Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 12, 2012 17:41:54 GMT
Do I spy a Lorlin rotary power switch a la Naim Well done for getting it finished (for now) and ship shape
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 12, 2012 22:07:13 GMT
Do I spy a Lorlin rotary power switch a la Naim Well done for getting it finished (for now) and ship shape Hi Chris Well Done Lorlin rotary mains switch it is. I'm going to need something a little more substantial in the long run but for now it's fine. take care
|
|
pagan
Been here a while!
Posts: 512
|
Post by pagan on Nov 13, 2012 6:45:33 GMT
Seems that pic was quite popular... reached limit on imagebucket (free-one) so here's the same pic on different host Uploaded with ImageShack.us
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 13, 2012 14:07:50 GMT
Hi All the 3V3 Tent shunt arrived this morning so pretty fast delivery on that. the down side is I'll need to wait until my birthday to try it out on the clocks. swings and roundabouts take care
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 13, 2012 16:49:22 GMT
Domenico has just told me by email that the XOs draw typical 2.5mA and max. 12mA.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 13, 2012 17:16:01 GMT
Hi Javier thanks for the above information which is useful to know especially for cricket and shunt users. any news your end take care
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 13, 2012 18:06:39 GMT
Not yet, I'll have to get very lucky to have it back this week, I guess most probably early next one.
|
|
|
Post by gommer on Nov 13, 2012 19:12:54 GMT
Aha Javier, could you also ask if the XO's disable the oscillator or only the output?
A datasheet would also be mighty handy. I'm very interested in duty cycle, rise and fall times and other timing data.
Few people realize or even know that the rise and fall times determine the real bandwidth of a signal, not the frequency of the signal itself. e.g. a 1ns rise time (not uncommon, some go down to 10's of ps) implies a 1GHz bandwidth).
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 13, 2012 19:18:56 GMT
I'll ask about the XOs disabling, he told me he has no "redistributable" datasheet (NDA? )
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 14, 2012 12:20:03 GMT
Hi All So after ‘’borrowing’’ a Salas from Will’s PK DAC to power the Amanero I decided that it might be a good time to try SYNC mode. I’m waiting for my Birthday pressy to replenish my supply of Salas Shunts. Anyway to the point ;D It was a pretty quick job to solder in a couple of extra wires to the Teleporters and five mins latter it was good to go. And Well the first thing that surprised me way how much the presentation had changed. I’m getting more air and sparkle on the top end which was nice but IMHO to my ears it did tend to make that part of the sound a little ‘’hot’’. Female vocals for instance took on a slightly more sibilant quality which I found just a bit too much at times. More on that latter. The mids seem to have more presence and ‘’texture’’ with more overall detail which helped to separate out instruments and gave a deeper impression to the soundstage. Bass seemed just a gnats drier to my ears but at the same time more detailed and tuneful and went down as low as I remember it did in ASYNC. Ok On the top end I tried the same OP filter arrangement as the one stated in the DATA sheet so 4.7nf Polly caps but no resistors. (Hope that makes sense) IMHO this seemed to reign in the top end a little without losing that air and sparkle. I just soldered in some turned turret sockets and plugged the caps in as Leo suggested in his build which made it pretty easy to try with and without. thanks Leo I have more of those little caps should anyone wish to hear for themselves. And just to be clear. The above is in no way intended to question or criticize Will’s PK or the design choices made. Let’s not forget that his implementation was primarily aimed at SPDIF input and it was only some (awkward) folk that wanted USB/I2S. Will it’s a great DAC and I’ve been more than happy with the SQ. (I’ll be on that little thing we talked about as soon as I have an extra Salas) IMHO the USB/I2S route does offer different choices as far as clocking goes and IMHO that is one of the key advantages with this method. One central controller (amanero/WAVEIO or whatever) dictating the whole process which works pretty well in this instance. That re-clocker that Marc and Javier have posted does interest me and having one through the Teleporters nice and close to the DAC interests me still further. but Just a little more thought needed on the implementation. Mmmmmmmmmmmm So all of the above IMHO, to my ears, in my system as usual. Thanks for the encouragement to try Javier, Chris, JK Take care
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 14, 2012 13:16:55 GMT
Hi Shaun, Could you remind me exactly the implementation on this to save me trawling through the thread. By strange coincidence I pulled a little bag containing 4.7nf polys from my "DAC-stuff" box just the other day, thinking "what the hell were they for" Cheers
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 14, 2012 15:27:35 GMT
Hi Chris here is Leo's implementation i273.photobucket.com/albums/jj220/kingbusoms/ESS%20dac/P1130937.jpgthanks Leo caps go in C12-13. i used some of those nice turned pins stripped out of a 8 pin DIP socket. solder them into the board and then the caps just plug in. a good way of experimenting with different cap values. then once you have a keeper UN-solder the pins and solder the caps in place. take care
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 15, 2012 14:48:50 GMT
Thanks Shaun!
