|
Post by jeffc on Feb 20, 2010 9:49:48 GMT
And another update As mentioned, a few days ago I swapping out the plastic LME49710s I'd used in the I/V stage for the kit OPA134s which brought immediately improved SQ.... better overall clarity, vocals smoother, better 3D and soundstage depth etc... these impressions with them green and after a few days burn in it is obvious that transparency and drive are also much better. This with the metal can LME49710HA opamps I had fitted up front at the output in the expectation they would better the plastics ... as Alex has now confirmed. As I am now quite in tune with the sound, before I fit AK-crickets, I thought I'd test the less severe output filter setup used in the XDAC-V3 as suggested by Alex in Post#83 and tested by Alfred with favourable comments. Well well .... this elevates what I thought was a darn good sounding DAC into something I'm sure wouldn't be embarrassed by any of the big guns. The changed filter opens up the sound beautifully, soundstage expands in all directions and it simply sounds less constricted and more real. Quite a leap up in overall SQ and enjoyment. Soldering iron still warm Alex, get it fired up, I think your XDAC-V3 will have a serious contender with the same filtering scheme in place. ;D Now getting very excited about just how nice this DAC is going to sound when AK-Crickets and CVXO are fitted. More updates soon. cheers.. jeffc
|
|
|
Post by jeffc on Feb 21, 2010 11:25:43 GMT
AK-Crickets P-to- Guys, been a bit slow getting these in use as I really didn’t have the right zener diode type. Alex fixed that with more generosity in the way of a package of BZX55C-5V1 and BZX55C-3V3 zeners that arrived during the week. Plus other bits and pieces he’d sent earlier. Anyway, here’s some pics of my P-to-P twin baby AK-Crickets on some cheapo small vero boards from eBay that came as a pack of 10 for $8AUS delivered and have proved the ducks guts for building these little voltage stabilizers. AK-Cricket top side: V and gnd wires to the central pins will supply both Crickets AK-Cricket bottom side: robust wire used and some points wire is looped over, although its hard to see this at the happy snap angle And for those anal ones reading resistor codes, yes I buggered this one up by placing the 10K and 2.2/2.7K resistors arse about Caps used with a 10uf Wima PP with 5 mm wire spacing a 4.7uf BG-NX and 10uf Oscon for the low ESR output. Used because I had them. Best caps for this application.. not sure.. but hopefully they’ll work well. Opamp is a LME49710 because Alex generously sent me a stash of them. And tips for young players like me. Tested one 3V3 zener with a 15.12V supply, V out 3.45V @2.0K, 3.41V @ 2.2K and 3.32V @ 2.7K. So I built a pair of AK-Crickets and V out on one 3.32V, 3.20V on the other. Swap out the 2.7K for a 2.2K on the 3.20V Cricket and V out now 3.28V. Lesson, don’t trust these little buggers to all test exactly the same and match up zener diode and resistors pairs before building. OK swapping out resistors on an etched PCB, but P-to-P is more a PITA due to the extra solder points and can stuff up what you thought was OK workmanship. Now will I ever get to testing them. ;D cheers.. jeffc
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Feb 23, 2010 12:46:28 GMT
Top stuff, Jeff, keep us informed on how it's going. Those little vero boards are excellent, and spot on for the job, by the looks of it. Really need to pull my finger out on this one....
