So I received the DAC a few days ago, shipping from China was less than 2 weeks. $150 with shipping, wired up the transformer and tested, worked perfectly out of the box. Automatically follows the incoming sample rate, audible relay clicking as the unit changes sample rate. Plugged in the src board and have used it since - I am a believer in 192k sample rate as the higher and more gentle filters sound less ofensive to my ear and the benefits of jitter reduction are well known. Overall the DAC sounds excellent, far ahead of my AKM4393 daily DAC, a very highly tweaked Behringer SRC2496. (For reference my preamp is a CJ PV12A with an amazing pair of Mullards, Herbies tube dampers, naked Vishays for the signal path, v-caps, etc. Accurate (with some caveats) and not euphonic, this is a simple class a, triode, no feedback preamp that I have compared extensively to a wire and know its sound well. Upstream changes are very audible especially solid state components as their distortions are very different and not masked by similar distortions in the preamp.)
Step one: remove second buffer. There are two bridges at c71 and c72 (I'm guessing where insterstage coupling caps used to be on a different revision) soldered my hot wires to the bridges, ground to the original rca output ground, worked great and sounded better with one less opamp in the path. Expected results with one less opamp - better width, more detail, less "electronicy" etc. VS older dacs I'm impressed with how clear the CS chip sounds. Less "jumbled" and artificial sound. Also, the notes finish, meaning the sounds have follow through and ring or sustain much longer than before.
Step two: swapped the remaining 5532 opamp with lm4562. (Sound of 5532 is well known - flat, punchy, a bit bright and gritty...) The sound of lm4562 is also well known - clear, resolved, not terribly engaging. Result was much clearer sound, nice width, deep bass, good dynamics. Room ambiance is very obvious and reverb sounds are much more intense and clearly overcooked on many pop recordings.
Step three: soldered pair of 2.2k resistors to lm4562 outputs to v- for deep class a, soldered 1uf film across v+ and v- (standard mod for me with lm4562) with usual results - slightly more organic sounding, better details, more organic.
We are starting to get somewhere. Recordings sound more different from each other than I realized. I can hear space around drums, particularly the hi hat like I can in real life.
Step four: pair of LT1360 on brown Dog adapter, 2.2k resistors for deep class a, pair 1uf and very large high end 10uf caps across v+ and v- on each chip. I didn't expect this arrangement to sound as dramatically better as it did! Huge soundstage, very organic, dynamics for days, detail that blew my mind... This was surprising because I have always found the lm4562 and LT1361 to be peers of sorts, different but one not vastly better than the other. Perhaps the big difference was the pair of lt1360 vs both stereo channels on one chip?? A very different sound from the usual lm4562 sound and so far the best I have heard.
Step five: wired up a pair of Jensen JT-11P directly to dac outputs. This is very easy as the chip is on a daughter board with large solder points to attach wires to. with seperate outputs to my preamp comparison between the opamp output and transformer output is a simple turn of the selector knob. The output level between the two is exactly the same which made comparison even more simple and relevant. First impression - I could live with the lt1360 outputs as they really sound excellent. The jensen outputs simply sound fabulous. Bass is deeper thatn with the opamps, more detail is present. The frequency spectrum is slightly different. The opamp output is slightly brighter over 10k (noise and harmonics?) a touch thinner overall. The jensen output sounds correct and by comparison the opamp output sounds a bit "fake" if that makes sense. The opamp output has no noticable noise but after comparison with the jensens there is definately something happening low level that should not be there and it dampens the ambient and room tones and little nuances that make many recordings interesting. Overblown pop is suddenly more tolerable and extremely dynamically compressed material sound better and not so noisy. Many recordings sound clearly bad, with poor eq choices, obvious edits, etc. Using transformers (while not absolutelu transparent) to remove an active stage has reminded me of how muich damage active stages can do to an audio signal and why guys like Nelson Pass can charge good money for components that pass along audio without beating it completely to death...
Bottom line: shockingly good sound for under $350. It would be interesting to compare this to something extremely high end like a lavry. I don't know what if any benefits are yet to be had with the power supply sections and welcome some comments! -Pasquale, Orlando FL
Step one: remove second buffer. There are two bridges at c71 and c72 (I'm guessing where insterstage coupling caps used to be on a different revision) soldered my hot wires to the bridges, ground to the original rca output ground, worked great and sounded better with one less opamp in the path. Expected results with one less opamp - better width, more detail, less "electronicy" etc. VS older dacs I'm impressed with how clear the CS chip sounds. Less "jumbled" and artificial sound. Also, the notes finish, meaning the sounds have follow through and ring or sustain much longer than before.
