HOLO Audio MAY DAC

[lpost]

8 hours ago, scintilla said:

 

So I don't think there are two 'sides' to break in. 

I don’t really know but everything I’ve read talks about separate paths. I don’t believe DSD can be processed thru an R-2R ladder.

 

hoepfully others that know can chime in.

[madfonzy]

Some text from the crux audio website

 

”Patented R2R technology. Discrete DAC that has Linear compensation and this allows for ultimate music reproduction accuracy. Separate R2R network for PCM, and  R2R for DSD.

HOLO Audio is the world’s first to support DSD natively on R2R DAC, so far the only one. This is not the DSD converted to PCM before digital-analog converter, but directly by the discrete components of the DSD digital to analog converter. Supported currently on MAC (DOP)and Linux (DOP), and Windows/PC (Direct Native and DOP).“

 

so yes it has separate ladders.

[GoldenOne]

10 minutes ago, lpost said:

I don’t really know but everything I’ve read talks about separate paths. I don’t believe DSD can be processed thru an R-2R ladder.

 

hoepfully others that know can chime in.

 

1 minute ago, madfonzy said:

Some text from the crux audio website

 

”Patented R2R technology. Discrete DAC that has Linear compensation and this allows for ultimate music reproduction accuracy. Separate R2R network for PCM, and  R2R for DSD.

HOLO Audio is the world’s first to support DSD natively on R2R DAC, so far the only one. This is not the DSD converted to PCM before digital-analog converter, but directly by the discrete components of the DSD digital to analog converter. Supported currently on MAC (DOP)and Linux (DOP), and Windows/PC (Direct Native and DOP).“

 

so yes it has separate ladders.

DSD Cannot be processed natively on an R2R ladder.

Rockna converts DSD to 768khz PCM. Denafrips converts to 6-bit (unknown sample rate)
Holo has a completely separate DSD converter which is genuine 1-bit.

It does not use the R2R ladder for DSD

[lpost]

and this from Kitsune website. It's rather ambiguous 'based on R2R however we don’t refer to it as R2R since the design is far superior.'

 

'The May DAC is the most recent product and is capable of DSD native on an Discrete Resister Ladder DAC!  This May DAC is capable of DSD1024 native and actually running on his bespoke discrete ladder DAC! This is a unique custom build discrete resistor network dac with linear compensation and advanced architecture based on R2R however we don’t refer to it as R2R since the design is far superior.'

 

Regardless of the tech, it sounds wonderful.

[GoldenOne]

1 hour ago, lpost said:

and this from Kitsune website. It's rather ambiguous 'based on R2R however we don’t refer to it as R2R since the design is far superior.'

 

'The May DAC is the most recent product and is capable of DSD native on an Discrete Resister Ladder DAC!  This May DAC is capable of DSD1024 native and actually running on his bespoke discrete ladder DAC! This is a unique custom build discrete resistor network dac with linear compensation and advanced architecture based on R2R however we don’t refer to it as R2R since the design is far superior.'

 

Regardless of the tech, it sounds wonderful.

I imagine it might share some similarities with stuff like the tambaqui or this: http://puredsd.ru/

[lpost]

35 minutes ago, GoldenOne said:

I imagine it might share some similarities with stuff like the tambaqui or this: http://puredsd.ru/

Yes, very likely. I believe this is Jussi Laako’s open hardware design for DSD.

[scintilla]

An addressable resistor network can be configured by logic in numerous ways by simply changing the addresses.  You have to understand that there are not fixed networks on silicon, but that they are formed by controlling logic.  Likely that the fpga controls the network based on the input signal, so it forms ladder dac for pcm and low-pass network for dsd using the same set of resistors. It an elegant solution.  Jussi has previously stated that this is similar to the dsc-1 design, but I think that Jeff has done his own refinements and improved upon the linearity with dynamic compensation in ways not anticipated by the dsc-1 proposal.   

[Jean Paul D]

On 4/16/2021 at 10:56 PM, Quadman said:

May would be 2x maybe 2x+ over used T+A,  if you can stretch, the may is the clear choice.  It is that much better almost an endgame dac.  My SME30 mk2 TT and dynavector XV1s cartridge just sit idle as my digital has surpassed it in SQ (even the T+A at 512 was the equal to it). I briefly tried the T+A direct to my amplifiers like @fds does and it was very good, but I have a buffered preamp with a very simple circuit and very high parts quality and really heard no difference between direct to amp or through my buffered pre.  So I keep my buffered preamp in the chain.

 

I also, last weekend, made an upgrade to my L2 May (making it almost a KTE minus the copper R2R board covers) and will report on that tomorrow after one more long listening session.  Dang I just love the MAY.

someone compared our TTs. My tonearm is different though. With mono albums I can make the hum and idler rumble disappear but with stereo it's far from perfect black background. But very exciting sound https://www.inner-magazines.com/audiophilia/sme-30-2a-versus-emt-930st/

 

How about the almost KTE upgrade?