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Feb 24, 2010 5:01:13 GMT
Wider Bandwidth Filter, as per X-DAC V3As Jeff has verified, the less severe rolloff of the simplified filter stages results in an improved, more natural soundstage. I have also modified my SC DAC to this schematic as discussed early in the thread, with similar results. The new component values are shown alongside the components. In addition to the new 1% MF resistors, 4 x 2n2 polypropylene capacitors are recommended for the original 2n7 locations. The 2n7 polypropylene capacitors are now relocated to where the 8n2 capacitors were. If not too expensive, I recommend LME49710HA metal can opamps for the 2 output devices. The improvement over the plastic LME49710 is quite noticeable. Alex URL=http://img641.imageshack.us/i/scdacoutputpcbmodified.jpg/] [/URL]
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Feb 26, 2010 1:06:39 GMT
I had a listen to this DAC yesterday and this morning through the main system while the house was quiet for a few minutes, and I was blown away by the purity of Renee Olstead's voice on a few .wav files that I had downloaded and burned to a Kodak Gold CD-R. I would recommend that anyone starting with the kits go straight to the simplified filter circuit as in the DSD1792 data sheet as posted here. This modified DAC is good enough already to reveal some differences between files, despite them being degraded due to a few trips around the globe. Hopefully, a further report from JeffC shortly. Alex
|
|
|
Post by jeffc on Feb 27, 2010 8:04:17 GMT
Crickets chirping, must be rain on the way. Hooked up a CVXO at last, had to mount it on the dual 3.3V Cricket board as the DIP8 socket pins were too wide for the Xtal holes. Still, I had space for it, just, no holes spare. I connected the CVXO Pin4 and Pin5 via wires to the holes of the 33 pf ceramic cap that was removed, and which linked the input and ground of the original Xtal that was removed. Link wire to ground across the top 33 pf ceramic cap holes was then not needed. I left the 100R on the CVXO input to the DIR9001 in place. The second 3.3V Cricket on the dual board was connected to Pin 1, gnd wire to Pin 2, of the ribbon cable supplying the DAC chip 3.3V Vdd. The link wire to the 3.3V Reg 4, which also supplies the DIR9001 3.3V inputs, was removed. No smoke on power up. So how does the SC-DAC sound now, everything is simply better. The problem is now I want to play music louder as its 'oh so clean, stable, detailed, textured and smooth' but my little Charlize T2020 T-amp has maxed out of headroom to drive my NXT panels. I need more power. ;D With each mod, this DAC is turning into a serious BIG GUN challenger, no joking. As Alfred coined previously, these AK-Cricket supplies are indeed "Little Miracle Workers". I was very tempted to connect up the other dual 3.3V/5V Cricket that I now have working, and be gone with both Reg4 and Reg5, but you know who would have .,,,, ;D. I will soon. cheers..jeffc
|
|
leo
Been here a while!
Team wtf is it?
Posts: 3,638
|
Post by leo on Feb 27, 2010 15:19:24 GMT
I'm after a dac for the PC, hmmm should I or shouldn't I It would be interesting to see how this with the cheap mods compare against the modded Buffalo32s with PH regs What I'd do is try it with the standard dac chip and then try the 1792 to see if it improves sonically, SMD is no problem for me
|
|
leo
Been here a while!
Team wtf is it?
Posts: 3,638
|
Post by leo on Feb 27, 2010 15:26:38 GMT
BTW did anybody try LT1028's with its compensation caps in the differential to line out as in the 1792 drawing?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Feb 27, 2010 21:10:57 GMT
Leo Not that I am aware of . It seemed rather pointless when we already have 6 OPA134 singles in the kit. I also had some LME49710 from a Aus/NZ GB in DIYAudio where we got them for <AU$2 each ! I had enough spare to send a few to Jeff, although now he has the 2 x LME49710HA, he was able to use 2 of those in the "cricket". The kit comes with decent DIL sockets for the DAC PCB so you can play around.The beauty of this particular kit is that it is so easy to work on because of the open design with separate PCBs Alex
|
|
leo
Been here a while!
Team wtf is it?
