Motu 8A Measurements

[pos]

The trim to act on ESS internal digital volume control.
Every dB you trim there will be deduced to the DNR. There is no magic: ESS volume control is no different than any (properly implemented) digital volume control in front of it (in the Motu internal mixer, or in the PC)

 

[dwkdnvr]

It may be that the 112dB spec of the UltraLite mk4 is of the stereo monitor outs where one ES9016S chip is used while the 8 line outputs have another single ES9018S chip to cover all 8 outputs.
Not expecting better performance from the Ultralite mk4 for the line outputs, just hoping they measure the same as the 8A but at a lower cost

Did you ever circle back and get measurements of the Ultralite Sabre version? It really offers a whole lot of functionality for the money - if you switch the ADAT port to spdif, you can get 14 channels out which is about the max I can see myself needing even if all my ambitious plans come to fruition (which they probably won't)

 

[jaynyc]

@March Audio Hi - did you ever measure the headphone output impedance of the Motu 8A?

 

[March Audio]

Hi

Actually no I haven't, but from a subjective POV FwIW I don't recall it being anything write home about, but don't trust subjective memory

I will see if I can measure it later today

 

[March Audio]

@jaynyc

OK just for reference 8A in balanced loopback. Hits 111dB THD +N

Headphone out (unbalanced) to balanced input. THD +N 98dB

 

[restorer-john]

OK just for reference 8A in balanced loopback. Hits 111dB THD +N

That's fabulous performance.

 

[March Audio]

That's fabulous performance.

Its loopback so I am not sure which is the limiting factor, the A to D or D to A.

I seriously considered getting the RME ADI2 pro for measurement but then wondered if it was actually any improvement.

 

[restorer-john]

I seriously considered getting the RME ADI2 pro for measurement but then wondered if it was actually any improvement.

It's pretty much at the 'academic' interest only level isn't it?

So you could theoretically set up 4x2 inputs, externally fixed and padded down at 10x ranges to cover a decent voltage range for power amp tests and switch between them via the Motu's software?

 

[dwkdnvr]

Becoming harder to ignore this as a great candidate for a multi-channel interface for an active speaker project. I've been going down the road of interfaces with one or more digital outputs to take advantage of the well-measuring DACs we see around here, but I'm increasingly thinking there may not be any reason to go more complicated/expensive than this.

Wish I was more confident in Linux support; I'm currently running JRiver on Windows, but would prefer the option to go with a Linux approach. It seems that some of the AVB interfaces do work, (and should, since apparently they're class-compliant) but real-world reports are pretty scarce.

 

[jaynyc]

@March Audio thank you for this. Is there a way to know the headphone output impedance of the 8A similar to how Amir does it, ie: measured in ohms?

 

[March Audio]

@March Audio thank you for this. Is there a way to know the headphone output impedance of the 8A similar to how Amir does it, ie: measured in ohms?

Yes I will try to get time to measure this but I can already tell you it's relatively high. When plugging in my Oppo PM1 which are a 32 Ohm almost purely resistive load, the open circuit output voltage dropped from something like 600mV tp 400 mV!

 

[somebodyelse]

Wish I was more confident in Linux support; I'm currently running JRiver on Windows, but would prefer the option to go with a Linux approach. It seems that some of the AVB interfaces do work, (and should, since apparently they're class-compliant) but real-world reports are pretty scarce.

There's a long thread at LinuxMusicians that's primarily about the Ultralite AVB, but the problems appear to be common to the whole AVB range as other models get a mention from time to time. My reading of it is that unfortunately it's still a bit of a lottery, but you might be lucky. The windows driver doesn't use the class compliant mode, so it only gets tested with Apple stuff as MOTU don't support linux. The most interesting thing to come out of it may be work on an AVB driver for ALSA. From what I know of AVB I assume you'll need an ethernet port that supports hardware timestamping to connect directly to the 8A or other AVB device. That's relatively cheap ($20-30) if the one on your motherboard doesn't support it already. A common suggestion from existing MOTU owners is to buy RME as they actually support linux.

 

[March Audio]

Becoming harder to ignore this as a great candidate for a multi-channel interface for an active speaker project. I've been going down the road of interfaces with one or more digital outputs to take advantage of the well-measuring DACs we see around here, but I'm increasingly thinking there may not be any reason to go more complicated/expensive than this.

Wish I was more confident in Linux support; I'm currently running JRiver on Windows, but would prefer the option to go with a Linux approach. It seems that some of the AVB interfaces do work, (and should, since apparently they're class-compliant) but real-world reports are pretty scarce.

I use Roon > Acourate > Motu for my dsp active speakers. It's rock solid. Can I ask what is your reason for wanting to go Linux?

 

[LTig]

Its loopback so I am not sure which is the limiting factor, the A to D or D to A.