Step two: swapped the remaining 5532 opamp with lm4562. (Sound of 5532 is well known - flat, punchy, a bit bright and gritty...) The sound of lm4562 is also well known - clear, resolved, not terribly engaging. Result was much clearer sound, nice width, deep bass, good dynamics. Room ambiance is very obvious and reverb sounds are much more intense and clearly overcooked on many pop recordings.
Step three: soldered pair of 2.2k resistors to lm4562 outputs to v- for deep class a, soldered 1uf film across v+ and v- (standard mod for me with lm4562) with usual results - slightly more organic sounding, better details, more organic.
We are starting to get somewhere. Recordings sound more different from each other than I realized. I can hear space around drums, particularly the hi hat like I can in real life.
Step four: pair of LT1360 on brown Dog adapter, 2.2k resistors for deep class a, pair 1uf and very large high end 10uf caps across v+ and v- on each chip. I didn't expect this arrangement to sound as dramatically better as it did! Huge soundstage, very organic, dynamics for days, detail that blew my mind... This was surprising because I have always found the lm4562 and LT1361 to be peers of sorts, different but one not vastly better than the other. Perhaps the big difference was the pair of lt1360 vs both stereo channels on one chip?? A very different sound from the usual lm4562 sound and so far the best I have heard.
Step five: wired up a pair of Jensen JT-11P directly to dac outputs. This is very easy as the chip is on a daughter board with large solder points to attach wires to. with seperate outputs to my preamp comparison between the opamp output and transformer output is a simple turn of the selector knob. The output level between the two is exactly the same which made comparison even more simple and relevant. First impression - I could live with the lt1360 outputs as they really sound excellent. The jensen outputs simply sound fabulous. Bass is deeper thatn with the opamps, more detail is present. The frequency spectrum is slightly different. The opamp output is slightly brighter over 10k (noise and harmonics?) a touch thinner overall. The jensen output sounds correct and by comparison the opamp output sounds a bit "fake" if that makes sense. The opamp output has no noticable noise but after comparison with the jensens there is definately something happening low level that should not be there and it dampens the ambient and room tones and little nuances that make many recordings interesting. Overblown pop is suddenly more tolerable and extremely dynamically compressed material sound better and not so noisy. Many recordings sound clearly bad, with poor eq choices, obvious edits, etc. Using transformers (while not absolutelu transparent) to remove an active stage has reminded me of how muich damage active stages can do to an audio signal and why guys like Nelson Pass can charge good money for components that pass along audio without beating it completely to death...
Bottom line: shockingly good sound for under $350. It would be interesting to compare this to something extremely high end like a lavry. I don't know what if any benefits are yet to be had with the power supply sections and welcome some comments! -Pasquale, Orlando FL
You do know there is already a 2400 post thread about this dac on this forum.
I completely agree with your findings, and if you are up to it you can improve it further by pulling the CS8416 and wiring in I2S directly from your source to the ASRC chip. It is another marked improvement, but limits its versatility.
Best, Bill
I completely agree with your findings, and if you are up to it you can improve it further by pulling the CS8416 and wiring in I2S directly from your source to the ASRC chip. It is another marked improvement, but limits its versatility.
Best, Bill
Bill, I've read the thread you mention. It's very long and a fun read and it branches in many directions and covers several revisions. My thread is not meant to break new ground, but simply to offer a "meat and potatos" commentary on the currently available version (april 2010) of the Gigaworks DAC. Long live DIY>
Yes, I have mine set up like this now. Level matching is the only problem you could run into, as I have. The older Philips chips output 5V I2S and the 8421 can only handle about 3.5V. I used voltage dividers on the data and bit clock lines, 1k in series, 2k to ground. The best setup would be installing the board in the player, so I'm ordering another board.
Making subjective comparisons as to the sound, it has improved everything, width, depth, low level detail, bass slam, clarity, imaging, dynamics, etc.
Right now I have a 4 wire bundle about 8" long connecting the units, I used the 8416 socket to plug the resistors and wires into the dac board so I didn't have to do any hacking. It's surely not the best way but it's just temporary.
I'm never going back to SPDIF to play CDs.
Not just soundstage, everything improved.
I haven't tried a clock in this one yet. I did put a clock in the one I was using for SPDIF along with actually reclocking the SPDIF stream, that made quite a difference. I will have to get pretty creative to get a new clock in along with mounting a dac board inside, but I'm going to try. It's an old cd2000, non oversampling, pretty much ideal. I have read that the clock is not that big of a deal with I2S, but I'm going to find out.
I haven't tried a clock in this one yet. I did put a clock in the one I was using for SPDIF along with actually reclocking the SPDIF stream, that made quite a difference. I will have to get pretty creative to get a new clock in along with mounting a dac board inside, but I'm going to try. It's an old cd2000, non oversampling, pretty much ideal. I have read that the clock is not that big of a deal with I2S, but I'm going to find out.