Posts: 3,638
|
Post by leo on Feb 27, 2010 23:54:36 GMT
I wouldn't want too many opa134's in there and rather keep them in the first stage only, if there anything like the 132 using too many gives a dark signature. Reason I asked about the LT1028's is because I already have a couple , I don't have any of the LME's
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Feb 28, 2010 0:16:17 GMT
I wouldn't want too many opa134's in there and rather keep them in the first stage only, if there anything like the 132 using too many gives a dark signature. Reason I asked about the LT1028's is because I already have a couple , I don't have any of the LME's Leo Which is exactly why I chose to use the LME49710 at the output right from the start.I even had them ordered before the kit became available. I didn't even bother trying to use the whole 6x OPA134 as supplied, for that reason. Otherwise you end up with a typical "house"sound. Alex
|
|
|
Post by jeffc on Feb 28, 2010 2:38:21 GMT
Leo, Once Spirit gets his Buffalo DAC with PH regs firing and he's happy with it, I'll have him over so we can compare the DACs and take notes. Although Alex's encouragement was part of my choice to build one, the other was the DSD1796 DAC is used in a friends modded Denon 3910 multiplayer that I'll say again, betters another mates top of the range Meridian unsampling CDP, and is the the most analogue sounding digital source I've heard. So.. for me, the SC-DAC has provided a means of trying to replicate this sound, for relative cheap, and relative ease with the receiver and DAC PCBs separarted, which provides lots of real estate for add-ons. With the "AK-Cricket" supplies left on overnight, I am even more pleased with the sonic improvement today. Getting to the point of microscopic insight into the recording event, just fantastic. A step up in SQ from the el' Cheapo CS4397 DAC I'd pretty much modded to the max. cheers...jeffc PS. Will, sorry to eat into your AK-Cricket PCB market, but I figured they'd be easy enough to build P-to-P, and if I'd taken more care with which resistor went where, and where the CVXO would fit best, I might have even done a reasonably neat job of them.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Feb 28, 2010 3:24:09 GMT
Jeff I wish that Will was making them. It seems like there may be plenty of places they could be used where a very low noise DC supply up to about 30mA is required . Even as high as 9 or 12V with the appropiate Zener diode and a 15V supply?
Alex
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Feb 28, 2010 9:34:31 GMT
I wish I had a cheap and plentiful way of getting them made as well! It would certainly help things this end I must admit I don't want to etch and supply myself, but if there was enough interest from the people here, I'd be happy to investigate getting some made (either the PSU board and/or the osc board) I'm not in this to make money, so they would be supplied at cost. Jeff, no probs with your vero board supplies The circuit is simple enough to make on veroboard, nice and quick. Has anyone got any thoughts on a discrete IV for this dac? I know I've still got to power mine up but it something I'd like to investigate, along with swapping out the 1796 for the 1792 I have sitting here (along with it's greater current output)
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Feb 28, 2010 10:08:44 GMT
I wish I had a cheap and plentiful way of getting them made as well! It would certainly help things this end I must admit I don't want to etch and supply myself, but if there was enough interest from the people here, I'd be happy to investigate getting some made (either the PSU board and/or the osc board) I'm not in this to make money, so they would be supplied at cost. Jeff, no probs with your vero board supplies The circuit is simple enough to make on veroboard, nice and quick. Has anyone got any thoughts on a discrete IV for this dac? I know I've still got to power mine up but it something I'd like to investigate, along with swapping out the 1796 for the 1792 I have sitting here (along with it's greater current output) Hi Will I was just stirring ! We would need to get a few more people to prove their worth before we even thought of something like that. I have just printed out a transparency to use for making some of those, but I hope that I don't stuff this lot like the previous lot. Sometimes I get fabulous results with the photo method, but other times I seem to lose part of the pattern before the rest of it has fully developed . Regarding the discrete IV stage, I would suggest that you try what Jeff and I have now before going there. The LME49710HA really is superb at that location. In fact, I have wondered how a pair of those on a browndog adaptor would go with the SC HA with JLH. Alex
|
|
leo
Been here a while!
Team wtf is it?
