some cut a side of an uf-l at one side when no-ufl are present at one side, others prefer coax on a connector... sure some digikey/mouser/farnel could be interrest us to match the size pitch pins of the Soekri's inputs...

some good readings about glass fiber and short interconnects : DIYHiFi.org • View topic - Optic fiber Is Stereophile mag knowing more things than Jocko ? Well finaly advise the SB Touch owner to stay with SE SPIDF output to profit from 192 Khz...as the toslink on it is limited (my understanding) at 96 Khz though the upgrade at 192 Khz (in the particular case of SB Touch's toslink...)

Well after some readings on Slimdevices : both squezebox server is limiting the Duet receiver at 44 Khz as also the hards architecture embeded in the Duet receiver.... (with the Subbu/Duet combo 75 ohms interconect and some good tweaking I have a as good quality than a good SACD player and an other CD 24/192 CD player... so I will considere to swap with a squeezelite server on a BBB feed from an external clock one later day, not yet :), boring all this wires and micro devices...:yell:
 
Last edited:
E

Eldam

Banned
Joined 2012
bambadoo said:
Just ordered a gustard u12 usb-i2s converter. Will try to see how it performs with soerens dac
Click to expand...

Have you not expensive administrative customs declaration fees from DHL/UPS with it (around 27 euros here for their administrativ task :rolleyes:) + customs taxes: + 30% or with a minimum below 25 euros ????

I given up to buy in USA with such DHL/UPS shipment because all the hassles.... more money spent with taxes than with the products !
 
E

Eldam

Banned
Joined 2012
What sort of pot is needed for the volume control please ?

log or linear ?

Is it a mono plot which command 2 channels (stereo) or each canal need its plot (dual plot or two mono plot for a fine balance correction :) ) ?

10 k ? 100 K ?

Sorry if already asked elsewhere in the vendor thread
 
S

soekris

Member
Joined 2009
soekris.dk
Paid Member
Eldam said:
What sort of pot is needed for the volume control please ?

log or linear ?

Is it a mono plot which command 2 channels (stereo) or each canal need its plot (dual plot or two mono plot for a fine balance correction :) ) ?

10 k ? 100 K ?

Sorry if already asked elsewhere in the vendor thread
Click to expand...

Any mono linear one from 1K to 10K, preferable high quality conductive plastic to avoid noise....
 
Bones13

Bones13

Member
Joined 2006
Paid Member
abraxalito said:
That's a bifilar wound toroid. Assuming you can get a core with enough inductance and low enough loss (a bit tricky) then you'll have a large interwinding capacitance. This couples common-mode noise fairly effectively. With a core as large as they've chosen it'll probably still have more leakage inductance than a SC, despite being bifilar.
Click to expand...

I ordered an AN one from HiFiCollective, I figure that they already have done the work of making these well.

I noticed some from Lundahl as well :

LUNDAHL LL1574 TRANSFORMER Digital audio impedance matching, PC mounting, 100/75 ohms
 
Last edited:
spzzzzkt

spzzzzkt

Disabled Account
Joined 2005
Eldam said:
some good readings about glass fiber and short interconnects : DIYHiFi.org • View topic - Optic fiber Is Stereophile mag knowing more things than Jocko ?
Click to expand...

The comments about glass and short interconnects were not made about TOSLINK.

Jocko was responding to this comment:
Now, if you would use a real laser (VCSEL) and glass fiber (even if 'only' OM1 or OM2), I am sure the results of this discussion would different.
Click to expand...

These are Fibre Optic formats used in telecommunications and networking. OM-1 can be used to carry 1Gbit ethernet up to 275m and 100Mbit ethernet up to 2000m. Jocko was arguing that these and other "glass" formats were not suitable for 1-2M connections.

If you actually read the thread the argument is made that while jitter reduction should make the type interface irrelevant it is hard to do well, even for professional designers.

Thorsten Loesch (AMR/iFi designer) comments that

Well, option one, we remain in the real of standard devices for audio (read Tosh-Link [pun intended]). The answer is - No. We cannot improve optical. We can improve clock recovery from the SPDIF (or I2S/EIJA) stream, regardless of source, but the methods that deliver real performance gains are not simple, or easily amenable to DIY, unless someone makes a pre-packaged module available to DIY'ers.
Click to expand...

It is important to understand that a lot of the debate in the thread is based on the DIYer being stuck using an off the shelf DIR (CS841x, DIR9001, Wolfson etc) and is forced to use jittery clock recovered from the incoming data. COAX SPDIF is recommended because it offers the best jitter performance and easiest implementation for a DIYer under these circumstances.

The DAM1021 has very good (but perhaps not ultimate) FIFO setup that is very effective in rejecting jitter on incoming data streams, which is precisely one of the "methods that deliver real performance gains". The more you improve jitter rejection, the less sensitive the DAC becomes to jitter on the interface, so difference COAX, TOSLINK, AES/EBU becomes less of an issue to the point with perfect jitter reduction there should be no difference at all.

The argument made in this thread by TNT, Julf, Abaraxlito and others is that because the jitter reduction of the FIFO used in the DAM1021 there will be no audible difference between RCA and TOSLINK SPDIF. That isn't disputed in the thread you have linked to.