[fds]

On 4/16/2021 at 11:36 PM, Jean Paul D said:

But -3dB for headroom (HQP), -6 for DSD vs PCM with a Holo DAC, - 8 9 10 dB maybe more if I choose convolution in HQP

This is exactly the main point that makes me hestiate to try the May. To my knowledge, the "-6 for DSD vs PCM" is not present with the T+A whose balanced XLR outputs I am using. Thereby, it still seems to me that a preamp cannot be avoided with the May to get optimal results for DSD playback in particular when one wants to use Audiolense-generated filters in HQP convolution.

[Quadman]

4 hours ago, fds said:

To my knowledge, the "-6 for DSD vs PCM" is not present with the T+A whose balanced XLR outputs I am using. Thereby, it still seems to me that a preamp cannot be avoided

It is true despite the Holo having a higher output voltage SE (2.9V) vs the T+A SE (2.5V) (balanced should be 2x) I find I have to use higher volumes (when up sampling to DSD) in HQP to play my usual loudness compared to the T+A.  OTOH if you think the T+A direct to amp will still beat the Holo through a preamp then Aerosmith wrote the perfect song for you, Dream On.  And seriously what is your amps sensitivity?  Balanced Holo out is 5.9V, -6db I believe is still 1.475V most amplifiers reach full rated power well below this figure, My amp reaches full power with .75V.  Do you still need to set HQP at -3 if the signal is already -6?

[Quadman]

As I mentioned Friday, last weekend I performed surgery on my Holo L2 dac to make it as close to the KTE as possible.  the main difference's between the L2 and KTE are as follows.

*  Kte replaces the IEC copper wire and regular faston connectors with 1.5MM silver wire and silver/rhodium plated faston connector

    they also remove the "connector" to the Power PCB and solder direct to board.

*   The stock Shurter gold plated fuse is replaced with the Red Nano fuse

*   Standard Vishay capacitors are replace with special KTE branded capacitors

*   Hand selected Dac modules with black coper metal shields over the boards

*   KTE branding logos

 

*  I replaced the standard copper wire and standard faston connectors with DH Labs Silver Sonic high performance silver plated          continuous crystal 14ga hookup wire and WBT gold plated copper faston connectors, soldered direct to the PCB.

*  Previous owner of my L2 had already swapped the stock shurter fuse with a synergestic Orange Fuse.

*  I replaced the 2-Vishay MKP 1848s 10uf, 700v capacitors with Auricap XO 10uf, 160Vdc capacitors

 

If you look at all the pictures posted on the Kitsune Holo May page the first group show the L2 dac then in the middle they show a full assembled dac with the fox KTE logos then after this all internal pictures are from the KTE version so you can further study to see the differences.

 

Power wire harness before and after.

 

Vishay Caps on input board near front of dac and replacement Caps

 

Now I just have to figure out if it is worth covering the R2R modules with some type of RFI/EMI covering.  For now I listen.

 

The KTE special Capacitor they use only has a black wrap and a fox logo so brand is unknown.  I do know that the solder points for the Vishay's are 37.5mm apart and in kitsunes photo's the new KTE cap has at least 2-3mm inside those dimensions making the cap 31.5mm - 33.5mm long.  Searching the web trying to find 10uf caps that fit that size there are only 2 I could find that would make possible replacement, a Nichicon or an Illinois' capacitor.  I am very confident it is not a Teflon capacitor as the film is thicker and any Teflon cap I could find is bigger than the solder points.  I choose a very high end Auricap XO which is about 35mm wide, the pair were $80.  Is it better?  no idea its really good though.  Is the Synergestic orange fuse better than the red nano fuse I don't know either, some KTE owner would have to try and report back.  FYI, Tim told me NO caps are in the signal path!

 

As I completed this fun project, he says tongue in cheek, I let the dac run with a signal from HQP continuously .  The first serious listening session was 50 hours in and through Friday night I had completed 150 hours of burn after that session.

 

The dac is better with the mods.  I find myself listening mostly to DSD512 up sampling as it is so organic, so real and so believable so palpable. I call it an audio Hologram.  With the mod the already huge stage expanded a bit more and individual images which already where very present had more weight and and better definition, subtle sounds before part of the mix now came forward with more of a presence themselves.  Bass further tightened up and the leading edges of it had better definition.  The lead performer was more present and the definition of their vocals was heightened in a very real and more believable way, very intimate.  Overall the new presentation was more real more satisfying.  Something you wouldn't be aware of until you hear it.  The mod was worth it, a lot of work and worth it.  I noted as the week went on and the hours increased on the cap very slight improvements.  I am not sure when this cap is done forming I suspect by now it is 99.x% of the way there.