I've had a little play since, I had several appropriate values of polypropylenes in stock so was able to try in steps.
My impressions are as follows, on my kit, with my ears and IMHO.
First in were some 4n7s, for me, straight away NO. I felt robbed of the very top end and couldn't really listen past the problem.
Next in were 2n2s, much better but still a little shy at the top end although there was an attractive fuller tone to acoustic instruments but less convinced of the effect on voices being ever so slightly "didgeridoo'd".
Last in, only because it was my last smaller value, was 1n5. The top snapped back into place and the voice effect above was gone. Very pleasant.
Now back to no caps. Well, a hard choice. Basically if I have a recording with a hot top then I'd like to switch these in. For recordings that are well balanced or maybe slightly dull then I'd like the caps switched out.
I'm guessing actually putting them on a switch would be counter productive!
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 15, 2012 18:37:55 GMT
Hi Chris interesting remarks. i must say that i did not get such a pronounced change when i put those caps in but each set up is different i guess I've since gone back to ASYNC with Will's PK as i felt that the cost outweighed the benefits SQ wise. in my system to my ears in SYNC mode sounded a little tooo dry for my taste and prefer the fuller flavor that i get from using the PK ASYNC. I'll wait and try again with SYNC once the new buffer for the PK arrives. the Amanero is sounding pretty good with the Buffalo but that's in ASYNC also so i may try to SYNC it at some point just to compare. on the Salas's after a gentle nudge from Alex i swapped out the Panny FC for Silmic II in the sense circuit which really helped add a little musicality to the sound. that was another surprise. take care
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 15, 2012 20:08:44 GMT
Shaun, the 100MHz Crystek CCDHD-950 on the BIII has much better phase noise performance than the Amanero's XOs and doesn't have the CPU added jitter so hardly a chance of better SQ in sync mode. Also the Amanero's XOs speed is much lower allowing a max SDM of 64x at 192KHz sampling rate (24.576 = 128fs@192, 22.579 = 128fs@176.4) which is somewhat low and certainly waaaay lower than what can be achieved with the local 100Mhz clock. I'll be very, very surprised if you prefered the BIII in sync mode.
|
|
|
Post by gommer on Nov 15, 2012 20:54:09 GMT
Javier, by any chance, do you know more about ASYNC mechanisms?
Simply resampling any signal with a 100MHz clock would introduce jitter as large as 10ns, so there must be more to it if a high spec DAC resamples.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 15, 2012 21:41:01 GMT
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 15, 2012 23:14:32 GMT
Shaun, the 100MHz Crystek CCDHD-950 on the BIII has much better phase noise performance than the Amanero's XOs and doesn't have the CPU added jitter so hardly a chance of better SQ in sync mode. Also the Amanero's XOs speed is much lower allowing a max SDM of 64x at 192KHz sampling rate (24.576 = 128fs@192, 22.579 = 128fs@176.4) which is somewhat low and certainly waaaay lower than what can be achieved with the local 100Mhz clock. I'll be very, very surprised if you prefered the BIII in sync mode. Hi Javier all good points that i totally agree with. take care
|
|
pagan
Been here a while!
Posts: 512
|
Post by pagan on Nov 16, 2012 1:21:40 GMT
there was a talk going on in a thread that was about using the b3 in sync mode, diyaudio or Ian's FIFO reclocker or other,didn't bookmark at time. Do a search on this, "usb i2s sync mode b3" comes up with relevent threads I think... also talks about coax for clock and/or i2s signals. 100Mhz clock on the b3 for the use of oversampling may not be as goood as sync mode and lower clock. I've yet to try it. another trade off for phase noise might be between shorter distance between clock and dac and better clock further away. allan ps typing on a tablet is a pita :-) hifiduino.wordpress.com/2012/08/21/ians-fifo-reclocker-and-buffalo-iii-dac/
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 16, 2012 7:20:11 GMT
100Mhz clock on the b3 for the use of oversampling may not be as goood as sync mode and lower clock. I've yet to try it. Alan, IIRC Ian's FIFO board with the programmable Si570 can do: 11.2896, 12.2880, 16.9344, 18.4320, 22.5792, 24.5760, 33.8688, 45.1584, 49.1520, 90.3168, and 98.3040 MHz With the last ones (90.xxx and 98.xxx) 512fs is achieved for 176.4 and 192 so a real SDM of 256fs (max SDM rate = 1/2 XO speed). Even at 352.8 and 384 those XOs have speed for good SDM though oversampling would be limited to 4x instead of the 8x for any thing less demanding. IMVHO, if one plays not only 16/44.1 but also Hi Res and/or DSD, the minimum XOs freq to get the ES9018 singing in sync mode would be 45.xxx and 49.xxx. If one only plays Std. Res. then even 22.xxx and 24.xxx should be OK.
|
|
pagan
Been here a while!
Posts: 512
|
Post by pagan on Nov 16, 2012 8:12:38 GMT
there was a talk going on in a thread that was about using the b3 in sync mode, diyaudio or Ian's FIFO reclocker or other,didn't bookmark at time. Do a search on this, "usb i2s sync mode b3" comes up with relevent threads I think... also talks about coax for clock and/or i2s signals. 100Mhz clock on the b3 for the use of oversampling may not be as good as sync mode and lower clock. I've yet to try it. another trade off for phase noise might be between shorter distance between clock and dac and better clock further away. allan ps typing on a tablet is a pita :-) hifiduino.wordpress.com/2012/08/21/ians-fifo-reclocker-and-buffalo-iii-dac/Ok I'm back Alex, thanks for attaching the link. The search I did was this one... www.google.com.au/#hl=en&output=search&sclient=psy-ab&q=usb+i2s+sync+mode+b3&oq=usb+i2s+sync+mode+b3&gs_l=hp.12..33i21l2.3075.3075.0.9525.1.1.0.0.0.0.252.252.2-1.1.0.les%3B..0.0...1c.3rlgB9jwWIs&pbx=1&bav=on.2,or.r_gc.r_pw.r_qf.&fp=77d97114e097c10e&bpcl=38625945&biw=1280&bih=607 It was either the "hifiduino" or "diyaudio" threads. As for trade-offs with phase noise... The distance between the clock and the dac can impact on jitter/phase noise as far as i've read. So using an off-borad super clock may not be as beneficial as a slightly less spec'd clock on-board. As for the B3's oversampling,,,, The theory is supposed to be that the oversampling is done at such a high frequency that the noise from it is way too high to be audible. I believe they were looking for oversampling harmonics down to 200Khz, not 20Khz as for most older sigma delta dacs. Interesting thought now..... Years ago the DAC clock was the master, and the input chain was sync'd of that. But that was CD Payers. Allan
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 16, 2012 10:34:47 GMT
there was a talk going on in a thread that was about using the b3 in sync mode, diyaudio or Ian's FIFO reclocker or other,didn't bookmark at time. Do a search on this, "usb i2s sync mode b3" comes up with relevent threads I think... also talks about coax for clock and/or i2s signals. 