Jump to content
IGNORED

NVMe SSD designed for audiophiles


Recommended Posts

I am intrigued by the sort of lack of specificity as to what characteristics are improved here.

 

I am sure that there must be some measurements that will show improved S/N or reduced jitter, or something, aren't there?

 

I have an external 10 TB MLC-based DAS storage system that is five feet away, housed in a separate cabinet, and is attached via RAIDed, redundant fibre channel cables; only one 8 Gb FC HBA though.  This system provides what we call 3-D RAID, one dimension comes from system software, the next from the RAID 5 master controller, and the 3rd from PowerPC microRAID chips on each of the 12 MLC boards contained inside (each such is also independently RAID 5).  With Latency in the 80 msec rage, and bandwidth at just 5 GB/s, I have yet to hear much benefit.

 

So - has anybody got real numbers, or are these products looking for something to fix?  I don't mean to be a troll, I am just an old man listening to music (truth to tell, I still have a couple of thousand records)

 

BadK

 

Link to post
Share on other sites

I guess if a music listener does not have the belief in certain components having an impact on Sound Quality then this entire Thread is a moot point...Right?

 

I can only say what I hear in my system hearing how the different Operating System Storage Devices affect the Sound Quality of digital playback from my PC.

 

A simpler way would be to look at the factors that may impact on SQ of the digital playback system. Could it be the power delivery to the Storage device, latency, data transmission rate, grounding, impact of more accurate clocking have influence on the way the OS device operates and hence the resultant SQ.

 

At present I am still burning in the NVME M2 from @darkforce. I can already hear differences in SQ between the Samsung 860 EVO which is the reigning champion and this new M2 drive.  I have experimented with different storage devices in different setups of my 2 PC rigs, a Intel 6700k and a Intel 10700k.

 

My finding is that NVME drives are plagued by proximity to the Motherboard and the Hi-Freq and RF emissions and noise from the various Mobo components. The higher read rates are way in excess of even DSD 512 music reproduction. All that electrical activity produces pulses and that is noise.

The SSD type device are at least shield in a drive bay.

Some on this forum even prefer mechanical type HDD's. 

 

In my experience, HDD's are noisey mechanically and vibration of the read and write heads and platters impact SQ in my system.

Also taking the SSD out of the PC and wrapping both the OS and music drive SSD's in Mumetal shielding with further grounding has taken the SQ to a level I did not hear before. RAID and external NAS may be in the same external benign enviroment being away from the PC. Of course the RAID controller, firmware, powering quality etc are other factors that one would have to address. 

 

I guess there is no hard and fast answer that applies to all systems. It is the details and how the components are put together that matters. 

Link to post
Share on other sites
4 hours ago, badK said:

I am intrigued by the sort of lack of specificity as to what characteristics are improved here.

 

I am sure that there must be some measurements that will show improved S/N or reduced jitter, or something, aren't there?

 

I have an external 10 TB MLC-based DAS storage system that is five feet away, housed in a separate cabinet, and is attached via RAIDed, redundant fibre channel cables; only one 8 Gb FC HBA though.  This system provides what we call 3-D RAID, one dimension comes from system software, the next from the RAID 5 master controller, and the 3rd from PowerPC microRAID chips on each of the 12 MLC boards contained inside (each such is also independently RAID 5).  With Latency in the 80 msec rage, and bandwidth at just 5 GB/s, I have yet to hear much benefit.

 

So - has anybody got real numbers, or are these products looking for something to fix?  I don't mean to be a troll, I am just an old man listening to music (truth to tell, I still have a couple of thousand records)

 

BadK

 

I have about 800 albums dating from the 70's when I first started collecting. And remember the time from 1985 -2001 when digital was supposed to be perfect yet no

one with a good TT wanted to give up their albums for digital. Measurements for digital are maddeningly ineffective other than for mi-fi qualification. Mostly you have to

keep looking and auditioning to find what improves sound then figure out why afterwards to see if its extensible. I do expect to  finally start retiring what I consider my

ordinary vinyl albums with this latest step.  But my original Reiner LP's and Direct to DIscs aren't yet matched by my digital playback as well as other really well mastered

vinyl recordings.

