Uptone Audio Regen - Clock Mod

[sandyk]

Last I asked John about this, he was not convinced that the better clock was the reason for the SQ bump. It may very well be that powering the clock separately is what is resulting in the change. The USB clock really should not need such low phase noise performance. This is not to take anything away from the fun and good results that folks in this thread are having with their REGEN mods. We watch and learn from all this.

 

Has anybody tried John Kenny's suggestion of powering the existing clock from a Li-ion battery ?

[Pepi]

Turns out terminating the clock with 50 ohm does not work - the DAC won't register and I assume the Regen is not receiving clocking signals correctly. Probably the clock signals are loaded down too much by the 50 ohm load.

 

75 ohm works as expected. Comparing to the unterminated version the 75 ohm terminating one exhibited more clarity. In my previous experiment with a cheap 75 ohm terminating resistor the sound is somewhat tense. This time however, no such shortcoming exists and it is clearly better then unterminated. I am not sure if this is due to the better quality resistor or due to another change where I split the shielding connection and added a resistor to the GND connection between the Regen USB output and the DAC USB input (will report that in a separate post).

 

I am leaving the 75 ohm in the Regen for a longer-term evaluation. Note the Regen in this case is a Green version.

 

Hi JL, thanks for doing this. I followed your lead and got 0.1% 50 and 75 ohm Vishays. The 50 ohm didn't work for me too - the driver could not connect to the DAC with it in place.

 

The 75 ohm works. The sound did change - warmer, more analog but I sense a slight loss of airiness/breathiness. Still good and will be listening further.

[Crom]

Thought I'd post an update after some more investigation. I purchased an amber regen and a couple more neutron star clocks to replace the clocks in my dac. These clocks are connected to iancanada's Fifo project which in turn drive dual mono bIIIse all powered by paul hynes stuff. Previously these were crysteks which had been the best of the bunch I cld afford at the time. Have to say that the dexa clocks literally blow the crysteks out the water. Wonderful upgrade and in my book worth the almost 10x cost between the clocks. I powered the neutron with stock psu's.

 

in fact, the change was much much in fact hugely more than the Change when I went to dexa clocks on the MB...I'm starting to doubt my previous experiments and think that as far as motherboards are concerned power plays a much bigger part than clock quality...which would make sense I suppose.

 

more experimentation awaits...

 

 

Oops, the whole point of the post was to say that the regen didn't do much in my system at all. Not sure whether I was expecting too much seeing as the usb out card is re-clocked alreadyalthough following blizzards comments I'd hoped it might have helped the xmos chip in the waveio card. Maybe Ill try to power the regen clock directly with a lifepo. Incidentally, regen was powered with 7.5v paul hynes.

[zerung]

Crom,

Thanks for the update. Your implimentation is good enough to not need the Regen. Specially at the price point. Another thing to note is that the Packet noise impact was also controlled, else the Regen would have made a difference.

Did you feed the Regen with LPS?

I wonder if there would be more gain if the clock in Regen was also upgraded? Perhaps not as some had also noted Regen not being required when the USB clock was upgraded.

 

Thought I'd post an update after some more investigation. I purchased an amber regen and a couple more neutron star clocks to replace the clocks in my dac. These clocks are connected to iancanada's Fifo project which in turn drive dual mono bIIIse all powered by paul hynes stuff. Previously these were crysteks which had been the best of the bunch I cld afford at the time. Have to say that the dexa clocks literally blow the crysteks out the water. Wonderful upgrade and in my book worth the almost 10x cost between the clocks. I powered the neutron with stock psu's.

 

in fact, the change was much much in fact hugely more than the Change when I went to dexa clocks on the MB...I'm starting to doubt my previous experiments and think that as far as motherboards are concerned power plays a much bigger part than clock quality...which would make sense I suppose.

 

more experimentation awaits...

 

 

Oops, the whole point of the post was to say that the regen didn't do much in my system at all. Not sure whether I was expecting too much seeing as the usb out card is re-clocked alreadyalthough following blizzards comments I'd hoped it might have helped the xmos chip in the waveio card. Maybe Ill try to power the regen clock directly with a lifepo. Incidentally, regen was powered with 7.5v paul hynes.

[Daudio]

... the regen didn't do much in my system at all.

 

And what does that system consist of ?

 

Adding your audio components to your CA Profile would be nice, as then folks can find it anytime, any post/thread

[sandyk]

And what does that system consist of ?

 

Adding your audio components to your CA Profile would be nice, as then folks can find it anytime, any post/thread

 

It's also helpful , provided you don't use your real name, if you include at least the City, or State and Country that you live in as Dave has done below his Avatar.

