Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 31, 2012 23:51:12 GMT
Sure, but whether you like it or not your Oppo, SC DAC and X-DAC all convert to DSD as the last step before analogue. They all use Sigma Delta chips. Javier That is clear from the data sheets for the DSD179x series for example, as well as the name. Alex
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 1, 2012 1:22:40 GMT
Sure, but whether you like it or not your Oppo, SC DAC and X-DAC all convert to DSD as the last step before analogue. They all use Sigma Delta chips. EDIT: IIRC neither of your DACs supports 24/192 so why should you care? Although the MF X-DAC V3 can only do 24/96 natively , it upsamples everything to 24/192. I can still hear a clear advantage of an original 24/192 DL over the same DL in 24/96. Also, I do have intentions of changing the Input PCB in the SC DAC whenever a suitable upgrade candidate becomes available. It would need to have at least one Toslink Input and a Coax SPDIF input though.My SC DAC has already been compared against other DACs such as the W4s which can do 24/192 natively, and the Oppo95 with Sabre DAC, as well as a B3.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 1, 2012 2:17:22 GMT
Chris and Javier Initially I would try a +5V JLH supplying the onboard voltage regulator. It may also be possible to bypass the onboard voltage regulator and supply different areas of the PCB from 2 different 3.3V JLH or Salas/ PH regulators. Anybody constructing this should be more than capable of replacing the connector at the PC end of the USB cable with an A type USB plug where the red +5V Vbus wire is not connected.If using a JLH, a J506 or similar type CRD (about 1.5mA) should be used to replace the 10K resistor.You would need to check though that the PC recognises the card without continuity of the Vbus lead. Regards Alex
|
|
XTRProf
Fully Modded
Pssst ! Got any spare capacitors ?
Posts: 5,689
|
Post by XTRProf on Sept 1, 2012 5:16:54 GMT
Wah, after only a one night sleep so much had been written about this (positive and negative) already. I put my name down for this interface as it can do both stereo PCM up to 384-32 and DSD. To me DSD, although I would like to greedily have MC as well, is important to me as there are quite a lot of material that will never be available as hirez PCM. Take for example, those purely on DSD recording. Whether we can hear the harsh high frequency noise or not is really immaterial to me as the engineers had already filter them out well with the later DSD Dacs like Mytek. Anyway, how many speakers can do > 48khz? Even ribbon can only response to about 40khz due to it's mass. Also, most professional analog recorder like Studer can only do until < 40khz. Same for professional analog mics. Since we must use analog mics obviously to record before converting to digital, even professional digital recorder will not have a response pass 48khz. So all recorded material, past and present, will not have that high frequency problem as they will be filtered out without any loss in details being captured when using DSD. I will like DSD to be at DSD128 but will that help in any sense in present times recorded material? I don't think so even if the digital recorders can do it. So enjoy the music instead of speaking of all the technical challenges and going to war with it. To me I need DSD as well as PCM, just like Javier, as we have well recorded music in DSD. We will not let the DSD problem to stop us from enjoying the music in DSD when not available in hirez PCM. Btw, I agree with Javier that past music is better sounding musically then present and I have heard many. Anyway, this new interface, we are only talking of 39 Euro when the 60 mark is hit. So, like Javier had said, any BIG risk? Even if it don't work, we will have lost the most 50 Euro including postage. To many especially those in the DIY community, it's a risk worth taking as we will have the upside opportunity to learn more about the interface and have a better understanding of any future implementation. Anyway, I don't think I will like to go pass 384-32 as that will be purely academic already like I had wirtten last time. Can our hearing threshold even go pass 192-24 in the first place? Anway, it's a good interface to try for PCM even if the DSD cannot be implemented on my MF X-Dac V3 as that Dac doesn't have a USB interface. Must just find out where to tap the signals into the BB DSD1792 chip from the USB interface bypassing the BB 96-24 receiver and 192-24 upsampler chips. Yeah, I know that will be quite horrendous due to the small traces and I could destroy the X-Dac. But this is what DIY is all about. The difficult part and failure and finally the success . Also, I still can keep it for my future Sabre 3 or 4 dac when all things fail in the X-Dac.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 1, 2012 5:37:44 GMT
Chong FWIW, there is genuine musical content past 50kHZ in some of Barry Diament's recordings. Ribbon tweeters usually go out to around 40kHZ, so it could be worthwhile retrofitting them to an existing speaker system to take over where the existing tweeters plummet.Even Jaycar has 8 ohm ribbon tweeters for around $30 each. It wouldn't cost much to make a couple of 12dB Xovers for them either. www.jaycar.com.au/productView.asp?ID=CT2023There is also talk of a new LASER(!) microphone being developed with extremely wide frequency response. Barry has put his hand up to field test it when it becomes available. BTW, it wouldn't be a good idea to stuff around with the V3 trying to implement that, unless you have found a new well paying job. Something like Will's PK where you can further upgrade using better power supplies would be a far better idea. Alex
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Sept 1, 2012 9:12:44 GMT
What has interested me in this little pcb is that it enables me to try usb out from my pc for very little outlay.