There is no point arguing that because you hear a difference using a DAC with poor jitter reduction, or find a thread with comments based on using DIR's with poor jitter rejection that the recommendations made for the DAM1021 are wrong. In doing this you are ignoring important technical differences that significantly influence what advice is relevant and what is not.


cheers
Paul
 
Last edited:
abraxalito

abraxalito

Member
Joined 2007
spzzzzkt said:
The argument made in this thread by TNT, Julf, Abaraxlito and others is that because the jitter reduction of the FIFO used in the DAM1021 there will be no audible difference between RCA and TOSLINK SPDIF. That isn't disputed in the thread you have linked to.
Click to expand...

My argument is a little more nuanced than that. I'd argue there will almost certainly be no audible difference due to jitter. However with extreme jitter rejection (as we have here) the improved isolation of optical would be my choice over (wired) S/PDIF. In my experience with multibit DACs, the audible degradations I've heard have been tracked down to common-mode noise, not jitter. Optical by far gives the best rejection of CM noise.
 
spzzzzkt

spzzzzkt

Disabled Account
Joined 2005
abraxalito said:
My argument is a little more nuanced than that. I'd argue there will almost certainly be no audible difference due to jitter. However with extreme jitter rejection (as we have here) the improved isolation of optical would be my choice over (wired) S/PDIF. In my experience with multibit DACs, the audible degradations I've heard have been tracked down to common-mode noise, not jitter. Optical by far gives the best rejection of CM noise.
Click to expand...

Sorry if I misrepresented.

One final quote from Jocko in the thread Eldam linked to over at DIYHifi, which should clarify for once and for all that Jocko's comments about glass and short lengths were not aimed at TOSLINK.

Jocko@DIYHifi said:
No kidding TOSLINK works. A paper clip probably works about as well.

I am specifically addressing the idiotic notion that you can use a SM fiber system for 1-2 m.

Is that clear enough now?

(Ironic part is Toshiba's later versions actually worked better than the early ones. Until the factory was flooded, and they stopped making them. The other brands are worse. To make things worse, Sharp will not sell their version in the USA. It isn't in their catalog, unless you go to a EU site. And it says "Not for sale in USA". They are doing us a favor.)
Click to expand...

So if you are going to use TOSLINK, use Toshiba RX modules: TORX177 for 5V and TORX147 for 3.3V
 
Last edited:
E

Eldam

Banned
Joined 2012
Hummmm, my bad ! I missunderstood the thread !:(... didnt know some used fiberglass interconnect without Toslink interfaces !

Ah ! believe some will make the experiment with their SB Touch between the 96 limited Toslink (if the information i read is still good : they say after the 192 plugin correction !) and the 192 Khz spidf...

Btw, I am limited at 44 with my prehistoric Duet... hopping it makes no great difference with the few 96 Khz and more I own ! Never tried any FIFO indeed ! Should be good not to hear difference between wire, bordeaux glass bottle wire, caps, etc...
 
S

SebastianL

Member
Joined 2013
Eldam said:
Btw, I am limited at 44 with my prehistoric Duet... hopping it makes no great difference with the few 96 Khz and more I own !
Click to expand...

My experience with R2R dacs really made me reconsider the usual critisism concerning the limitations of 44.1 redbook material. A good recording in 44.1 format with a good R2R dac, and you can forget all about hi-res files imo. So I’m really looking forward to hearing the dam1021 :)
 
E

Eldam

Banned
Joined 2012
Ah, had the same idea : a unity gain valve stage with a good trafo for lowering the voltage... directed heated tube with some good colorations and nice harmonics (or a 1/2 5687 ?!) Just for lowering a little the voltage for an even better impedance adaptation if no preamp (or no tubes preamp?). (have a 100 K input amp ! is a 1 to 10 ratio not better in voltage impedance adaptation?)

What about also a little two channels 60 euros Lundhal traffo just for the SE (or the balanced with an other shematic)! 1:1 ?

It could be interessant to compare the bare SE output with such a good buffer valve with a good traffo (off shelf ? are there any acceptable devices yet without Taylor made Tango, Tribute, Tamura ! amorph Lundhal... have to sell my car viewing the prices of such trafffos !).


Is there an edited ID list to know if members are in first or second grup ? I mean some have paid thinking they are on the first grup while being in the second ?! Or maybe the day when the number of the first grup ended to let see us by ourselves if we recorded ourselves before or after this date ?

Soren, I'm not specially in a hurry and I'm sure you will take all the time you need for checking for the best quality, programming, feeding the rom, etc , but shall we have a mail to us by UPS when they will record the shipment ?

But mainly forgett my asking if lacking of time :)

PS : anyone for a good cheap pot ref. with the quality advised by Soren to update the first page... imagine some have already bought all the parts needed while some others will wait the manual provided with the DAC... !

regards
 
Last edited:
R

rol

Member
Joined 2010
Yes, thinking of adding a switch between unbuffered output and tube output to compare the sound. Most of the R2R DACs I tried, played beautifully with tube stage. I am kind of bored with popular ones like 5687,6dj8,12ax7... still finding something more special. I have some very skillful friends who are able to wound superb OPTs, just need ideas.
 
E

Eldam

Banned
Joined 2012
Lucky you are, few people are able to make little good audio OPTs... really not the same job than industrial traffos. One able to make the firsts can make the seconds but not the opposite most of the time !
Do you think about a direct heated tube ?

When people will have their DAC, this modest waiting thread to buy the basic parts needed will go to its end and I hope some fellows will open some threads with ideas like your !
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom