0 Members and 1 Guest are viewing this topic. Read 165918 times.
Removed the USB input assembly in totality
Does removing this and the optical input reap any audible benefits or only theoretical? The reason I ask is that the first one of these that I modded(output trafos) was the basic non-USB/upsampling version. For some reason I had also removed the optical input. I'm currently using a "fully loaded" upsampling version board with optical and USB inputs still attached. Maybe my memory is playing tricks on me but it seems that the old board sounded better. Before I start butchering this one I would just like to make sure that I won't be doing it for nothing. BTW - I'd really like to perform some of those other mods you did but for that I'd need to have the procedure spelled out for me.
You've gotta be a doctor with that moniker If you aren't going to use it I'd say yank the sucker out!I will do a detailed modification document with pictures when I take it all apart to put into the case. I have a stock one here to shoot before and after photos of to help it along.
I ordered the gigawork board without the transformer last night, as I am not going to use the op-amps. I know I can just use a cheap 9v wallwart to power the digital side, but I wanted to see if it was worth it to look into a better power source? If so, is there some place I should look? I would be game for building it myself if it would not be too hard and actually worth my time.
Time and money spent on a power source won't be of much benefit unless you also replace the crappy voltage regulators.
In the spirit of the thread why don't you take one of the pics already provided (or off the web), show exactly which regulators you are talking about, then provide some links to replacement voltage regulators that you recommend.
For applications where the recovered input clock, output on the RMCK pin, is required to be low jitter, use a separate, quite, analog +3.3 V supply for VA, decoupled to AGND. Make certain no digital traces are routed near VA, AGND, or FILT as noise may couple and degrade performance. These pins should be well isolated from switching signals and other noise sources. Extensive use of power and ground planes, ground plane fill in unused areas and surface mount decoupling capacitors are recommended. Decoupling capacitors should be mounted on the same side of the board as the CS8416 to minimize inductance effects, and all decoupling capacitors should be as close to the CS8416 as possible.
The type of capacitor used for the PLL filter can have a significant effect on receiver performance. Large or exotic film capacitors are not necessary as their leads and the required longer circuit board traces add undesirable inductance to the circuit. Surface mount ceramic capacitors are a good choice because their own inductance is low, and they can be mounted close to the FILT pin to minimize trace inductance. For Crip, a C0G or NPO dielectric is recommended, and for Cflt, an X7R dielectric is preferred.
I’m afraid any specific suggestion I make will be disparaged as my previous suggestion was. After all, you guys think you know it all.
I’m afraid any specific suggestion I make will be disparaged as my previous suggestion was. After all, you guys think you know it all.Some at diyAudio prefer shunt regulators and other like the so-called super regulators. Either of those would be a significant improvement over the cheapo series regulators supplied with the kit. They also have the advantage of being drop in replacements. I prefer ultra low noise LDO regulators. The disadvantage is they are not drop in replacements. They are a different form factor and usually source less current. But, at least for me, the advantage is that they can be sprinkled around to provide clean power to individual or small groups of chips that need clean power that is not contaminated with the noise spewed out by every other part of the circuit. That was the basis for my earlier suggestion that the daughter boards for the DAC and DIR be reworked to include ULN LDO regulators.I can’t tell what to use because I don’t know what voltages are being used and what input DC voltages are available. Besides, if you are not willing to do PCB layout and SMD soldering, there is no point because, as I said, they are not direct replacements for the existing regulators.I use Linear regulators. I think they were the first on the market. National also makes them. I mount them on tiny PCBs with the required input, output, and bypass caps making a convenient module that can be added where needed, can be easily replaced, and takes up very little space on the PCB to which it is added.I repeatedly bring up this subject because I am flabbergasted that you are investing considerable time and money in an attempt to turn a sow’s ear into a silk purse but seem to be willing to settle for a burlap bag. The voltage regulator and related components are some of the most critical parts of the circuit. This is what Cirrus has to say about it.They are equally specific about the components in the PLL filter. The Gigawork DAC meets none of those requirements. If you are serious about getting the most performance out of an inexpensive kit that is poorly designed in the first place, than you better get really serious about the modifications you are planning to make, otherwise you are wasting your time. That’s why I suggested you avoid the first two DAC kits pictured in the original post. The photo of the third one is not clear enough to reveal details of the circuit implementation but it looks a lot closer to the ideal then the others. On the other hand, if you just want to have fun replacing caps and opamps and think that will bring you audio nirvana, have fun. It takes more than simple parts swaps to really improve a digital circuit. It’s your money and your time, do what you want. A word to the wise: Output caps are there for a purpose; to block DC. Amps don’t like DC. That’s why they often have input caps, but you probably remove those too, don’t you? Signal transformers don’t like DC either, but you probably already know that, don’t you? What you may not know is that some Sigma-Delta DACs have considerable DC offset in their analog output, by design. If you don’t remove the DC in the reconstruction filter or with output caps, you are asking for trouble.Why is it audiophiles, who can’t read schematics and who don’t read datasheets, think they know more than the engineers who designed the components, drew the schematics, and wrote the datasheets?