The guy who designed it has used sensible solutions to the little niggles that could come up (the LC networks on the XO supplies) and has not tried to 'over-egg' the design, and just produce something that does the job in a sensible way.
As the circuit has been shown in public, it helps to understand how the supplies might help improve it - an almost definite dedicated supply for the XO's springs to mind, as does a clean 5V to the main pcb.
As for DSD/i2s or async or not, I don't care to much. I do now that I can connect this to my Buffalo32S or PK dac and music will come out. Nice.
Interestingly, for those with the SC Dac - one of the clocks on this little board runs at the frequency that the Dac uses on the SC Dac. As the SC Dac has the I2S signal some from the input board to the Dac/IV pcb via ribbon cable, it would be bodgeable to connect one of these USB devices to the SC Dac. Needs a bit more thinking out though.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 1, 2012 10:27:25 GMT
Interestingly, for those with the SC Dac - one of the clocks on this little board runs at the frequency that the Dac uses on the SC Dac. As the SC Dac has the I2S signal some from the input board to the Dac/IV pcb via ribbon cable, it would be bodgeable to connect one of these USB devices to the SC Dac. Needs a bit more thinking out though. You could even use both, the already present inputs plus the USB adapter, If you add one these: twistedpearaudio.com/control/otto.aspx
|
|
XTRProf
Fully Modded
Pssst ! Got any spare capacitors ?
Posts: 5,689
|
Post by XTRProf on Sept 2, 2012 0:17:11 GMT
Ah, that's what I need to switch between SPDIF receiver and this USB interface for the X-Dac. Have that in mind for some time to mode the X-Dac with USB interface, as highlighted to Alex some time back, but didn't think about where to get it yet as don't have any intention of moding the X-Dac now before this USB interface.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 2, 2012 0:49:30 GMT
Ah, that's what I need to switch between SPDIF receiver and this USB interface for the X-Dac. Have that in mind for some time to mode the X-Dac with USB interface, as highlighted to Alex some time back, but didn't think about where to get it yet as don't have any intention of moding the X-Dac now before this USB interface. Chong Javier would know more about this than I do, but you would still be limited to 24/96 from USB, and there is also the existing firmware to think about ? You would also need to use a JLH or similar for it's USB power, and modify the USB lead, as the X-DAC V3, unlike the recent models uses high quality coax SPDIF transformers and is far lower Jitter than originally if the TCXO and a few other simple component changes are made. Have you investigated fitting a .3PPM Vanguard TCXO to replace the mediocre 24.576MHZ Xtal oscillator ? A couple of people have already done that, and also used a "Cricket" to supply the TCXO with VERY clean and stabilised +5V. The "Cricket" could be put together on veroboard if need be. What I am suggesting here is to use the USB-i2S converter with something more amenable, such as Will's PK, and fit them both in a much larger case to permit all the additional PSUs to make them really "sing" The X-DAC V3 is too cramped internally to do too much more with it unless you fit it in a rack case with a toroidal transformer and regulated PSUs and JLH as david2vk from Sydney has done. Alex Attachments:
|
|
XTRProf
Fully Modded
Pssst ! Got any spare capacitors ?