[Crom]

Thanks for the replies guys:

 

Sorry about not giving precise system details. I did put what I thought were the relevant bits into the post (see para 1) but the digital end changes quite regularly as I tweak. The source is a ridiculously tweaked dual PC (JPLAY: control and play) system using 12 separate DIY linears and all sorts of clock replacements. The stable bits of the system are a Vitus SS-102 with ART Alnico Signature speakers. Software is currently JRiver 19 / Bughead / Roon - or it will be when I get a new ipad ;-(

 

The regen was powered with Paul Hynes power supply.

 

@zerung, yep, you're right. Bottom line, I think it's a bloody clever idea of the Uptone guys to stick a decently powered, nicely clocked USB hub directly in front of the dac input and I wanted to see if it made a difference. I'm also a fan of Swenson's work and approach. What I might do is to move the modded clock from my usb card into the regen or I might remove the regen clock and power it separately. Previous experiments would indicate that separate clock power 'should' make a difference but I could easily be just messing up what was previously a well balanced set up - it's all an experiment!

[AudioBang]

I'm looking for a potential drop in 100MHz clock upgrade for the Vega. Can you provide any options? I noticed on the NewClassD website that they appear to use an Abracon 100MHz oscillator on the Neutron Star. On page 5 of Abracon's spec sheet http://www.abracon.com/Support/ABLNO-Ultra-Low-Phase-Noise-Osc.pdf the phase noise curve looks pretty similar to the Crystek 950 so I'm not exactly doing backflips over the prospect that as a drop in or as implemented in the Neutron Star that it would necessarily provide any significant improvement. But I'm more like a backyard mechanic in electronics and admit that I don't know the ins and outs on this topic. As soon as I hear from the dealer on how to get the 256 DSD firmware upgrade for the Vega, I'd like to do the clock mod if it makes sense. Going back 20 years ago, I had a factory clock upgrade on the Wadia 27 - the technical explanation I received was that they went from a 20ppm to a 5ppm which was Connor Winfield's latest and greatest. I'm sure this is another one dimensional spec that doesn't tell the entire story but I do recall the entire soundstage snapping into a new level of focus after the upgrade. Powering the clock off battery stepped the performance up again. Any recommendations highly appreciated.

 

Rich

[petalthecat]

I'm looking for a potential drop in 100MHz clock upgrade for the Vega. Can you provide any options? I noticed on the NewClassD website that they appear to use an Abracon 100MHz oscillator on the Neutron Star. On page 5 of Abracon's spec sheet http://www.abracon.com/Support/ABLNO-Ultra-Low-Phase-Noise-Osc.pdf the phase noise curve looks pretty similar to the Crystek 950 so I'm not exactly doing backflips over the prospect that as a drop in or as implemented in the Neutron Star that it would necessarily provide any significant improvement. But I'm more like a backyard mechanic in electronics and admit that I don't know the ins and outs on this topic. As soon as I hear from the dealer on how to get the 256 DSD firmware upgrade for the Vega, I'd like to do the clock mod if it makes sense. Going back 20 years ago, I had a factory clock upgrade on the Wadia 27 - the technical explanation I received was that they went from a 20ppm to a 5ppm which was Connor Winfield's latest and greatest. I'm sure this is another one dimensional spec that doesn't tell the entire story but I do recall the entire soundstage snapping into a new level of focus after the upgrade. Powering the clock off battery stepped the performance up again. Any recommendations highly appreciated.

 

Rich

I also own a Vega and have been looking at the dexa clock as a replacement but also not sure how much this clock would improve the sound. The Pulsar clock might be a better alternative but the manufacture is having trouble exporting it because of its potential weapons usage. Links below

Pulsar Clock & Power

Pulsar Clock - Ultra Low Noise OCXO - Page 5 - diyAudio

Perhaps dacs and replacement clocks should be different thread. Anyway AudioBang if you decide to go ahead with a replacement clock in the Vega I would love to know the difference.

[AudioBang]

I also own a Vega and have been looking at the dexa clock as a replacement but also not sure how much this clock would improve the sound. The Pulsar clock might be a better alternative but the manufacture is having trouble exporting it because of its potential weapons usage. Links below

Pulsar Clock & Power

Pulsar Clock - Ultra Low Noise OCXO - Page 5 - diyAudio

Perhaps dacs and replacement clocks should be different thread. Anyway AudioBang if you decide to go ahead with a replacement clock in the Vega I would love to know the difference.