Posts: 3,638
|
Post by leo on Feb 28, 2010 10:25:18 GMT
Leo, Once Spirit gets his Buffalo DAC with PH regs firing and he's happy with it, I'll have him over so we can compare the DACs and take notes. Although Alex's encouragement was part of my choice to build one, the other was the DSD1796 DAC is used in a friends modded Denon 3910 multiplayer that I'll say again, betters another mates top of the range Meridian unsampling CDP, and is the the most analogue sounding digital source I've heard. So.. for me, the SC-DAC has provided a means of trying to replicate this sound, for relative cheap, and relative ease with the receiver and DAC PCBs separarted, which provides lots of real estate for add-ons. With the "AK-Cricket" supplies left on overnight, I am even more pleased with the sonic improvement today. Getting to the point of microscopic insight into the recording event, just fantastic. A step up in SQ from the el' Cheapo CS4397 DAC I'd pretty much modded to the max. cheers...jeffc PS. Will, sorry to eat into your AK-Cricket PCB market, but I figured they'd be easy enough to build P-to-P, and if I'd taken more care with which resistor went where, and where the CVXO would fit best, I might have even done a reasonably neat job of them. Thanks Jeff, that sounds quite promising. Its taken a lot of work tbh to get this Buffalo32s to where its at now, I wasn't happy with the stock version, because theres no schematics I've no idea how a BuffaloII would compare with or without the PH regs , theres obviously other things to tweak about to apart from just the regulation I guess what I'd find interesting is how the SC can be made to compare on the cheap, it should be more of a tempting project for more people Leo
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Feb 28, 2010 18:27:18 GMT
Alex, you may have been stirring, but it'd still be nice to get a batch of boards made, as they have a lot of uses external to this dac as well. I could use a set on the sc dac, plus some more elsewhere.
I will be trying the modified output stage certainly, it's just that I'd like to try out a discrete stage with it as well. I was wondering if the more experienced here had any pointers in something that might be worth a punt, for fun if nothing else.
As for etching using the photo method, I had similar problems, and ended up diluting the developer so that it wasn't so aggressive, and using a brush to do a bit og the board at a time. Of course, it could be a bad batch of pcb's also. BTW, not trying to teach you to suck eggs here, just what I found when I had problems.
|
|
leo
Been here a while!
Team wtf is it?
Posts: 3,638
|
Post by leo on Feb 28, 2010 19:34:04 GMT
Hi Will, I know I wouldn't mind a few of the boards if theres enough interest, I can think of lots of uses for those
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Feb 28, 2010 20:21:19 GMT
Alex, you may have been stirring, but it'd still be nice to get a batch of boards made, as they have a lot of uses external to this dac as well. I could use a set on the sc dac, plus some more elsewhere. I will be trying the modified output stage certainly, it's just that I'd like to try out a discrete stage with it as well. I was wondering if the more experienced here had any pointers in something that might be worth a punt, for fun if nothing else. As for etching using the photo method, I had similar problems, and ended up diluting the developer so that it wasn't so aggressive, and using a brush to do a bit og the board at a time. Of course, it could be a bad batch of pcb's also. BTW, not trying to teach you to suck eggs here, just what I found when I had problems. Will I had great results originally with many PCBs, then a string of failures. It may have also been a batch of crook developer . My new lot of developer has come from a specialist supplier, so the results may be very different. Before we race in with PCB manufacture, I would prefer to see at least confirmation from you that the boards are working as well as a couple of reports indicate.(I would be awfully embarrassed if my circuit 's benefits were more of a placebo effect!) In that case, I would certainly be interested in more also. I also liked Jeff's idea of a double PCB with a common 2 terminal power input connector. I was playing around in Photoshop last night to see if it was readily done with the existing singles design. Alex
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Feb 28, 2010 21:06:19 GMT
Alex, No race as such, but yes, I'd want to etch it and trial it first, along the same lines as Jeff and yourself has done. I've a few boards building up that need etching, so will do a couple of crickets My projects are building up big time at the moment, really need to win the lottery so that I don't need to goto work anymore. That'd give me plenty of tattin' time
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Feb 28, 2010 21:17:35 GMT
Alex, No race as such, but yes, I'd want to etch it and trial it first, along the same lines as Jeff and yourself has done. I've a few boards building up that need etching, so will do a couple of crickets My projects are building up big time at the moment, really need to win the lottery so that I don't need to goto work anymore. That'd give me plenty of tattin' time Will Although I came up with this variation, so far only Jeff and Alfred have actually tried it.We need more confirmation before Leo builds a batch of them . Alex
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 2, 2010 4:29:40 GMT
Update.I now have a 3.3V "Cricket" supplying the DAC PCB via the connecting cable from the Input PCB. The link to the existing 3.3V feed has been removed, so it only feeds 3.3V to the DAC PCB. There has been quite a large increase in sound stage and resolution.DTV audio via Toslink is now noticeably improved in the soundstage and ambience areas. I also tried playing a CD from both the Oppo and Pioneer DVD-A/SACD players using the same coax cable into the SPDIF Input of the DAC. I am absolutely amazed at the HUGE difference in sound between both players when used solely as transports. It is more like the difference between Analogue Out, between a good player and a mediocre player, which you would expect. I would be amazed if both of the players were not outputting the exact same stream of digital 1s and 0s.