Posts: 5,689
|
Post by XTRProf on Sept 2, 2012 2:02:42 GMT
Javier would know more about this than I do, but you would still be limited to 24/96 from USB, and there is also the existing firmware to think about ? You would also need to use a JLH or similar for it's USB power, and modify the USB lead, as the X-DAC V3, unlike the recent models uses high quality coax SPDIF transformers and is far lower Jitter than originally if the TCXO and a few other simple component changes are made. Sorry, I don't understand what you mean by the USB limited to 96-24. Isn't the USB interface can output up to 384-24 and also DSD? I mean by pass the 96-24 SPDIF reciever and direct into the DSD1792 Dac chip so that we can have SPDIF as well as the USB with the switch that Javier pointed us to. Also, we can change the receiver to 192-24 to upgrade the X-Dac V3 to 192-24 direct. The 192-24 upsampler chip will not do anything when the output is already at 192-24. Have you investigated fitting a .3PPM Vanguard TCXO to replace the mediocre 24.576MHZ Xtal oscillator ? A couple of people have already done that, and also used a "Cricket" to supply the TCXO with VERY clean and stabilised +5V. The "Cricket" could be put together on veroboard if need be. In fact, I have and was thinking about the Tentlabs clock way back (when I started my MF X-Dac V3 mode thread here in RG) when I wrote that the clock would improve the Dac sound and also the superregulators would as well but you were writing then that the clock would make not difference. I didn't do that as I don't have much time to think about implementing it as it would involve a lot of time to research and reading up. Also, I had planned to move the rectifiers outside of the X-Dac into another PS box and put in superregulators of this form factor, as shown in the link, into the X-Dac to replace all those 78XX and 79XX. So cramp X-Dac will have no problem for super power. www.partsconnexion.com/regulator_belleson_spj78.htmlYeah, 2A, ok? What I am suggesting here is to use the USB-i2S converter with something more amenable, such as Will's PK, and fit them both in a much larger case to permit all the additional PSUs to make them really "sing" The X-DAC V3 is too cramped internally to do too much more with it unless you fit it in a rack case with a toroidal transformer and regulated PSUs and JLH as david2vk from Sydney has done. Yes, you are right there. But then I will have to make another Dac which will create another unit in the house that family members will not be too happy about as now I already use up a lot of territory in the house for hifi. Since I'm interested to make the X-Dac V3 to current MF Dac specs with USB, upgrading the X-Dac V3 will be a better proposition until I get the Sabre Dac. But by then, I will like to have MC for the Sabre which this current USB intereface will most probably cannot do.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 2, 2012 2:28:13 GMT
Chong I think you will find it isn't that simple with the X-DAC V3. l will leave that to Javier to comment about.In the meantime you could read the DSD1792A datasheet where you will find it isn't so easy to implement both PCM and DSD . Fitting a Tent Labs oscillator originally would heve likely been a waste off time unless you upgraded the whole PSU area. Even the analogue area of the original cct. had NO voltage regulation.Fitting close tolerance polyprops to the filter pins of the DIR9001 would need to be done too, to reduce Jitter. A TCXO or Guido Tent oscillator on it's own can't work miracles. The super regs still need to be close to the area to be supplied. I had to fit the JLH externally for space reasons, which means you still need to keep the internal VREGs , which then have little work to do but pass clean DC. If it was all in a larger case the Super Regs or JLH could be mounted quite close to the areas to be supplied and the existing VREGs bypassed. Alex
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 2, 2012 9:23:22 GMT
Chong, I'm afraid Alex has a point there. Modding the X-DAC can be done but it will be far from easy and implies some serious risks for your X-DAC. To implement the OTTO 2:1 switch you will need to cut the traces from the TI4192 I2S output connect it to one of the inputs of the OTTO and then the OTTO's output to the DSD1792 input that should be rather straight forward if you have the proper skills but from that point onwards it becomes far, far more complicated. The X-DAC's circuit is designed for PCM only and to make it work with DSD a few important and challenging changes would need to be made. The DSD1792 belongs to an older generation and uses different input pins for PCM and DSD plus you need to signal the chip if there is a change in input type, then there is clocks/sync, etc. The Amanero board seems have provision for Reset, Mute and DSD/PCM signaling but making any use of them is waaaay beyond me. Newer chips like the ESS9018 or the AK4399 share the same pins for PCM and DSD and can autodetct incomming signal but with older DACs is not such an "easy" task. Someone with very deep understanding of these matters should see the schematics to asses whether this mods are even feasible or not and if they are, how difficult they'll be. Another limitation is that according to its spec sheet the DSD1792 stops at 216KHz PCM so forget anything over 24/192, still, you get to have 176.4 and 192 which you can't do now. The only