EVS modified Gustard X20/Gustard X20 thread

[Quadman]

Mark,

Like Ric I made my own cable, I used about 6" of pure silver wire, I then slid some tinned shield over this and soldered two grounding wires on the shield. I then wrapped all the in teflon tape. Direction is very important so you need to experiment which direction it is obvious which is better.

I used a abracon OCXO AOCJY1 100mHZ clock, no need for the 106 clock as the gustard is not native 48K capable at 512, it down samples to 44K levels. Look at the LKS 003 thread on head fi around P16 b0bb shows what he did. (Note this requires an external PSU and high speed regulator, as a OCXO demands much more amperage than a normal none OCXO clock). Since the clock is dip14 I soldered some very short wires to a dip 14 socket and now can swap dip 14 clocks at will. The Abracon is about $65, cheaper than a pulsar, but not as good, but better than the crystek or accusilicon which Gustard uses. The clock mod replaces the coax cable, I have a picture on head fi of my clock installed (maybe without the dip14 socket), but with an ohm meter pretty easy to figure out what pads are signal and ground for the clock, and the power for the abracon would come from an external PSU. Only 3 of the 4 pins on the clock are used (signal, ground and power)

Bottom of boards are just solder no components. No I did not play with those diodes you mention, thought about it but at 512 music is so good I hate to put the dac on the bench.

I build my own PC's for streaming a i7-6700K or even 7700K would run about $1000 parts cost for a decent machine, not the SMG $16K server level but quit adaquate. I have built 5 audio PC's so far, learn something on every build. Connect direct to dac with USB they will work as dac's come and go. Good investment.

I have read through every X20 thread on every site I could find and probably have lost track of it all. I wish there were more people exploring new mods, trying them out and reporting back. Maybe diyAudio is the place for that.

What high-grade microwave coax did you use for the 100 MHz. clock?
What clock board? Link (besides the $400 Pulsar!)? Instructions anywhere for installing one?
Is a 106.25 MHz clock (instead of 100 MHz) worth a try to better handle DSD512?
Are there any components on the bottom of the DAC/buffer board? I've not needed to take mine out yet and it seems like there ought to be more/bigger bypass caps around the DAC chips.

Have you noticed / replaced the M4 diodes (D12, D13, D14, D15, D26, D27, D28, D29) on the edges of the DAC/buffer board down by the output buffers? Are they worth replacing with SMT Schottkys? Yeah, I know Lindahl transformers are the way to go, but these might be polluting power rails if on them.

I've resisted getting into the whole music server and software thing because it sounds like a longer crazier journey than soldering iron mods. $15K for a tweaked PC from Monaco? I'm also trying to be rational on investment when new DACs and CPUs come out all the time (like the i7-7700 due out this quarter). I'd like to be able to take out a $400 clock and Lindahl transformers if I put them in my X20 and use them in some future ES9038 DAC, without ruining the X20. As of now my tab is only about $20 inside and $100 for a AQ power cord.

Another challenge is the DAC sounding so good that I end up listening to it instead of putting it back on my bench! If I had a second one to leave playing while working on mods, I'd probably still get distracted by the music.

BTW, I'll try to attach a pic of my back panel which shows my mounting of the transformers using 1/4"x2" nylon bolts, nylon fender washer, and 1/8" thick rubber washer. No metal in there with two corrugated cardboard and the supplied thin rubber disk under the transformer.

 

[Ric Schultz]

Mark, Have you read older posts on this forum? I went through 2 -X20U's doing Ric's level 1 mods. The second one I compared the sound to the first one which was full Level 1 of Ric's suggest mods, at each stage of Ric's mod path. These comments start around post #31, I did not begin to mod the second dac until I had close to 200 hours on it with signal. The coax cable is a huge change, even bigger than the diodes. An even bigger change is to put a clock right at the coax tower on the output board. Then, of course, the biggest change of all is to get the dac to play DSD512, which I describe a bit on this forum and in more detail on Headfi. DSD512 is where the magic happens. BUT, you need a big boy (i7-6700K) PC to handle HQPlayer at DSD512, though I did manage with my i7-4790K PC with HQP's less demanding 2s filter sets. You still have a ways to go to see what this dac is about. It is a very fun and rewarding journey.

The Gustard cannot do DSD512 with its stock usb board. Quadman has hardwired the DIYINHK DSD512 usb board into the Gustard in place of the stock usb board and is powering it from an external regulator. He is the only person, that I know of, that is listening to a Gustard at DSD512. This cannot be done easily......this is for advanced tweaks (took him quite awhile to make it work and have it be stable). Gustard might make a usb board that can take DSD512 in the future......but powering it separately will still sound better. However, the Gustard can do DSD512 via its I2S input. Singxer makes a usb to whatever (including I2S) converter that currently will pass DSD256. It is supposed to be upgraded via firmware? or have a new unit available soon that will pass DSD512. When this is available....hopefully soon, then everyone with a Gustard will be able to hear upsampled DSD512.