|
|
|
Post by r1a2s3t4o5 on Mar 3, 2010 15:11:09 GMT
Update.I now have a 3.3V "Cricket" supplying the DAC PCB via the connecting cable from the Input PCB. The link to the existing 3.3V feed has been removed, so it only feeds 3.3V to the DAC PCB. There has been quite a large increase in sound stage and resolution.DTV audio via Toslink is now noticeably improved in the soundstage and ambience areas. I also tried playing a CD from both the Oppo and Pioneer DVD-A/SACD players using the same coax cable into the SPDIF Input of the DAC. I am absolutely amazed at the HUGE difference in sound between both players when used solely as transports. It is more like the difference between Analogue Out, between a good player and a mediocre player, which you would expect. I would be amazed if both of the players were not outputting the exact same stream of digital 1s and 0s. After reading this thread, You've made these changes to the DAC: - replacing crystal with crystal oscilator with voltage regulator on separate board - changing output I/V caps and resistors - changing filter and replacing output OPA134 with canned LME49710HA - providing 3.3V to DSD1796 from input board? jeffc wrote, that he build AK-Crickets P-to-P voltage regulators. Where are these cricket voltage regulators used? +-12V ? Have You considered separate supply for analog circuit and digital circuit on output board? joining grounds through let's say 100R resistor? Problem I have with this DAC, that when no audio is playing I get 285kHz sine on output: which of course is not played by my amplifier. I'm just curious whether You get it as well. I have not done any changes to the original circuit.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 3, 2010 20:43:43 GMT
r1a2s3t4o5 Re the Xtal oscillator. I am unsure as to why it should make an improvement, as it is only used as a reference in this case. However, 2 other members have also reported improvements when using it. The filter circuit used is now as per the DSD1792 schematic, which is virtually as used in my MF X-DAC V3. We all found the original design to sound very lacklustre, with very little soundstage on good material. The +-15V supply for the DAC PCB is via a John Linsley Hood PSU Add on, which is very low noise , and has a low output impedance to past 300KHZ.This appears to markedly improve channel separation too. The "Crickets" are also supplied from this very clean +15V. Unlike a couple of other members, so far, I have only used the Cricket to supply the 3.3V to the DAC via a link removed near pin1 of the Input PCB and through the IDP cable as before. I have not played around with the earthing arrangements of the DAC PCB, although one member has done so with a reported small improvement.I am more conservative, and like to fully evaluate each change,and not do too many at a time . However, the SC DAC is now at least as good as my highly modified MF X-DAC V3 which clearly outperformed a friend's new Benchmark USB DAC. I would be most surprised if Silicon Chip's prototypes had such a high level of 285kHZ at rest as those guys are so much into their test instruments . Alex P.S. You will notice that I lived with the original version of the circuit, but with polypropylene capacitors in key areas, for quite a while before even trying the simplified filter from the DSD1792 data sheets. Why not send Silicon Chip an email, and include the photo? I noticed that the author of this project, Nicholas Vinen, now works for Silicon Chip, and they have now published a VERY low distortion Audio Oscillator to 48kHZ that uses a DAC. P.P.S. At maximum sensitivity on my old 6MHZ CRO , I only get a couple of divisions of what appears to be wideband noise with Toslink 1 cable plugged in,a little less from Toslink2 with no cable plugged in, and close to full scale from SPDIF at maximum sensitivity with coax plugged in to a SMPS powered DVD-A player. Results from the player when paused are pretty well meaningless because of rubbish due to the SMPS. My 50MHZ Frequency counter does however appear to see something small around 130 KHZ . (?)
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Mar 4, 2010 9:03:54 GMT
Alex, I know you've done a lot of work to your X-Dac, but I don't know what these mods entail. Is there anything along the lines of the crickets in there? Would this further lift your X-Dac?
|
|