Regards,

Dave

 

Audio system

Link to post
Share on other sites

I figured that using a storage layer that is isolated by a physically separate PS (actually four of them), and fibre cables for data, I would be about as isolated as I could get.  Has anyone else used FC-based storage?  These are available on ebay for pennies, relatively speaking.  Prices this morning like $149 for a 24 drive array with FC-to-SATA controller.

 

I was also a little concerned for you folks who "burn in" flash based storage.  I guess, on reflection, my so-called concern is silly, because, it probably doesn't matter at all, when used as an audio library (i.e. primarily for "playing" the info), but used as actively used OS, cache, or general purpose storage, despite what the manufacturers will tell you, it is subject to eventual burn out. 

 

I know that we all have laptops with it - I do, but this was done when OEMs decided to trade speed and power use for longevity.  They can always charge for the eventually needed replacement parts.  My IT life was in really large enterprise compute systems.  Think in terms of The Earth, or Jaguar, or similar systems.  Here the death of MLC flash is a real event, experienced most every day.  My most recent server here in the Cat Cave, has redundant everything, mirrored RAM, and an internal array of mirrored 15 k RPM disks; I keep all the music external, on that large MLC array.  In the last 30 years, I have not had an unrecoverable data error on anything other than internal flash, so my back-ups have not been needed (except for the laptops).

 

Of course, now that I have mentioned it - something will surely die this week.  I better back it up tonight.

 

Link to post
Share on other sites

1. Same here as my Clearaudio Avantgarde Magnum T/T setup with my Direct Cut Sheffield Labs and Direct to Discs still sounds better than my Digital setup but only just better in some areas like realistic dynamic range.. but my digital set is totally silent and 10x more user friendly!

i still keep my 600 Lps from golden age days😅

No i am not burning in the NAND but the capacitors and power elements on the M.2..

Should be burnt in by tomoro👍

Link to post
Share on other sites

something else to share on this drive, I can't use Euphony core isolation with it, this is what I get when I try to set gstp 3-5 stylus 6-7 where 1 is xhci interrupt

 

1023 /usr/bin/udevil : 0-7
1106 gstp : 0-7
306 /usr/bin/lvmetad : 0-7
418 /usr/bin/haveged : 0-7
430 /usr/bin/acpid : 0-7
456 /usr/bin/dhcpcd : 0-7
758 nfm : 0-7
768 stylus : 0-7
842 /usr/bin/smbd : 0-7
856 avahi-daemon: : 0-7
891 /usr/bin/smbd : 0-7
893 /usr/bin/smbd : 0-7
910 /usr/bin/smbd : 0-7
914 /usr/bin/nmbd : 0-7

Regards,

Dave

 

Audio system

Link to post
Share on other sites
16 hours ago, kelvinwsy said:

1. Same here as my Clearaudio Avantgarde Magnum T/T setup with my Direct Cut Sheffield Labs and Direct to Discs still sounds better than my Digital setup but only just better in some areas like realistic dynamic range.. but my digital set is totally silent and 10x more user friendly!

i still keep my 600 Lps from golden age days😅

No i am not burning in the NAND but the capacitors and power elements on the M.2..

Should be burnt in by tomoro👍

Sheffield's "Growing up in Hollywood Town" is one of my favorites. The CD version is so wimpy in comparison.

Regards,

Dave

 

Audio system

Link to post
Share on other sites

I have the  CD version of Sheffield Lab's Thelma Houston I've Got The Music in Me ! It's a piece of Digital Crap! 

 

Back on topic --- First comparison with Intel Optane shows similar findings to @david256 hears.

The bass and tonal range is expansive compared to the NVME M2 under test.. 

In my experience the accurate clocking gives more focused defined SQ.. still burning in and needs more tine..

 

Next swap will be the Samsung 970 Evo Plus 

Link to post
Share on other sites

In comparison with the Direct Cut I've Got The Music in Me is what I. left out!!