chips I know that can do PCM higher than 192 are the highest end Sabre 32s (9018, 9016 and 9012, not sure about the 9006 and 9008) and the latest from the TI stable, the PCM1795 - which can also do DSD - and Musiland found out that although undocumented the PCM1798 will also reach 352.8/384. A very interesting project could be a DAC based on the DSD1793 which is a full Vout chip like the ESS9022/23 therefore no need to design a I/V stage, just a simple opamp based LPF. It is still limited in PCM mode to 24/192 but takes up to DSD256. What do you reckon Will? Changing the SPDIF receiver on top of adding the USB input would a little too ambitious a project IMVHO. About clocks, there is a lot of talk about the Crystek CCD-HD957 in many threads at DIY, it looks to have tremendous performance at a very reasonable price (26 USD at Mouser) and they come in audio freqs. Some talk about it with graphs and figures at Hifiduino's blog: hifiduino.wordpress.com/2012/08/27/crystek-cchd-957/Plus, a nice comparisson table of the latest breed of regulators there too (including the latest TI's TPS7A4700 which he says is great for audio-> hifiduino.wordpress.com/2012/08/29/new-ti-tps7a4700/) ): hifiduino.wordpress.com/2011/03/29/new-breed-of-ultra-low-noise-regulators/and 7805 talked here: hifiduino.blogspot.com.es/2010/03/comparing-noise-figures-in-linear.html
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 2, 2012 10:09:22 GMT
Javier I am not sure if going even lower with Jitter will result in further audible improvements, but I can say that Sydney member Geoff , who has an otherwise virtually identical modified X-DAC V3 to mine, replaced the standard Xtal Oscillator in his with the recovered 1PPM Vanguard TCXO from mine, as well as the added "Cricket" for the +5V PSU, and reported a quite worthwhile improvement. However, when compared side by side with mine using the .3PPM . Vanguard TCXO, mine still had a clear edge with more 3D sounding on high res material.It wasn't such an obvious difference with 16/44.1 though. The X-DAC V3 does however use a 5V Xtal Oscillator. Alex
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 2, 2012 23:26:12 GMT
The ENOB of ESS DAC is around 19 bits, perhaps with the most optimal PCB design it may even reach 20 bits. This is true for most DAC's by the way. You can view the 32 bits 'handling' is more in the line of it can 'accept' the format. I am convinced there will never be a DAC reaching an ENOB of 24bit. a matter of physics I am afraid. Certainly not 24bit but the new ESS9018 based Exadevices e20 (2,499 USD) claims over 21bit ENOB at 128dB SNR (A-Weighted, 2Vrms) which, if true, isn't too bad. THD+N is 0.0002% -> 113.98dB so barelly 19bit ENOB: www.exasound.com/e20DAC/e20DACSpecifications.aspxThey provide some measurements too: www.exasound.com/e20DAC/Measurements.aspx
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 3, 2012 6:52:45 GMT
A very good PCB design (the ESS series MUST have a properly designed PCB to reach those values) can get the max out of the chip. Cheaply designed Chinese e-bay PCB's will never reach those levels though (nor is it needed in practice). I suspect the analog filter section behind the ESS is not just a simple RC filter either. Also I think they 'paralleled' 4 channels to increase performance, otherwise why would one choose the 9018 over the 9012. Note that jitter isn't a problem anymore in this e20DAC and other better NEWER designs (nor needs not be because of jitter reduction circuits in the ESS chips) Now we need headphones, microphones and amplifiers that reach the values these high-priced DAC's reach. Even the best headphones/microphones resonate, swing-out and distort (non-linear as well as I.M.) thousands of times more. And that is NO exaggeration. ENOB's are rarely published for obvious reasons. A manufacturer that markets a 32 bit DAC won't like to mention it can't resolve more than 20 to 21 bits in the end so leave it out of the spec sheet. Bad for sales so to speak, they only like to give impressive numbers that appear to tell something about the SQ but in reality is more a marketting strategy. Playback is only as good as the recording (the final product I mean). It's at THAT point most degradation occurs and cannot be undone.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 3, 2012 7:27:20 GMT
Frans What kind of standalone DAC are you using, or are you still using a soundcard ? Are you speaking here from practical experience with DACs, or basing what you say on theory ? Both Javier and myself are suggesting that very high stability, low phase noise Xtal oscillators are desirable, and are likely to result in improved performance provided that they are well integrated into the design. Are you suggesting that this is "snake oil" and things are already as good as it gets in most commercial DAC designs? Regards Alex
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 3, 2012 16:30:45 GMT
A fellow Spaniard has already received his Amanero board (he ordered it before the DIYAudio GB) and is quite surprised about the excelent build quality. He has also purchased and received at the same time a WaveIO so he'll be doing some comparisons connecting both to a "highly modded" Buffalo III. He says the WaveIO is twice the size of the Amanero board and build quality is also better even if hand built.
He ordered the Amanero without the LDO so he can feed it with a very high quality 3.3V PS.