What Quadman does not have is the latest DAC chips (way, way better....old DAC chips sound like mud!), the better Pulsar clock and he is still listening to the stock output stage....so, we are all always in transition. Yesterdays great toy is todays bathroom system.....he he.

What blows my mind is how incredible this Pro thing now sounds at level one mod (This is playing a 16/44 disc using my modded Oppo as transport through a custom one foot long digital cable). I did two new things that take it even further out and I can only play one cut because I am so orgasmic listening to it that it freaks me out....really! Never heard anything like this before....except live. I can only imagine how it will sound with the Pulsar clock...better output stage and then upsampled to DSD512.......Actually, I cannot imagine it sounding much better than it does.....but I will let reality be as incredible as it wants......"Its only just begun"!!!!!

Pulsar clock scheduled to be delivered on Friday......if it does then a fun weekend ahead.

I know I said this once before but this is important......the stock Gustard is just good. You will not get orgasmic listening to it. If you had one of the old ones and new ones and A/Bed you would think the old one was just a distant cousin......and the new one is a little better. But modded....OMG, OMG.....goosebumps, tears, orgasmic shouts of joy....yes really. The three areas that Gustard messed up are the AC power, the DC diodes on the DAC board and the clocking......you clean up those areas and add some Quantum dots and damping and it is like night and day. The curtains have come off. I have had the same electronics and speakers through all the level 2.5 modding of the old Gustard and now this level one mod of the new Gustard. It is only now that my mind is blown. This is the first time in years that I am excited about my stereo. I knew my DAC was limiting the transparency, but now I really know what I was missing. This thing rocks!

 

[MarkK02474]

Yeah, this DAC has made my system sound better than ever, giving me imaging and details I didn't know possible. Didn't know the rest of my system was that good either. The imaging is so amazing with speakers that headphone listeners are missing out.

I'm still pondering diodes. There are 2 Vishay Fred Pt surface mount models that are drop in replacements for the bad M4 diodes. One has less forward voltage drop but a little slower and the other trades those two. The lower Vf one is 78-VS-3EMU06-M3/5AT. Cheaper and tidyer than through hole TO-220 freds. I'm going to order some for the buffer section.

For adding extra power supplies, I'm tempted by Ifi's iPower 15 Watt wall warts for $50. Only go down to 5V, so need a 3.3V regulator after depending on use. The amazing thing is a claimed 1uV noise figure. I have their $150 SPDIF scrubber on order.

 

[MarkK02474]

My new ifi SPDIF iPurifier just came. Took maybe 2 seconds to hear a veil or two removed and the soundstage grow yet taller, wider, and deeper. $149.99 well spent. It came nicely packaged with adaptors and the 1uV noise, 5 volt iPower wall wart that I was thinking of using to power a clock board some day. This little iPurifier device also doubles as a coax <-> toslink converter for the input challenged, though I think purified coax still sounds slightly better than purified toslink.

 

[MarkK02474]

Seems I didn't get the ribbon cable fully seated, so don't have accurate results on the diode change. Sorting things out.

 

[Ric Schultz]

From my experiences of replacing the output stage....in stages....the discrete output buffer has practically no sound of its own....best to look elsewhere for better sound. Replacing the entire output stage is what I am doing. I already have the transformer output stage designed and now I have a solid state circuit I will try soon as well and A/B.

I just put in the 950 series 100 meg Crystek clock (industry standard "Femto" clock) using the stock power supply but using one of my modded .33 Wima caps as a bypass cap right on the clock. Now will let it burn in a few hours and listen again. First listen cold was nice. I will also add a WA Quantum chip to the Wima and see what this does. If this turns out to be a noticeable gain (and I think it is) then I will offer the Crystek clock as part of level one mod (no increase in price).

The $400 Pulsar clock and power supply arriving Monday.....fun stuff ahead soon!

Listening to some 24/96 DADs (remember those? first high rez recordings ever produced to sell....Classic records....late 90s). Unfortunately, the Oppo is just sending 48K to the Gustard....still, after more disc treatments Red Rodney 57 and Pulse (John Cage percussion music) sounding pretty dang great!

Nice improvement with the WA Quantum chip on the .33 bypass cap. Burning in nicely. Will burn it in till I get the Pulsar clock and power supply Monday then change the power supply first and listen and then change the clock to the Pulsar.....then comes the output stage.

