Chords New M -Scaler

[Crustacean]

A question about a hypothetical M-Scaler scenario has arisen for me. I apologize if this inquiry sounds like the pursuit of an unadvisable option or has already been discussed (I tried a quick search of "BNC" but didn't see this raised yet here), and I also apologize for not having the technical knowledge to have much of an idea yet what the answer might be. Also, I realize there's some debate over how much benefit may or may not result when using an M-Scaler with non-Chord DACs. Furthermore, I realize there are whole other discussions regarding cable variables such as quality and length. Given those caveats, the basic thing I'm wondering is:

 

1. With two non-Chord DACs located close to each other, e.g. in adjacent rooms like a listening room and then an office for headphone use, could a single M-Scaler feed them both via the two BNC outputs, with one cable each?

 

2. If so, what's the maximum advisable distance a decent BNC cable could run from the M-Scaler to one of the DACs; would something like 10 meters be way too long for the M-Scaler's output or the cable's integrity?

 

If this could work without substantial SQ degradation over the longer cable run, presumably one needn't buy a separate set of streaming endpoint hardware for the second DAC, which might make the price of a single M-Scaler a somewhat more interesting proposition, even without Chord DACs in the mix.

 

But maybe this is a fantasy that isn't technically feasible, because the M-Scaler output's signal integrity would be quickly defeated by anything longer than a short BNC run between components in the same hifi rack. Thanks for any thoughts, even just to set me straight.

[seeteeyou]

Almost all non-Chord DACs could only accept 192kHz via (single) coaxial BNC / RCA inputs. As a result, we're only able to utilize 0.25 million taps instead of 1 million. Maybe not worth all that money?

 

There's something like ComTrue CT7601 from Taiwan that could accept 768kHz via single coaxial inputs, though most likely it's only limited to DIY stuff.

 

They're still working on DAVINA with USB output, that's what I'm waiting for myself.

[Arpiben]

Record the M Scaler BNCs' Outputs and later convert to USB, or else, and play where you want....

[Crustacean]

1 hour ago, seeteeyou said:

Almost all non-Chord DACs could only accept 192kHz via (single) coaxial BNC / RCA inputs. As a result, we're only able to utilize 0.25 million taps instead of 1 million. Maybe not worth all that money?

 

There's something like ComTrue CT7601 from Taiwan that could accept 768kHz via single coaxial inputs, though most likely it's only limited to DIY stuff.

 

They're still working on DAVINA with USB output, that's what I'm waiting for myself.

 

Thanks, I need to learn more about the Davina project.

 

And correction: Maybe I should have posed my question as: using short Toslink for one DAC and a longer BNC cable via the S/PDIF for the other. According to the M-Scaler manual, at least, the dual BNC outputs should not be used at all for non-Chord DACs.

[greenleo]

2 hours ago, Crustacean said:

 

Thanks, I need to learn more about the Davina project.

 

And correction: Maybe I should have posed my question as: using short Toslink for one DAC and a longer BNC cable via the S/PDIF for the other. According to the M-Scaler manual, at least, the dual BNC outputs should not be used at all for non-Chord DACs.

The Davina will be a long wait I believe.

[kennyb123]

14 hours ago, seeteeyou said:

Almost all non-Chord DACs could only accept 192kHz via (single) coaxial BNC / RCA inputs. As a result, we're only able to utilize 0.25 million taps instead of 1 million. Maybe not worth all that money?

 

 

My Hugo TT accepts up to 384kHz via a single coaxial BNC.  But you could be right that most other DACs only go up to 192kHz via coaxial BNC.

 

A very nice improvement can be heard from just the half-million taps.   I'm expecting a bigger bump though after my TT2 finally arrives.

 

[Crustacean]

On 3/19/2019 at 1:47 AM, kennyb123 said:

 

My Hugo TT accepts up to 384kHz via a single coaxial BNC.  But you could be right that most other DACs only go up to 192kHz via coaxial BNC.

 

A very nice improvement can be heard from just the half-million taps.   I'm expecting a bigger bump though after my TT2 finally arrives.

 

Great to hear that you're getting such an improvement at 384kHz.