That Dirrct Cut Disc is my treasure from 30 years plus ago.. Tip top condition and I plsy it to remind myself a good digital playbk system  does not win against Vinyl everytime

Link to post
Share on other sites
14 hours ago, davide256 said:

something else to share on this drive, I can't use Euphony core isolation with it, this is what I get when I try to set gstp 3-5 stylus 6-7 where 1 is xhci interrupt

 

 

 

If "gstp 3-5 stylus 6-7" is your exact isolation string then you are missing the default isolation that should be in the first place. Correct string would be "0-2 gstp 3-5 stylus 6-7"

Link to post
Share on other sites
14 minutes ago, c-w said:

 

If "gstp 3-5 stylus 6-7" is your exact isolation string then you are missing the default isolation that should be in the first place. Correct string would be "0-2 gstp 3-5 stylus 6-7"

perfect, I made the wrong assumption a default  didn't need to be specified. Now I get

 

1034 /usr/bin/udevil : 0-2
1108 gstp : 3-5
307 /usr/bin/lvmetad : 0-2
425 /usr/bin/haveged : 0-2
434 /usr/bin/acpid : 0-2
481 /usr/bin/dhcpcd : 0-2
751 nfm : 0-2
761 stylus : 6,7
831 /usr/bin/smbd : 0-2
846 avahi-daemon: : 0-2
886 /usr/bin/smbd : 0-2
887 /usr/bin/smbd : 0-2
904 /usr/bin/smbd : 0-2
909 /usr/bin/nmbd : 0-2

Regards,

Dave

 

Audio system

Link to post
Share on other sites

Alright I think it's time for the next report of the review.

I use Window Server 2019 GUI and HQPlayer 4.10.3 the ranking of the NVME M.2 drives as per my sonic preferences are:

 

1. Intel Optane - 3rd = Last in SQ

2. Samsung 970 EVO Plus - Mid

3. NVME M.2 under review - Best 

 

First a short discussion on Power Supply current capacity and quality and its effect on SQ from these M.2 drives.

a. There is a clock on the NVME M.2. and capacitors as well as higher memory rating. - this results in higher current drive requirement from the PSU.

b. There is a 3 pin jumper on this card, Position 1 and 2 will give MOBO power to the M.2 card. Position 2 and 3 allows power supply injection.

c. The card is sensitive to grounding. Good Grounding by using a Earth Shunt to a Grounding terminal - best would be a Grounding Box.

 

The power units that I tried with this M.2 card is a 1.5 amp LT3045 PSU with Transformer and the second is a 25watt 5V 3.5 amp PSU that has claimed 1 uV purity. THis is actually a USB hub PSU with floating ground function. 

The M.2 test card shows the noise from the PC when the USB is connected to this Hub. Floating the ground helps but the M.2 sounds best without any USB Cable connection - just as a pure 5V PSU.

The 1.5 amp LT3045 sounds cleaner than this USB Hub PSU but lacks the current drive. The M.2 sounded Anemic . Bass sounded weak and the Sound overall was lifeless. Changing to the USB Hub with its 3.5amp capacity brought back the bass line and overall life to the music. 

So while the M.2 design quotes better grounding, the dirty ground of the MOBO gets into the music and can be heard as low level hash. With the ground floated, the music sounded cleaner with better PRAT.

 

How is the SQ of this M.2 drive? Bass line is punchy and full but clearly delineated. Mids are not laid back but slightly forward. The highs are as clean as the PSU quality but is still sounding a bit dry.

 

The M.2 drive is in my opinion still needing more run time to sound its best.

 

Happy Listening

 

IMG_0955.jpg

IMG_0954.jpg

Link to post
Share on other sites

Next step in the test.

I swopped out the Win Server 2019 GUI for the Datacenter Core Edition and ran my AudipOptimizer 3.0 to reduce the no of processes down to 46, leaving the internet and remote metwork control functions intact. This sounded cleaner than the Win Server GUI version.

 

I can hear the differences in the PSU floating ground, Ground On and with and without USB connected to the HQPlayer Server. The M.2 drive is that transparent.