 

[Ric Schultz]

Just this morning received an email and part of it is this:

Unit now has over 100 hours and has improved some more. I do not find it lacking stock apart from a very slight dryness and lack of fullness.

In fact it is now more detailed and three dimensional than the Esoteric K01X when fed from my Aurender N10. The Esoteric sounds a touch more natural tonally though the difference is very small.


He is talking about a stock Gustard Pro ($869) A/Bed using AES digital with the digital section in the Esoteric K-01X ($20,000). Pretty cool! So that means that level one mod would completely trounce the Esoteric in every way.....wow! He wants the whole shebang at level 2plus. He can sell his Esoteric and get many modded Gustards......or even take a trip around the world. The times they are a changin'

Just as I way typing this I heard the UPS truck drive up.....Yup, the Pulsar clock and power supply have landed....oh boy!

 

[SJM60]

Greetings! I am new to this site and hope this is the appropriate place to post this. I am the proud owner of a Gustard x20u with complete mods by Ric (also have an OPPO 105 with mods by Ric). I love it! I also recently purchased the Gustard u12 because I read on another forum that it will also improve sound by connecting to the x20u via the IIS connection. It was a cheap unit, so I'm giving it a try using standard HDMI cable. I'm using JRiver and the standard XMOS driver for USB.

In JRiver I'm upsampling to DSD, but I can only get DoP to DSD 128. The DAC is capable of much more through the IIS connection per it's specifications, so I'm wondering what I'm doing wrong. I read that the XMOS driver is only capable of DSD 128, so is that my limitation? If so, how does one get higher resolution DSD files through the DAC if the XMOS driver is the only driver? I can't find any switches or check boxes in JRiver to enhance the signal.

Very much appreciate the help!

SJM60

 

[Brucemck2]

Ric, any updates?

 

[gstew]

Seems I didn't get the ribbon cable fully seated, so don't have accurate results on the diode change. Sorting things out.

Any updates on the diode change. Looking for a good SMD diode for upgrades.

TIA!

Greg in Mississippi

 

[Ric Schultz]

Greetings! I am new to this site and hope this is the appropriate place to post this. I am the proud owner of a Gustard x20u with complete mods by Ric (also have an OPPO 105 with mods by Ric). I love it! I also recently purchased the Gustard u12 because I read on another forum that it will also improve sound by connecting to the x20u via the IIS connection. It was a cheap unit, so I'm giving it a try using standard HDMI cable. I'm using JRiver and the standard XMOS driver for USB.

In JRiver I'm upsampling to DSD, but I can only get DoP to DSD 128. The DAC is capable of much more through the IIS connection per it's specifications, so I'm wondering what I'm doing wrong. I read that the XMOS driver is only capable of DSD 128, so is that my limitation? If so, how does one get higher resolution DSD files through the DAC if the XMOS driver is the only driver? I can't find any switches or check boxes in JRiver to enhance the signal.

Very much appreciate the help!

SJM60

The driver that comes with the new Pro can do DSD256. hopefully it will work with the non pro too. I have it on my computer. I can send it to you.....email me. I don't know if the driver only works with Windows or can work on a Mac.

 

[MarkK02474]

Any updates on the diode change. Looking for a good SMD diode for upgrades.

TIA!

Greg in Mississippi

Sadly, I think the "ultrasoft recovery" TO-220 Vishay Fred I tried sounds a little better than the "soft recovery" SMD Fred I tried. That is VS-HFA04TB60PBF sounded better than VS-3EMU06-M3/5AT. I wish it weren't so as a SMT device is so much neater. A difference between the parts is the 4A, 600v TO-220 part is faster with a greater forward voltage drop. The SMT I chose is slower with less voltage drop. There is another SMD 3A, 600V SMT that is a little faster with more voltage drop. Its not as fast as the TO-220 one. Its possible faster could be the key to better. I'll leave for others to do more comparisons.

Either one though sounds vastly better than the stock diodes.

I have an Oppo Sonica DAC on the way to me! I had signed up on the Oppo web site to notify me when it was available and they emailed me yesterday with a personalized link to pre-order, which I did. Scheduled delivery by Fedex ground is Wednesday.

[Edit: After some more break in time and listening, I better appreciate the virtues of the SMT diodes. So, while there isn't as much depth with the SMT freds as the TO-220 freds, what they do have is more dynamics. Some of that is from less voltage drop giving the storage capacitors a little more voltage, and thus requiring break in at the new voltage. There is also probably less resistance with a SMT component than TO-220 with copper leads despite the SMT rated for 3 amps and the TO-220 rated for 4 amps.