100Mhz clock on the b3 for the use of oversampling may not be as goood as sync mode and lower clock. I've yet to try it. another trade off for phase noise might be between shorter distance between clock and dac and better clock further away. allan ps typing on a tablet is a pita :-) hifiduino.wordpress.com/2012/08/21/ians-fifo-reclocker-and-buffalo-iii-dac/Hi Allan yes i've been trying to read through Ian's FIFO thread but it's a bit of a monster. as far as i can tell with the BIII in SYNC V's ASYNC the differences have subjectively been found to be presentational ones. people seem to be choosing on that basis. I'd have to say that the clock used on the BIII seems to be a very high quality unit so it's going to take some doing to better it. but without trying it's hard to know for sure. the Idea of re clocking close to the DAC really interests me mostly because I'm using Teleporters. theoretically the Teleporters introduce very little in the way of jitter but my natural instinct dictates that any jitter is too much. have you managed to find time to plug your Amanero in yet? I'm looking forward to hearing your thoughts on that. Alan whilst we are on the subject did you notice that TP have just released an updated AVCC module? it is supposed to be closer in design to the Trident 3. I've ordered one up but things usually take 3 weeks or so from TP.(birthday pressy from mother in law ) yup i get a little P'ed off with hifi elitism and try to involve the whole family in my hobby. ;D '' i don't know what to get you for your birthday'' ha don't worry on that i know just the thing I'll report back once i have it fitted. take care
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 16, 2012 10:57:15 GMT
Sure beats a couple of pairs of socks and crap catchers! ;D
|
|
|
Post by gommer on Nov 16, 2012 11:16:18 GMT
I haven't read threads on the subject other than in RG. But my knowledge on the subject tells me exactly the above. I do have to do my homework about upsampling techniques and how master clocks are exactly used. But what i do know about signal and power integrity and jitter is this: * More complex chips (more gates) will produce slower signals and more propagation delay jitter than 'discrete' logic. * Layout and cables for HF signals (think rise time) have to be very carefully designed, handling every signal as a trasmission line and taking care of HF return paths, avoiding antennae, impedance mismatches and the lot. If not taken care of properly, reflections and power plane issues can severely damage the transmitted signal. If clocks are to be very clean, then a very carefully connected coax cable at the source oscillator, reclocked right before the DAC and a DAC that does little or no processing with the master clock, might, or should lead to the best results (with a decent master oscillator off coarse). My 2c on the subject to add to the confusion and doubts Cheers, Marc
|
|
pagan
Been here a while!
Posts: 512
|
Post by pagan on Nov 16, 2012 12:28:03 GMT
I haven't read threads on the subject other than in RG. But my knowledge on the subject tells me exactly the above. I do have to do my homework about upsampling techniques and how master clocks are exactly used. But what i do know about signal and power integrity and jitter is this: * More complex chips (more gates) will produce slower signals and more propagation delay jitter than 'discrete' logic. * Layout and cables for HF signals (think rise time) have to be very carefully designed, handling every signal as a trasmission line and taking care of HF return paths, avoiding antennae, impedance mismatches and the lot. If not taken care of properly, reflections and power plane issues can severely damage the transmitted signal. If clocks are to be very clean, then a very carefully connected coax cable at the source oscillator, reclocked right before the DAC and a DAC that does little or no processing with the master clock, might, or should lead to the best results (with a decent master oscillator off coarse). My 2c on the subject to add to the confusion and doubts Cheers, Marc Marc My view on upsampling is. If the track wasn't recorded at that frequency/bit depth then where is the extra data coming from. Someone's mathematical formula. Use the data that's been recorded (Known) If you want a smoother sound, and a valve.. ;D:o As for adding extra components to clocks/reclocks... Any components can add or subtract something. Usually adds, very hard to subtract. It was surprising to me that the subject view was that they preferred the lower clocked (with extra cable), sync mode. Allan Ps "* More complex chips (more gates) will produce slower signals and more propagation delay jitter than 'discrete' logic." cpu clocking 100Mhz bus over 3Ghz in processor. On-dye things maybe easier, but then better jitter tolerance.
|
|