 

A next step was to disable all internet and remote network control functions. This brought the number of processes down to about 25. A further step up in transparency of the SQ. The M.2 drive sounds really nice and clean now.

 

I guess maybe another 24 hours of powering will bring the sound of the M.2 card to its best stable level.

 

Kelvin

Link to post
Share on other sites

The NVME M.2 femto under test is almost ready.

The components are getting "burnt in"

Today's testing session started with an hour of warm up - which brings the total burn in time to about 50 hours.

The Bass line was now powerful and articulate

Mids are less pronounced and blend in with the bass

The Highs have lost the "dryness' and is now well balanced with the mids and the lows. 

All in all, a smooth powerful sound!

 

Now for the Sound stage. Before the burn in period, position of vocals and instruments were not so focused. 

Now the focus have snapped in and between the extreme left and right positions, location of voices and instruments is quite pin point. 
Extreme left and right - outside of the HPs or speakers is still not there yet.

Depth of Sound stage is still not as expected.


I guess another 30 -40 hours with the Crystek Femto clocks powered up for say 24 hours running before serious listening tests will be needed.

 

A quick brief on what I have been A-B comparisons between my Samsung 860 and the NVME M.2 Femto

Both OS drives are on Win Server 2019 Core with AO3. Both powered by the same 5V 3.5 amp ultra low distortion PSU.

I have not formed any firm conclusions yet as the FEMTO M.2 drive is still running in. But my impressions can be summarized as follows

 

If the NVME M.2 is a 100watt Solid State Amp then the Samsung 860 is like a 150 watt Tube Amp 

meaning the 860 is

1. Slightly louder

2. Warmer in tone

3. Holographic in Sound Stage especially outside of Speaker and HP sonic images

4. More depth to Sound Stage

 

The Femto M.2

I believe will outclass the 860 within the next 1 week once it fully run in with the Femto clock running on power for at least 24 hours non stop. I think it will change into a good 100 watt Class A amp once it is fully powered up and femto clock at steady temps. Steady solid bass, clean mids and highs with no dryness or sibilance.

The 860 will remain as a good 150 watt Tube Amp = slightly louder, a bit more fuzzy but nice and warm sound

 

Keep everyone posted in my next report.

 

 

Link to post
Share on other sites

@kelvinwsy do give Euphony a whirl with buffer track to RAM options disabled, cache to drive enabled, paying attention in particular to treble. Was having some resonance like

problems with treble on "Ready Player One " 96/24 soundtrack that this helped, will be trying other recordings today with lots of treble energy.

Regards,

Dave

 

Audio system

Link to post
Share on other sites

abWas Euphony the same software Everyone was playing with 2 years ago - it cost about 29 Euros then. Pierre?? I think was the founder of the SW.

Tell me more about your perceived benefits of Euphony over other OS's. 

If I download the trial SW, I may have to prepare one of my many drives as a double boot device. Say my Optane or my Samsung 970 or even this Femto M.2.

What do you think is the best way to try this out.

Thks

Link to post
Share on other sites
1 hour ago, kelvinwsy said:

abWas Euphony the same software Everyone was playing with 2 years ago - it cost about 29 Euros then. Pierre?? I think was the founder of the SW.

Tell me more about your perceived benefits of Euphony over other OS's. 

If I download the trial SW, I may have to prepare one of my many drives as a double boot device. Say my Optane or my Samsung 970 or even this Femto M.2.

What do you think is the best way to try this out.

Thks

Euphonys claim to fame is midrange timbre integrity. Cost now for license is ~$300 USD so its not cheap. Features include run OS in RAM and buffer queue to RAM

options that I'm beginning to find aren't as good as sourcing OS/caching song queue to NVMe disk. Have ordered some Apacer RAM to see if RAM choice is a factor.
Not sure that it would be helpful in your testing as you have to live long enough with something to build a baseline of experience for reference. Hopefully a next

tester will have Euphony to see if they have same/different experience.

 

Euphony requires license to transfer system from USB boot drive to system drive. But I think you can get around this for a trial if you have an empty M.2 slot

in your desktop PC so that you can flash image to an m.2 drive directly, I recall doing this with a SATA SSD drive. You would have to redo from a USB boot drive

later to complete licensed installation to a system drive.