Hi Pasquale,
I just replaced the original opamps with the LM4562 and I think the sound is already great.
I was suggested (on an Italian forum) to use a transformer directly connected to dac output, mainly the Lundhal LL1690.
I am not an expert about transfomers, so I wrote an email to Brian Sowter who replied this:
"I recommend our type Sowter 3603 which is a 1:1 isolator. This will have some filtering effect and will provide a floating output which can be loaded balanced or unbalanced"
Instead you used the Jensen (which is also cheaper). I wonder which one will be the best.
A friend of mine used cheap Monacor transformers but I didn't like the sound, it was compressed with respect to my LM4562 setup.
I am still thinking about which upgrade path to follow!
Best Wishes
Pietro
I just replaced the original opamps with the LM4562 and I think the sound is already great.
I was suggested (on an Italian forum) to use a transformer directly connected to dac output, mainly the Lundhal LL1690.
I am not an expert about transfomers, so I wrote an email to Brian Sowter who replied this:
"I recommend our type Sowter 3603 which is a 1:1 isolator. This will have some filtering effect and will provide a floating output which can be loaded balanced or unbalanced"
Instead you used the Jensen (which is also cheaper). I wonder which one will be the best.
A friend of mine used cheap Monacor transformers but I didn't like the sound, it was compressed with respect to my LM4562 setup.
I am still thinking about which upgrade path to follow!
Best Wishes
Pietro
I will post here because it is not 2400 post long, although the answer may be there.
Just got a gigawork up and running. It works when feeding a signal through the coaxial input but I am not having luck with the USB Port. CAN ANYONE HELP ME. Does those switches need to be changed that are on the cs4398 dac board to get sound out of the USB port????????? HELP WOULD BE APPRECIATED. Thank you in advance.
Just got a gigawork up and running. It works when feeding a signal through the coaxial input but I am not having luck with the USB Port. CAN ANYONE HELP ME. Does those switches need to be changed that are on the cs4398 dac board to get sound out of the USB port????????? HELP WOULD BE APPRECIATED. Thank you in advance.
There are three inputs to this DAC and are selected via the dipswitches. You need to read the documentation that came on the cute little mini CD. A clear description of the switch positions are there.
I finished my player mods and also installed a Tent clock so now the I2S from the Saa7210 is fed directly to the CS8421 feeding 24/192 to the CS4398 and out to two Jensen JT11-DM OP trafos with a couple single pole filters for good measure. I hacked the dac board down to about a third the original size, eliminating everything I didn,t need.
I added new RCAs and XLRs for outputs, everything fit with plenty of space, the clock even has its own trafo.
It is very analog sounding, and very revealing. If the CD has any recording artifacts on it you can hear them, but other than that it sounds like very good vinyl.
I'm sure there are better Redbook players out there, but I havent heard one yet, and I'm sure I couldn't afford it anyway.
P.S. There are two jumpers you move to change the board input, and you can wire the pins to a front panel rotary switch for convenience.
Pietro, I've tried the following transformers with this dac:
Digitec, UTC-A10, Lundhal LL1540, Monacor and Lundhal LL1690. Worst are the monacor. The best are the LL1690 and UTC A-10 (they are bit smellow than LL1690). LL1690 are really expensive compared to the UTC-A10 (over 200 euro with SS)
I've got LL1690's in my DIY dac and I'm really impressed with their performance both on the scope and to my ears. They aren't cheap, in fact at $250, they are more expensive than the whole rest of my dac. But man are they making good music. Currently my dac is being fed by an Apple airport express and it feeds a custom tube amp and into a set of rebuilt Quad esl63's. It is a killer setup for not much money.
Stokes, did you have to do much with them to get them to behave or are the recommendations from Lundahl pretty much right on?
I really don't think I'll ever have the opportunity to get the 1690s with my budget so it's kind of moot any way though. We are probably at the level of comparing flavors, not technical excellence anyway.
I really don't think I'll ever have the opportunity to get the 1690s with my budget so it's kind of moot any way though. We are probably at the level of comparing flavors, not technical excellence anyway.
Cap out also
I have Jensen JT11-FMCF transformers which sell for $50 running direct out from the dac chip with no filters or series damping resistors and the sound is better than any opamp output I have ever tried with the CS43122 or AK4395. I also like the LT1360 better than OPA627 but haven't ever tried any of the newer National amps as I don't have any opamps in my signal path any more. A cheaper, and even better sounding? alternative is to cap couple one leg of the direct dac outputs with 3.5uf (2uf or greater) of good caps such as the Dayton foil. The caps sound even more transparent than the transformers as long as the following load is 10k or greater. The transformers are better into my 4k series/ shunt attenuators, probably because they utilize both legs of the dac chip.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.