 

I am still learning this dac but I find most of my listening is at DSD512 and I ask myself why is everybody else not hearing this, why is there not universal agreement like the T+A?  All I can come up with is Room and speaker placement within.  My large full range speakers are 8.5' into the room, a luxury most people do not have (see my profile page about my set up, its long a maybe boring but gives you a good idea) so the stage is really large in my room which plays into DSD 512's strengths, even 1024 with its so 3D palpable images.  It just trounces the more 2D, less palpable PCM 1.5 MHz which although very transparent with excellent bass feels like paper cutouts playing compared to full 3D with 512.  DSD256 EC7 is interesting and I need to spend more time there but every time I go back to 512 I relax a bit more.  What is shocking to me is un-up sampled PCM and how good it sounds, I may even prefer it to PCM 1.5MHz as the images are a bit more 3D un up sampled.  I set HQP to none for filters, Dither set to TPDF and sampling rate to 192 (the highest in my library or on Qobuz) music then plays back at its native rate and I like it a lot not 512 a lot, maybe better than PCM 1.5.

 

There are so many ways to play back your music with this dac, I am still learning.  I am having a blast but still learning and loving every minute.  My love of this dac just gets stronger as I listen more.  You cannot go wrong with this dac.  L2 is great but KTE is better.

 

 

 

 

[lpost]

Your work is solid. Well done. 8.5' you lucky son of a gun. Best I can manage is 3.5' and I'm too close to my back wall.

 

My KTE has 763 hours of continuous play save for 3 times I've powered it down for 4-8 hours. If for no other reason than to verify it starts up from cold without issue. I suspect component failures will be exceedingly low on the Holo gear. 

 

For my ear and room we really liking the dynamics and punch of 1.5M PCM. It is very nice however my HQPe hw can't do better than DSD256 5EC which I find to lack the hit of high-rate PCM. Like the middle core muscles have been made smaller/weaker. The treble extension is also softer. It's there but I can't hear as much of it. 54 years with some hearing damage in each ear but different damage in each ear. My setup is asymmetrical too.

 

We all hear and process audio information differently and of course we're all in different spaces. I think this dac is true. What bits go in comes out fundamentally how it was.

[Extreme_Boky]

@Quadman

 

I have been playing with electronics all my life... here's my experience and few suggestions

 

Bypass the 3 switches in the power supply unit:

 

1. The mains switch, as part of the IEC socket at the back of the unit. 

2. The 110/120 <-> 230/240 switches (2 switches) for the transformers (located on the PCB)

 

The wire you use will determine the sound colouring, but in general, the solid core copper / silver (especially ribbons, and especially silver ribbons, annealed) will sound best.

 

The biggest improvement comes from omitting 3 sets of contacts (contact resistances). I definitely deem all 3 switches as bypassable, unless you are planning to move overseas to EU / AUS.

You will notice  a huge improvement.

The 2 capacitors in question are there pretty much just to provide the means of colouring the end-result (the sound) to suit individual needs / the rest of the system / room acoustics..... simply because the long legs on those caps will provide very little benefit in terms of power supply rails decoupling. Here, the quality (type of dielectric, and metallization material/metal used) is more important (quantity is almost irrelevant). So, you do not have to stay with 10uF capacity... lower values will also work, which means that any cap that you like in your system, will work in the exact same way when used in May power supply. You could try silver in oil, silver in wax.... and so on.... I usually solder few little wires as extenders, and then I roll the capacitors until I find one that suits the individual taste and the rest of system/room acoustics best.

 

 

 

[Quadman]

23 minutes ago, Extreme_Boky said:

The mains switch, as part of the IEC socket at the back of the unit. 

Yea that one is super easy to bypass, front panel switch I assume would still allow dac to go to standby if needed, I would assume. I would also still have the fuse in circuit.

 

23 minutes ago, Extreme_Boky said:

The 110/120 <-> 230/240 switches (2 switches) for the transformers (located on the PCB)

embarrassed I didn't do that, head was focused on match KTE.  And yes it would make a nice uptick in SQ.  That I assume would also be relatively easy to do.  I'll look into it, Thanks!

 

23 minutes ago, Extreme_Boky said:

So, you do not have to stay with 10uF capacity... lower values will also work, which means that any cap that you like in your system

I almost did exactly what you suggest, but in the end I found those really good Auria caps and so far I am really pleased.  Your thought on soldering small wire leads in the PCB to roll was exactly my thought as well should I decide to roll caps.  Remember according to Tim, NO capacitor is in the signal path.  Not even that Teflon 1uF, 1000Vdc near the output. (it replaced the Mundorf)

[Extreme_Boky]

Yes, the front panel switch will place the May DAC in standby mode - the back panel mains switch is not really needed. The mains cable can be unplugged, if required....