Regards,

Dave

 

Audio system

Link to post
Share on other sites
On 5/9/2021 at 10:11 AM, davide256 said:

FEMTO NVMe Drive review

=======================

     Drives are an essential part of an audio server. They hold the OS and they are needed for music storage.

Ideally they should be transparent, not affect playback of music. But in my experience that ideal is difficult

to achieve. Media storage drives in the USB audio out server seem particularly difficult to solve for the DIY builder...

I've tried internal NVMe, SSD,HD,SDXC and USB3 attached variations, in the end NAS solution prevailed

for cleaner, better sound.

     But the OS can't be outsourced, has to be physically attached/bootable. OS drives do affect sound,

in particular (a) media player speed,(b) transient attack and (c) digital irritants. With Euphony I've traveled

the usual path of first using USB3 boot, then SSD, and for some time now Ive used Optane M.2 memory

as the best drive option for OS storage. My recent attempt to better Optane using a Samsung 970 Plus NVMe

drive didn't pan out for me as (c) got worse with no appreciable benefits vs Optane.

    Enter the FEMTO NVMe drive as something different to try; I was intrigued as my recent purchase of

a Denafrips IRIS DDC with FEMTO clocking had been a success, perhaps FEMTO clocking on board would matter

for a drive also? The ability to power the drive externally was also promising given how NUC audio output

improved with better main-board power supply. The FEMTO NVMe drive has a nice set of features,

feel free to check the features posted earlier and there is anticipated addition of a heat sink option

 

 


    My installation was in an 8i7BEH NUC with 16GB Crucial RAM, powered by an Uptone Audio JS-2.
The NVMe board horizontal dimension is standard for M.2 but the DC power jack and added capacitors  make it taller.

I had to remove the SATA cage to be able to close the case  and using the DC in jack on the board required

leaving the case bottom off as I had no right angle DC jack adaptors that would allow closing the bottom.  
Digital out was Lush 2 USB to Denafrips IRIS over AQ Eagle Eye coax to Metrum Octave DAC

 

 

image.thumb.jpeg.37ecff09cfb76517c67c5443fc83bc11.jpeg

 

 

     For some reason this drive changes hardware signature with Euphony, did not experience this with

the Samsung drive. So a period of waiting for license transfer before one can transfer OS from boot
stick to new NVMe drive.

    First impressions of sound using main-board power is that this is something different. It's quite soft for sound using

the  Metrum Octave.  Worries start, is this damping transients? A quick swap to a Chord Mojo says no, plenty of transients

here and criticism of the Metrum has been its softer transient response

    Over longer period of time what sinks in is that this is headed in a better direction; the sound I was used to before with

Optane was exaggerating transients. But more importantly the digital irritants have been greatly reduced, listening gets closer

to the analog experience of even if its wrong, it doesn't irritate.

    A system deficit I recognized over the last several months was choral music, couldn't enjoy it when played as I was hearing

wrong harmonics, kind of a glaze on the sound with poor word intelligibility regardless of whether the OS resided on Optane or

in RAM. Had me agonizing over what could fix it...problem solved using the FEMTO NVMe drive

    Have always been on the fence about running OS in RAM vs off the Optane drive... RAM was more incisive
but Optane was smoother. But neither in my testing is as easy to live with as running the OS on the FEMTO NVMe drive.

What surprises me is that 44.1/48k rates in my setup benefit more than 96k and above rates for digital irritant reduction

    The DC power jack is an interesting feature, sound character seemed more solid when I switched to using an iFi 5V power

supply, switching to the second JS-2 output helps further smooth out irritants. But it kept nagging at me that the JS-2 was

also driving the NUC so this was less than perfect electrical isolation.
    The seller has advised that 5v/2.5A is recommended PS but I was able to get this working with an Uptone Audio
LPS 1.2's lower amperage with even cleaner sound and better transient attack on a 96/24 recording, probably
because native Stylus doesn't demand much. So it looks like the DC jack is helpful/necessary for best sound
with a separate good quality isolated supply.