 

I was intrigued, and a bit perplexed, to see that this reviewer felt that even at 192kHz into the Totaldac, for example, the M-Scaler at presumably only 0.25 million taps was delivering most of its goodness already; the reviewer called the difference when adding the TT2 with the full million taps only "subtle" and he liked the M-Scaler "regardless of DAC pairing." This was different from the conclusion in the review here at A/S, which found much less benefit pairing the M-Scaler with the Ayre QX-5 Twenty than with the TT2. I'm wondering if the difference could have to do with the Totaldac running in non-oversampling mode (I don't know about the QX-5 Twenty) and thus passing on more of the effect of the M-Scaler. If that's the case, perhaps non-oversampling DACs pair better even at only 192kHz, I don't know.

 

I'd still be curious if anyone has thoughts about the maximum BNC cable length that would be advisable running out of the M-Scaler.

[barrows]

8 hours ago, Crustacean said:

Totaldac running in non-oversampling mode

Bingo!  I suspect the above is correct.  The Ayre DAC runs proprietary oversampling in an FPGA at a 16x rate, and if one inputs 176.5/192 kHz into the Ayre it will still apply its own oversampling to get the rest of the way to 705.6 and 768 kHz.  

[kennyb123]

On 3/19/2019 at 9:52 PM, Crustacean said:

Great to hear that you're getting such an improvement at 384kHz.

 

(snip)

 

I'd still be curious if anyone has thoughts about the maximum BNC cable length that would be advisable running out of the M-Scaler.

 

I'm surprised that the Hugo TT has responded so well to the HMS at a half million taps.  All the improvements reported at a million taps are present here too.  I think I've been most amazed with the improvement to bass articulation and tone.

 

As far as BNC cable length, I believe the standard specifies a max length of 10m.  But your best bet is to post to the HMS thread on HeadFi on the with a callout to Rob Watts.  He's good about answering questions there.  

 

HMS thread:

https://www.head-fi.org/threads/hugo-m-scaler-by-chord-electronics-the-official-thread.885042

 

[mansr]

9 minutes ago, kennyb123 said:

As far as BNC cable length, I believe the standard specifies a max length of 10m.

If you're referring to the S/PDIF standard, there is no maximum cable length specified. As long as the signal levels and rise/fall times are within specified limits, the cable can be however long you wish. With any specific cable type, its loss characteristics will of course determine a max length beyond which the signal requirements can no longer be met.

[incus]

On 3/20/2019 at 6:06 AM, barrows said:

Bingo!  I suspect the above is correct.  The Ayre DAC runs proprietary oversampling in an FPGA at a 16x rate, and if one inputs 176.5/192 kHz into the Ayre it will still apply its own oversampling to get the rest of the way to 705.6 and 768 kHz.  

Yes - I agree this is most likely what is going on. Waiting for my HMS to arrive to pair with my Holo Spring II for this very reason. I can switch between OS and NOS mode to test this theory. Then once my TT2 comes I will throw that into the mix as well.

[incus]

Also - wondering if anyone here has tried the Blaxius^2 BNC cables with their HMS?

[matthias]

59 minutes ago, incus said:

Yes - I agree this is most likely what is going on. Waiting for my HMS to arrive to pair with my Holo Spring II for this very reason. I can switch between OS and NOS mode to test this theory. Then once my TT2 comes I will throw that into the mix as well.

 

What about software OS vs. HMS with the Spring2?

Thanks

 

Matt

[Miska]

1 hour ago, incus said:

Yes - I agree this is most likely what is going on. Waiting for my HMS to arrive to pair with my Holo Spring II for this very reason. I can switch between OS and NOS mode to test this theory. Then once my TT2 comes I will throw that into the mix as well.

 

Spring's OS (AKM chip) is not so much to compare against... It is best running in NOS mode with external digital filter and modulator. If you lack capability of running good external digital filter and modulator, then it is certainly better than nothing.

 

[Miska]

Note that if you are running Spring in PCM mode, to get best out of the Spring you need to send it about 20-bit (or less with noise shaping) to keep it linear.

 

So it is very important to configure correct word length to your upsampler to achieve full resolution out of Spring.

 

Here's Spring 2 doing 1 kHz -120 dB at 20-bit noise shaped at 1.4112 MHz sampling rate:

 

And here the same, but with 24-bit data and no noise shaping:

 

Even with 16-bit 1.4112 MHz with fairly aggressive noise shaping you still get good results:

[matthias]

37 minutes ago, Miska said:

Note that if you are running Spring in PCM mode, to get best out of the Spring you need to send it about 20-bit (or less with noise shaping) to keep it linear.

 

So it is very important to configure correct word length to your upsampler to achieve full resolution out of Spring.