You have been asked on several occasions to add to the discourse and you have yet to do so. A tutorial would be a good thing from you.
This captures most of the mods in one picture.........I auditioned opamps: AD826 - forward and hard, almost brittle sounding OPA2134 - smooth but not very articulate LM4562 - woofy and vague sounding, hard to define some instruments especially in the mid to upper bass Twin OPA627 (Gigawork) - as always very good with nice bass and smooth yet detailed, very dynamicThere you have it so far. The journey will continue............................... ......Dave
.... (to jb) Better yet, become an entrepreneur and whip up a batch of your super-duper gimswitches and offer them for sale. If the price were right I would buy some.
Exactly what kind of tutorial would you like?
Dave,Based on the above picture, can I assume that both channel's differential outputs are handled(combined/'single ended') by a single dual opamp, and that the output buffering for both channels are handled by the other (no longer there) dual opamp? or did you reroute the signals to use that single dual opamp?If your mods didn't require that signal rerouting an alternative might be replacing the opamp recieving the differential outputs with an OPA2107 (or that wonderfully priced dual OPA627) and using a LM4562/LME49720 for the output dual opamp. This would also facilitate removing all signal coupling caps. While I do like the sound of the OPA627, I subjectively prefer the sound of a 2107/627 coupled to a LM4562. Its not necessarily better, just voiced differently.Great suggestion! I'm in for that group buy. The LT176x/196x LDO regulators are excellent.Regards,Paul
....I am a 'less is more' guy and eliminating the buffer lifted another veil from the soundstage and it got wider and deeper with improved timbre. For what I've got in this DAC, it doesn't suck.Dave
Indeed, and obviously opamp eugenics is fraught with cosmic humor; one might get the worst of combined characteristics, rather than the best. But OPA 2107 works particularly well with the LM4562, fed from the CS4398, and the architecture seems to allow a simple swap of opamps in the extant sockets, and a bypass of signal coupling caps. IMO, preferable to a Zapfilter handling the CS4398 outputs..... during the height of eugenics popularity among the English, a lady suggested to her fellow dinner guest, George Bernard Shaw that they should reproduce for Society's benefit, to which he responded -'But Madame, suppose they get MY looks, and YOUR brains?'Regards,Paul
Anything that you would consider doping to educate us ignorant audiophiles would be a good start.
Anything that you would consider doping to educate us ignorant audiophiles would be a good start.Dave
I tried that. I described the importance of voltage regulators in DAC design quoting from the Cirrus datasheet. You complained that I was preaching. Before that I explained that unless you replace the voltage regulators on the Gigawork DAC you'll feed the DAC chip a dirty clock and dirty power and changing caps and opamps will not clean it up. I also ponted out the the Gigawork DAC did not meet Cirrus' advice regarding component selection and PCB layout and was a poor chice for further modding. Wushuliu didn't like hearing any of that. I think you guys just don't like hearing that you picked the wrong DAC. As I said, I think the third DAC pictured in the original post would have been a much better starting point for further modification. There was less wrong with it to begin with.Pick a topic and I'll share what I know. It appears no one here is comforatable with surface mount and all high performance digital audio hardware today is surface mount. That kinda rules out hardware discussions. You should be happy it hasn't transitioned to BGA or "chips without pins." It's no more difficult to work with than SMD, or so I'm told, but you need specialized equipment. That will leave hobbyists out inthe cold. Maybe you want to talk about software. I do that too.