Overall, I'd recommend this to the DIY crowd who use a NUC or PC server capable of M.2 boot if you are struggling

with digital artifacts, and edgy sound.YMMV if you lack a good quality power supply.

 

Per Darkforce they are finalizing channel arrangements at this time, including sale w/wo heatsink,
small batch availability only at the moment. Expected price somewhere in ~ $700

PM @darkforce if interested in early purchase

Transfer your NUC board to a fanless case and your digital irritants may go away. 

Link to post
Share on other sites
On 5/7/2021 at 10:59 AM, kelvinwsy said:

Super NVME M2 drive in my main rig

Gigabyte Aorus Z490 Ultra / 10700k

Pins in 2,3 position to take power injection from external 5v source - min 2.5amp needed

 

Running Windows Server 2019 GUI version with Hqplayer 3.25 and 4.10.3

 

initial impressions very favorable..

Tests will be as below:

One vs the following:

1, Samsung 860 Pro 256gb on 1.5amp LT3045 running Win Server 2019 Core / AO3 

This is the reigning Champion

2. Optane 64gb NVME Express  -Win Server 2019 GUI on MOBO power

3. Samsung 970 Evo Plus 512gb WonServer 2019 GUI on Mobo power

4. Intel SLC 64gb SSD 1st Gen  - Windows LTSC on LT 3045 

5. Sandisk SSD 120gb - Win 10 Pro on LT3045

6. Toshiba 2TB with 12v and 5v external power injection running Win Server 2019 Core 

as Base case

 

My main rig is pretty much optimized with full external fan power and PWM bypass using half speed resistors.. Fully damped chassis and fans are flexible mounted. Case and 860 Evo and Samsung Music SSD earthed and Mumetal shielded outside PC case

3 ATX psu's all Corsairs.. RM550 powering fans AIO  and fan controller. RM650 poweing 12v EPS to 10700k CPU.. HX850 on 24 pin ATX Connector

2 units of Elfidelity PCie power filters on PCie slots.

All fans are on Elfidelity fan filters

 

I am lucky as the Power Company 440v 3 phase supply line comes into my house then 220v single phase enters my house DB.

I had my electriciisn run 2 dedicated lines from my DB into my music room

1 dedicated circiut 30 amp to HP amps and tube preamp, tube monobloc power amps.

DAC is SMSL VMS D1 and Gustard A22 and iFi Micro BL 

Main HP amps 2 x 30 watt pure Class A direct Resistor equipped speaker tap connection. to Abyss AB-1266Phi/CC

 

Spkr System - 45 Tube / 6n1p Double C Transformer equipped Preamp driving 2 Audionote Quest Silver 300b Monoblocks on 2 x LOTH-X Aura Back Loaded Horns with Tannoy ST-50 supertweeters

 

Happy testing ahead

 

 

 

 

 

 

C6C320A8-9199-410B-AECD-CC0FF3BBAE06.jpeg

98B5BE86-8D6E-48E3-8CF9-96E090357A60.jpeg

44E72C6B-E70D-428A-B19F-5F798CD85FFB.jpeg

345A1ED7-944E-4678-837F-A0E2941F7D2C.jpeg

6381E81E-6AFA-4DB8-8245-E8FC50E5C4A9.jpeg

BD21E846-EC66-440F-BD02-04B6DE297DA5.jpeg

I compared my 16GB and 32GB Optane drives and the 16GB sounded much better in my system (more digititis with the larger drive) . I assume the 64GB Optane would be worst. BTW, I also have the Samsung Evo 860 Pro but is now relegated for Laptop use. 

Link to post
Share on other sites
On 5/14/2021 at 10:16 PM, kelvinwsy said:

abWas Euphony the same software Everyone was playing with 2 years ago - it cost about 29 Euros then. Pierre?? I think was the founder of the SW.

Tell me more about your perceived benefits of Euphony over other OS's. 

If I download the trial SW, I may have to prepare one of my many drives as a double boot device. Say my Optane or my Samsung 970 or even this Femto M.2.