 

Auricap XO is my favourite capacitor, at the given cost. It is neutral, not bloomy, not warm, not wobbly... has a firm rendering of bass frequencies and an extended top range. It will provide firmer sound at either end of audio spectrum, compared to those Vishay MKP's. I expect that separation and fine details 'rendering will also improve a fair bit. (This is what Kitsune does...for a pretty penny + the additional shielding of R-2R PCB's; it may be that they are also using the exact same Auricaps). However, you can "compensate" by bypassing the switches I mentioned before, which will bring great improvement overall. 

 

The capacitors we are talking about (you replaced) are used to de-couple the power supply rails (I think they are analogue power supply rails, but I am not 100% sure... will be once I decide to "go in" myself), i.e. they are not used for sound (audio) coupling. 

[Toni-Mang]

@Quadman thank you very much of sharing your method. This is very interesting and a large enrichment of this thread for me. Personally, I am not sure, how much added value i get from the cabling (and material) in front of the transformers, which are still plugged in and out. On the other hand, the vishay caps are specwise and technical very good caps. In this position, imho they do their job well. Your approach is very interesting for me, because i thought, I might tweak the May as you did...but soundwise i feel no need actually...and my technical reservations are the reason, why i bought the L1.

[Quadman]

14 hours ago, Extreme_Boky said:

The capacitors we are talking about (you replaced) are used to de-couple the power supply rails (I think they are analogue power supply rails

Tim said they had something to do with the power supply and I believe you are correct that they are analog.  I sure love the Auricap so far which may push me to not roll capacitors.  Getting to the undersides of the boards to solder is a lot of part removal, not difficult but tedious and one has to keep their wits about themselves.

 

For the 2 switches on the PCB (voltage selector) I would probably want a re-work station to safely remove.  I used an old solder removal iron with a bulb to suckup solder for the cap and connector removal.  Those solder pads are a bit larger for this switch.  Should be worth doing tho.

9 hours ago, Toni-Mang said:

I am not sure, how much added value i get from the cabling (and material) in front of the transformers, which are still plugged in and out. On the other hand, the vishay caps

The AC cabling I added probably does not add much and due to the work involved I decided not to remove those quick connectors the O type transformers use to connect to the PCB.  I do think what @Extreme_Boky said about getting the switches out of the signal path does have real merit and would definitely be audible and probably take my dac above the KTE level now.  Of course then your dac is stuck, for resale, in a same voltage country as you are currently in.

[mrkoven]

I am looking into a streamer that doesn't have USB output. Currently happy w/ USB on the May but I would have to switch to AES if I went with this streamer. Would love to hear people's experience w/ AES vs USB. Thanks

[Miska]

On 4/18/2021 at 5:51 PM, Quadman said:

Do you still need to set HQP at -3 if the signal is already -6?

 

Yes, you need to consider difference between digital levels and analog levels...

[Quadman]

40 minutes ago, Miska said:

Yes, you need to consider difference between digital levels and analog levels...

Thanks Jussi, I realized after I wrote it that its the May that lowers DSD 6 dB, in the PC and HQP there is no DSD output penalty.

[Miska]

17 minutes ago, Quadman said:

Thanks Jussi, I realized after I wrote it that its the May that lowers DSD 6 dB, in the PC and HQP there is no DSD output penalty.

 

And it is not straightforward -6 dB, but just 6 dB difference in reference levels. Because DSD can momentarily go to +3.15 dB (DSD), while PCM cannot exceed 0 dBFS. So the level difference is just 3 dB at maximum possible peak level.

 

[Quadman]

2 minutes ago, Miska said:

So the level difference is just 3 dB at maximum possible peak level.

ah that would explain why I ended up with my HQP min vol at -4 as I used to very occasionally see limiting when it was -3.  Until the Holo, 100% of my streaming was DSD512 via T+A.

[dkdali]

Got my KTE edition yesterday and was foot tapping all evening. Very clear improvement right out of the box compared to my RME ADI-2 fs.

Left it playing during the night and next day and found to my surprise that after being on for about 24 hours it is so hot i cant keep my hand on the side panels....

Is that normal?

i do have my Innuos Zenith on top of it but I didn’t expect it to get that hot. So now i have turned it off and will write Holo if that is normal...

[lpost]

Yes, it's normal. Give it space to breathe. I would not put anything on top of it. Try it on top of the Zenith. You should be able to keep your hand on it but not want to for long. Mine is ~100F/38C on the top surface.

[dkdali]

10 minutes ago, lpost said:

Yes, it's normal. Give it space to breathe. I would not put anything on top of it. Try it on top of the Zenith. You should be able to keep your hand on it but not want to for long. Mine is ~100F/38C on the top surface.

Thanks for the quick response!

And good to know it’s normal though quite surprising for a dac!

Guess i need to rearrange my gear again tomorrow 😉