 

Thanks for the response PTC. I did correspond with Auralic and they are looking at 3-4 months to do the 256DSD firmware upgrade for North America. The next hurdle for me is how to get the black metal shield off that I believe is covering the DAC and the clock. I suspect the entire board will need be removed where a few solder pins underneath hold the plate to the board.

[AudioBang]

I also own a Vega and have been looking at the dexa clock as a replacement but also not sure how much this clock would improve the sound. The Pulsar clock might be a better alternative but the manufacture is having trouble exporting it because of its potential weapons usage. Links below

Pulsar Clock & Power

Pulsar Clock - Ultra Low Noise OCXO - Page 5 - diyAudio

Perhaps dacs and replacement clocks should be different thread. Anyway AudioBang if you decide to go ahead with a replacement clock in the Vega I would love to know the difference.

 

Pulsar Clock is now delivering their clocks. It took months of waiting and formal paperwork but I finally got the part and now my mod is complete.

First of all, the Crystek 575 is an outstanding value for $25. But the Pulsar is an OCXO and a step beyond, making instruments more natural sounding, transient blur is significantly reduced and spacial reproduction is further refined. BTW, I apparently still do not have privileges to post a new blog which is why I am writing here...

I got to this page because I was interested in the clock mod for the Uptone regen. I eventually used a Connor-Winfield clock for $25 which was a big improvement but not an easy mod as a "drop in" replacement on the Uptone. In other words, it was not pretty but it was functional with extremely short leads.

Back to the Vega DAC. I researched the NEWCLASSD 100MHz and it is not an OCXO like all their other clocks. It is a TXCO. Their reply was at 100MHz an OCXO is not available. Plus, I was able to get a phase plot from Pulsar that is specific to my production unit and was able to compare specs to the Crystek 575. On average the Pulsar is about 10dB lower noise between 1 and 1KHz. Until you hear the ease of music delivery and improved transients and removal of blur can you appreciate the value of the mod. But don't expect it to be an "exciting" change.

I also modified the power filter on the Vega removing the packaged Schurter filter and replaced it with high quality WIMA X and Y caps - .1uF across positive and neutral and then 2KpF, 1KpF, 200pF and 20pF paralleled from positive to Gnd and Neutral to Gnd. I also added a silver mica 2KpF, 220pF and 20pF on the Pulsar user capacitor pins in addition to the 10uF and 1uF tantilum that come with the unit. The Pulsar requires 150mA at 3.3V for the first 90 seconds then it backs down to 50mA steady state. So I had to remove the second stage regulator which is likely only able to source about 80mA and insert the Pulsar regulator. BTW, the black metal cover illustrated below needs to be desoldered at the tabs to access the clock. Because there is a large degree of improv in terms of getting the black Auralic cover off and desoldering the Crystek you may want to think twice if this is your first mod.

 

I could not cross reference the numbering on the 5-pin SOT-23-5 regulator package.

The regulator is the black 5 pin package between the top and bottom silver can caps. This needs to be removed and replaced with the Pulsar power board. I tapped the 5V input from the top pin of the 5V resistor to the right of the regulator as it is a more robust pad vs the small pad to the 3.3V regulator.

Even using ChipQuik I could not fully remove the clock board - perhaps it was epoxied to the motherboard before it was soldered. As the clock cover and components disintegrated with heat I needed to be careful not to flick solder over the other components, especially the Sabre DAC chip. Leverage your soldering tip away from the SABRE!

Note that the image is rotated {sorry}. The Crystek clock remaining depopulated board is at the top.

I soldered a silver wire extension from the Pulsar Ground pin 2 and Clock output pin 3 [maybe 1/4" in length] to the surface mount pads partially accessible despite the board still being in the way. I inserted and soldered the power pin 4 from the Pulsar clock directly into the Power input hole of the Pulsar regulator and also ran a short ground wire from the regulator to the Pulsar clock and to the robust solder pads where the tabs from the black cover was removed.

Here is the final illustration of the mod -

[Crom]

@AudioBang - thank you very much for the write up. Fascinating and your appraisal of the OCXO's is borne out by others I've read across DIYA and other places. I received a note from pulsar today saying that they are able to take orders again and I am tempted to play a couple of Pulsars off against the Dexa clocks I have.

 

So...Has anyone compared the Dexa Neutron Stars against the Pulsars?

 

Having re-read a bit of this an update about my regen might be appreciated. As zerung correctly surmised last year - my regen didn't do much in my system....that is until I bypassed the regulation and powered the chip and clock directly with a lifepo battery, following a recipe posted by JKenny. Fabulous upgrade.

 

Cheers,

Crom