 

Here's Spring 2 doing 1 kHz -120 dB at 20-bit noise shaped at 1.4112 MHz sampling rate:

 

And here the same, but with 24-bit data and no noise shaping:

 

Even with 16-bit 1.4112 MHz with fairly aggressive noise shaping you still get good results:

 

Very interesting,

can you please post some with converting to DSD as well?

Thanks

 

Matt

[Miska]

17 minutes ago, matthias said:

Very interesting,

can you please post some with converting to DSD as well?

 

Output level is quite a bit lower at DSD inputs due to different reference level and other technical reasons, so in measurements the relative background noise comes up by about 8 dB in same test. Analog background noise level is the same in both cases, but "0 dB" output signal level is different. Otherwise looks the same, for DSD256:

 

Output level here at 0 dBr with PCM is 3.8Vrms while with DSD it is 1.7Vrms (7 dB difference). Since DSD spec allows momentary +3.15 dB levels it is good that DACs have headroom to accommodate for this (ESS DAC chips don't, AKM DAC chips in DSD Direct do by having -3.5 dB DSD output level). But when testing with continuous signals I don't go over 0 dB.

 

[Miska]

DAC's could apply 6 dB gain to DSD sources like ESS does. And then analog background noise vs 0 dB reference peak level is the same. But that has a risk when source contains momentary overmodulated DSD content (many SACDs do have that) where the output voltage would go +3 dB higher than maximum output with PCM and could clip analog stages somewhere on the way...

 

For PCM, 0 dBFS is a hard limit that cannot be exceeded. For DSD, 0 dB is not hard limit and it is just defined as 50% modulation index. And the maximum allowed momentary modulation is 75%. Just for reference, 100% modulation is just plain DC at the rail voltage...

 

[Crustacean]

14 hours ago, incus said:

Yes - I agree this is most likely what is going on. Waiting for my HMS to arrive to pair with my Holo Spring II for this very reason. I can switch between OS and NOS mode to test this theory. Then once my TT2 comes I will throw that into the mix as well.

Terrific, would love to hear what you find. (My DACs can also run in NOS mode, which is what is pushing my curiosity on a possible HMS pairing even though they are not Chord machines.)

[incus]

18 hours ago, Miska said:

Spring's OS (AKM chip) is not so much to compare against... It is best running in NOS mode with external digital filter and modulator. If you lack capability of running good external digital filter and modulator, then it is certainly better than nothing.

It was never my belief that OS mode on the Spring would sound better than the HMS into the Spring in NOS mode. I listen almost exclusively in NOS mode on the Spring and ordered the HMS for the TT2 - I just plan to play around a bit when it all arrives...

 

Thanks for all the info about the Spring, btw. Right now the only software upsampling I can do is either with ROON DSP or A+'s iZototpe 64-bit SRC. I currently do not have HQPlayer so I don't think I have the bit rate setting and noise shaping capabilities you mention. I actually find the Audirvana upsampling to sound quite good going into the Spring in NOS mode. And I prefer Roon DSP upsampling into the Spring in NOS mode to straight redbook in either NOS or OS...

[Miska]

15 minutes ago, incus said:

Thanks for all the info about the Spring, btw. Right now the only software upsampling I can do is either with ROON DSP or A+'s iZototpe 64-bit SRC. I currently do not have HQPlayer so I don't think I have the bit rate setting and noise shaping capabilities you mention. I actually find the Audirvana upsampling to sound quite good going into the Spring in NOS mode. And I prefer Roon DSP upsampling into the Spring in NOS mode to straight redbook in either NOS or OS...

 

Then it is better to stick to DSD256 output from those two instead of sending PCM there...

[matthias]

2 hours ago, incus said:

I actually find the Audirvana upsampling to sound quite good going into the Spring in NOS mode. And I prefer Roon DSP upsampling into the Spring in NOS mode to straight redbook in either NOS or OS...

 

So in the absence of HQP would it be correct to say that you prefer upsampling with A+ to upsampling with Roon (both into Spring NOS mode)?