He has promised to keep me up to date with his findings.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 3, 2012 16:51:32 GMT
Javier, Just for the sake of clarity, not that it matters much, but are you saying that the build quality of the WaveIO is better than the Armanero even though that of the Armanero is "very good"? Dave.
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Sept 3, 2012 18:38:41 GMT
A fellow Spaniard has already received his Amanero board (he ordered it before the DIYAudio GB) and is quite surprised about the excelent build quality. He has also purchased and received at the same time a WaveIO so he'll be doing some comparisons connecting both to a "highly modded" Buffalo III. He says the WaveIO is twice the size of the Amanero board and build quality is also better even if hand built. He ordered the Amanero without the LDO so he can feed it with a very high quality 3.3V PS. He has promised to keep me up to date with his findings. Looking forward to hearing what your mate says, Javier. Would it be Merlin ? It appears that I'm in the first batch, thanks to you giving us the heads up. I'm looking forward to comparing it to the spdif out from my Asus soundcard and in the long term, seeing how far I can push both. Something that I have pootling in the background is a SMD reg/JLH combo which should be quite useful for this, as I understand it draws 130mA total, so less per individual section.
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Sept 3, 2012 21:38:38 GMT
Think that this might help for the future... at the bottom of the picture is a rectangle (P3) - two circles and a square. One of them (the square, I think) is an external feed in for 3v3 volts, which is linked in to the on-board 3v3 line, currently generated by USB 5V via an adp151 reg. Removing the adp151 (to the left of the P3 rectangle, has caps C27 and C2 near it) or simply removing inductor L1 will isolate the circuit from the USB 5V, allowing you to feed in 3v3 directly. The benefit of doing this is that the nice clean supply you feed in will not get 'noisy' after passing back through the adp regulator. Having said this - You will get good results from simply using clean 5V into the usb socket over that supplied by the PC, without having to isolate stuff. I've only posted that for those that fancy going a little further.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 3, 2012 21:49:32 GMT
Javier, Just for the sake of clarity, not that it matters much, but are you saying that the build quality of the WaveIO is better than the Armanero even though that of the Armanero is "very good"? Dave. Yep, he told me the WaveIO looks like a 4 layer board with thicker tracks and better quality parts even if they are hand assembled one by one by Lucien. The Amanero looks like a thinner 2 layer with finer tracks but also he says it looks more professional/industrial build, parts being very good though not to the same standadrd as the WaveIO. I emailed Lucien recently asking him about the possibility of implementing DSD, he answered that he is first going to try 32/384 (the XMOS chipset cannot do 32/352.8) and if it works he'll try DSD. The problem is he is not Speedy Gonzalez when it comes to making changes to the WaveIO, he never has enough free time. Looking forward to hearing what your mate says, Javier. Would it be Merlin ? It appears that I'm in the first batch, thanks to you giving us the heads up. I'm looking forward to comparing it to the spdif out from my Asus soundcard and in the long term, seeing how far I can push both. Something that I have pootling in the background is a SMD reg/JLH combo which should be quite useful for this, as I understand it draws 130mA total, so less per individual section. As always, spot on Will, that was indeed Felipe (aka Merlin), a very nice chap. I've spoken to him a few times over the last days. He is extremely picky about PSs and has serious plans for both boards, it'll be interesting to learn about his findings though I'm afraid he is in the money no object camp when it comes to regs and that kind of stuff.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 3, 2012 22:16:02 GMT
Probably best to have the CL section at a couple of hundred mA more if the SMD reg can cope with that at startup . You may be able to use some of the Zetex devices for that ? Incidentally, for those wishing to marry this PCB with Will's PK, it may be worthwhile fitting it all in a 1U rack case with a couple of 6-0 and 6-0VAC 30VA toroidal transformers. This would give the possibility of at least 4 separate +5V or +3.3V supplies , or a combo of both voltages using JLH /PH/Salas regulators, assuming that LDO regulators are used. Alex
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 5, 2012 16:29:42 GMT
Size comparisson of WaveIO and Amanero boards (low PQ, hope to get a better pic soon):
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 5, 2012 17:07:25 GMT
BTW, have all who joined the GB sent their emails and payed?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 5, 2012 21:48:37 GMT
Hi Javier
I added my name to this GB but haven't been back to check progress. I'm on holiday at the minute with just a net book, no copies of home emails.
Would you please send me a link as I can't find the right place on DIYaudio.
Thanks
Syd
|
|