What do you think is the best way to try this out.

Thks 


You might be referring to Piero Olmeda‘s AudioLinux? That was going for around that price. 


Euphony’s a very different sounding OS. 

Link to post
Share on other sites

 

Before I fill everyone following this thread, a side bar on drives and Euphony.

 

a. I downloaded the Trial Euphony and used Rufus to make a bootable 32gb USB2.0 drive. It worked. Fiddled around b4 I realized you need to be on the internet. So ethernet port on my 6700k server hooked up; and a laptop for euphony.local and euphony: 8080 for music library. Easy if you read the guide to Euphony startup.

 

b. Played some PGGB 705.6 files on Stylus after copying them to the Euphony drive USB. Even Jriver MC26 on the 705.6 files sounded so much better. Was not impressed at all. Maybe I will install my Matrix Element H USB card, 12V LPSU, and other mods and try again. Maybe even HQPlayer Embedded Trial version to see how it sounds.

 

c. Right now - my optimized WinServer2019Core/AO3 is like Light Years ahead of everything I have heard. But still early days.

 

2. Drives - The drives themselves may not be the sole criterion for SQ. My firm belief is 

a. Power Current and Stability and Low distortion is key. You are after all feeding this power to the drive and that is what will directly affect the SQ you get.

b. EMI and RFI are beasts that need tackling. Putting a drive into the PC case whether in a drive bay on on the MOBO as in the case of a M.2 drive is actually NOT ideal.

c. Grounding is much more important than one realizes until you ground it, listen to the results and you will never go back to "groundless".

 

1. My Samsung 860 is the champion drive in my setup not just bcos of its innate superiority. It is due to all my efforts to shield, isolate, ground and feed it my best power source. At present, it is outside of the PC case, heavy shielded SATA cable with multiple farad clamps to connect it back to the SATA port. Same same with the 1TB Samsung SSD Music drive. Both drives earthed, wrapped with Mumetal and Aluminium Foil and Earthed again to the Earth port of my power conditioner with Star grounding to the entire PC case, PSU, PCIE brackets and then to a Grounding box.

Used as it is the 860 sounds 'BLEAH!'

 

2. I am getting a second Grounding box to earth the USB output of the USB PSU/Hub unit I am using to power this FEMTO M.2 drive. Grounding the earth of the power lead to the M.2 drive is tricky ... May short it somehow. Grounding the parallel USB output port on the USB PSU unit is a lot safer - should be common ground to the power lead to the M.2 drive but with some resistors and caps in circuit to prevent a power backsurge to the drive. A lot safer.

I will report back on the findings.

 

As for the burn in. Another 48 hours under the belt. Here is the listening test results:

 

3. The FEMTO M.2 drive is a very very good Drive.

a. Very neutral especially in the High frequencies. No digital nasties. Extended highs without any sibilance on hardness. This is totally dependent on the quality of the input power. (If you are thinking of using the MOBO power... Don't. The drive is transparent to the quality of the power input.  Even if you have a LPSU powering the MOBO, there are still DC-DC regulators on the mobo which can and will contaminate the power to the M.2 drive.)

b. Bass lines are clean and articulate. Impact of bass and drums is very good.

c. The mid range is neutral, not lacking any warmth and pace.

d. Stereo spread is now very good; you hear sounds extreme left and right just outside of the speakers and/or HPs.

e. Depth is good.

f. Positioning of vocals and instruments is spot on. The temperature steadied Crystek FEMTO clocks really shine hear. There are 3 clocks in my USB chain. OXCO in the Clones Audio USB card, the FEMTO clock on the M.2 and the Clock in the IsoRegen.

 

So the conclusion so far is this FEMTO M.2 drive is a winner. 

 

Next test will be using a USB C M.2 adapter card to mount the FEMTO M.2 drive outside of the PC Case. I can carefully try some shielding around the M.2 drive (my PC case is already shielding and earthed.) and do more listening tests. This will be with and without the USB earthing of the USB PSU Hub powering this M.2 Drive.

 

Happy testing so far

      

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×
×
  • Create New...