 

Matt

[incus]

It's not apples to apples because although both come through my modded and reclocked SOtM component/cable chain, A+ comes through my iMac while Roon comes through my modded and re-clocked SOtM server -- so OVERALL the Roon route sounds quieter and more refined because of the difference in 'source' quality... (iMac sounds better streaming Tidal through A+ than playing from local files - playing off the iMac's hard drive is a whole level worse than any other playback solution I have.) Nonetheless, I am impressed by how good A+ upsampling to DSD256 from the iMac (again streaming Tidal, NOT from local files) through the SOtM sMS-200 Ultra SE in DLNA mode into the Spring on NOS sounds. Happy doing that while I wait for my server to get some more tweaks. Again I use a lot of SOtM product so the overall sound through all the filtering and reclocking is similar. Still the nod goes to Roon... It will be very interesting to see what the HMS/TT2 combo does in comparison, as well as using just one BNC output on the HMS to upsample into the Spring...

[kennyb123]

I’ve had my HMS for a little over a month and a half.  My opinion of it has evolved since it landed.  

As I mentioned earlier in this thread, I was impressed with the improvements I heard when I first connected it to my Hugo TT using the OEM BNC cable that Chord provided.  The improvements from the WTA filter at a half million taps were easy to hear and made my music more enjoyable to listen to.  

But did I think the improvements I first heard justified the amount I paid for the HMS?  I didn’t really consider that question initially because I was anticipating the soon arrival of a Hugo TT 2.   The TT2, with its dual BNC input, would allow me to hear the HMS scaling up to a million taps.  There were many who reported that the big leap comes when making that jump.  I think my answer to the question as to whether the HMS justified the price at that point would have been "TBD".

 

But then news came that my TT2 would be delayed, as Chord was having trouble meeting demand. My TT was going to be with me longer than expected, so I stared thinking about making another planned move sooner.  There were many who reported hearing greater than expected improvements from upgrading BNC cables, so I figured I’d delve into this while I still had the TT.  

 

The first BNC cable I tried was a preowned Stereovox HDXV, which I had purchased from an Audiogon seller.  This cable had at one time retailed for $100.  Many years ago I had used the predecessor to this, the Illuminati, which was a very well-regarded cable back in its day, so I was optimistic that I’d hear good things from the HDXV.  Well, it would be an understatement to say that the HDXV exceeded my expectations.  I’m pretty sure that I’ve never heard a cable at this price point bring about such a big level of improvement.  If I would have been asked at this point whether the HMS justified the price I paid, my answer at this point would have been “it might come close to justifying it's price, even when only tapping into a half-million taps, but a better digital cable must be used”.

 

6Moons reviewed the HDXV very favorably back in 2003 and had this to say about it "The [HDXV] remained steadfast in the face of truly challenging material and showed itself to excel in the realm of retrieving minutiae.”  That’s what I heard too.  To be able to fully appreciate what the HMS brings to the table, I believe it’s critically important to use a BNC cable that can pass on all the goodness coming from the HMS.  The OEM cable that Chord provides is not even close to being able to do that, in my humble opinion (at least not when scaling to only a half-million taps).

 

The HDXV did not remain in my system very long.  I’ve been super-impressed with the bang-for-the-buck I’ve gotten from the Shunyata Alpha series, so I just had to hear what would happen were I to use their Alpha S/PDIF cable between the HMS and the TT.  I’m using an Alpha USB upstream of the HMS, so I figured the Alpha S/PDIF would compliment it very well.  My dealer was able to arrange to have a cable sent to me last Friday.  It arrived on Saturday, and roughly within an hour I told my dealer that there was no way that cable was ever going to be leaving my house.  HMS plus the Alpha S/PDIF have brought about the biggest step change from any upgrade I’ve made to my digital source.  Given their combined price, one would hope that would be the case - and it most certainly was.  If you were to ask me at this point whether the HMS justified the price, my answer would be “hell f’n yes - it’s a game changer when used with cabling that’s up the task, even when only tapping into a half-million taps."

 

As impressed as I am with the HMS, I think I may even be more impressed with my Chord Hugo TT.  It’s truly stunning how much its performance was able to scale with the HMS.  I can now honestly say that I’m in no rush to receive my TT2.  What I’m hearing now already far exceeds what I thought was possible to achieve - especially from 16/44.1 files.  Chord is really on a roll right now - and I'm a very happy customer.

[incus]

 Great to get this feedback from the front lines as I await my HMS and TT2 combo as well. Thank you. Could you perhaps describe in a little more detail what the Shunyata brought to the table, besides "minutiae" (which I assume means detail retrieval?)? What is the "goodness" you are hearing, if you don't mind sharing? It would be very helpful to me personally as I contemplate which BNC cables to get for my combo. Cheers (Also has your dealer given you a firm date for the TT2 to arrive?)