This is troubling. It shows that there is much room for improvement in the power supply. Bigger transformer, lower ESR and bigger caps, and hours of voicing different capacitors, diodes, and voltage regulators will produce rewards. The optimistic side of view is that the DAC chips are so revealing of everything that good care and feeding of them produces never ending benefits. There is still vast unrealized potential for the ES9028pro. The downside is that much time and work is needed to get the most out of these chips.

I'm pretty sure I chose the wrong surface mount diode to use. This one may be better: VS-3EMH06-M3. The Peak recovery current is slightly lower and Reverse recovery charge is half, meaning half the noise. The price is a little less voltage getting through to the storage capacitors. I will probably get some to compare eventually.
]

 

[Ric Schultz]

Mark,
Interesting story and I am sure you heard what you said. However, your story as to why is just a total guess. Might as well guess how many stars in the solar system. What I have found over and over again through listening tests is that even parts that have the same specs sound totally different. I am listening right now to a clock that measures exactly like the Crystek....doesn't sound like a Crystek! Thinking that voltage drop and/or whatever is the "reason" something sounds a certain way is really just a total guess. Wait till you hear a bunch of parts that measure the same and sound different.....your story will then will be "I really don't know ****, but I like this one best". That is where I am at now after almost 40 years of listening tests. But, guessing is fun.....so have fun. The mind wants to compartmentalize everything. I mean if there is no "reason" then we have to try every frickin part in the universe.....heck, I don't have time for that....let's just pretent I know something so I can just listen to a few parts and have a life. I know a person who almost lost his wife because he was frickin tweaking all the time. He should have been f.cking more instead of tweaking....he he. I don't currently have a partner so I can indulge in this game a lot. One thing is for sure, every part sounds different and every way you mount it makes it sound different. Infinite madness, this tweaking thing........and it never ends.....just like life, just like love....infinite and eternal.

I have been listening to clocks and output stages and new tweaks on the Gustard Pro here (actually two of them in rotation). I hope to have it all wrapped up sometime early next week and have all the mod info online and will post here as well. Sounding more outrageous all the time. This thing rocks!

 

[Val33]

Hi Guys,

I have just received an X20-pro. Its sitting on the bench running in at the moment. I've had the top off to hardwire my RPi3/Kali reclocker i2s direct to the USB header. I'm in the UK and have the same issue as Simon, in that the voltages are a tad high. I'll probably knock up a bucking transformer to start with, but may replace the torroids with some custom audio grade ones in the future.

I am confused by the clocking used and wonder if anyone has really worked out what is in use and when. I only listen to PCM so thats my main interest. I believe that when MCLK Mode is 'Normal' the DAC is in async mode and reclocking everything using just the 100mhz clock?
But, what is it using when in 'Auto' mode? Is it using the onboard 45/49 clocks or still just the 100mhz?

I guess the real question, is it worth upgrading the low quality 45/49 clocks as well as the 100mhz? DIYINHK have some decent NDK's for not much money so I may just swop them out anyway.

Finally, what's the general recommendation for diodes?

Thanks

Val

 

[Ric Schultz]

What I am doing is using a better clock right on the DAC......sounds way better than any of their two modes. However, which clock, how you mount/wire the clock and damp it and bypass the power supply is super important.

 

[Val33]

Hi Ric, What do you mean by 'any of their two modes'? Surely there are only two modes, Normal and Auto, no matter what you do to the clock?

 

[Ric Schultz]

When you put a clock directly on the DAC it is operating in Async mode, I believe. So there is only one mode.....the mode switch on the machine no longer does anything. You are removing the digital cable from the digital board to the DAC board.......running the ESS DAC like everyone else is, with a clock directly on the DAC.
Gustard was getting cute with their two mode try but having a clock on another board and using a junk digital cable and connector between the boards is severely limiting the performance.

 

[Val33]

Interesting, i hadn't thought of it like that. My clock should be delivered tomorrow, so I'll have a play.

Are you removing the coax tower when fitting the clock or leaving it there? Is there anything else to remove? I'll be powering the new clock with a Lifpo4 battery.

I may also try feeding the MCK from the Kali Re-clock board direct to the DAC chips too. Be interesting to compare.

 

[Ric Schultz]

Yes, removing the coax tower

No, nothing else to remove.

 

[Quadman]

@val33 are you talking about the clocks on the USB board (45/49), no one to my knowledge has replaced them, As to what happens with the digital signal Simon is the best one to answer that, he really was going deep down that route trying to basically bypass the digital board with a differential I2S signal to the LVDC chip on the output board but ran into big snafu when he discovered that the main traffic cop of the I2S lines is the big chip on the LED board. We both use 3rd party USB boards his a Wave I/O and mine the DIYinHK dsd512 one and when we hard wired into the I2S lines we both decided to use DSD only and no PCM provision.