I seriously considered getting the RME ADI2 pro for measurement but then wondered if it was actually any improvement.

I'd say no. See these balanced (XLR) loop back measurements of my RME ADI-2 PRO fs at 48 kHz samplerate, with 256 k FFT size and 32 averages, at +24 dBu and + 4 dBu setting:

Your Motu 8a has slightly better numbers as far as I can see.

 

[dwkdnvr]

I use Roon > Acourate > Motu for my dsp active speakers. It's rock solid. Can I ask what is your reason for wanting to go Linux?

Mostly that I have a particular system design in mind, and I know (more or less) how to accomplish this with Jack/Brutefir, but I'm not sure it'll fly with a Windows based system since it doesn't fit the 'headless appliance' application particularly well. It basically comes down to input switching - one place I'd like to use it is in our living room where we have DirecTV, FireTV, chromecast and a music player (currently squeezebox, but would almost certainly turn into an internal source for a PC based system). Since there is nothing currently on the market (that I know of) like the old Meridian 518 digital preamp, I'm faced with switching sources as part of the DSP system - ideally digital (e.g. something like Minidsp UDIO-8), but the 8A analog inputs are good enough not to worry too much about it as long as the signal is strong enough. Input switching is pretty straightforward with Jack - just run the external inputs into the soundcard and change the routes with Jack - easy to do with a bit of code and LIRC. JRiver doesn't handle soundcard inputs in most cases (since it has to open input/output as separate clients, and most drivers don't handle that too well unless you run through the windows mixer). I haven't looked at Acourate convolver to see whether it might be a better fit. In any event though, I'm not sure I trust windows to just boot up and play sound without some level of interaction / putzing, which would never fly in a 'used by the family' system - it has to 'just work'

Ultimately, I may have to just re-think the overall system. Something like the MiniDSP SHD / Studio (if I can live with a 2-way setup with a passive M/T xover) or the Hypex DSP amps might be a better solution for the living room where it has to be 'wife capable'. This is potentially even more reasonable given the experimenting I managed to do yesterday. More detail here https://www.audiosciencereview.com/...ital-active-speaker-systems.7731/#post-196090 but the upshot is that the results using a basic LR4 xover at 150Hz with no other EQ were so good that maybe I shouldn't over-think the problem.
If I revert to using the TV to switch HDMI sources, I'm left with only optical from the TV and whatever the music player is for inputs, and that should be a manageable problem - even the NanoDigi can handle 2 inputs.

 

[dwkdnvr]

There's a long thread at LinuxMusicians that's primarily about the Ultralite AVB, but the problems appear to be common to the whole AVB range as other models get a mention from time to time. My reading of it is that unfortunately it's still a bit of a lottery, but you might be lucky. The windows driver doesn't use the class compliant mode, so it only gets tested with Apple stuff as MOTU don't support linux. The most interesting thing to come out of it may be work on an AVB driver for ALSA. From what I know of AVB I assume you'll need an ethernet port that supports hardware timestamping to connect directly to the 8A or other AVB device. That's relatively cheap ($20-30) if the one on your motherboard doesn't support it already. A common suggestion from existing MOTU owners is to buy RME as they actually support linux.

Yeah - that thread was my original source of info on the AVB interfaces, but I hadn't visited in a while. The potential for an AVB driver is very intriguing. It does sound like many are getting reasonable stability with older firmware in USB mode, but that still seems a bit dicey to rely on. I have a Behringer 0204HD that I can use for prototyping/experimenting so I'm not in an immediate hurry, so will watch the AVB development with interest.

 

[sfabio]

As a follow-up, I have been testing a 24Ao today, and I must say I am a bit disappointed so far:

1. there is a turnoff noise with DC energy (meaning I cannot connect a compression driver directly, without a protection cap)
2. there are distinct noises on the outputs, not random (computer-like noises if you see what I mean) and quite audible at 2m with a 100dB/2.83V@1m woofer and a 26dB gain amplifier
3. these noises tend to reduce when audio is played, albeit I could not find a clear step: it seems to go and then come back.
4. these noises remain when no USB or ethernet cable is plugged.
5. changing sampling frequency takes a few seconds, even with the last firmware (probably due to the internal DSP processing re calibrating for the new sampling frequency)
6. the output trim settings are not really useful as they are done inside the DAC, meaning the 123dB(A) DNR becomes 111dB(A) when targeting a 8dBu max output level. Needless to say noises of points 1, 2, 3, 4 are not reduced by these trims.

I wonder if this particular unit might be faulty, but all output present the same noises, at any sampling frequency.
I also wonder if the motu 8A has the same issues (point 1, 2, 3, 4).

At this point I am not very interested in testing this unit further (linux configuration, toslink input, etc.)...

Sorry to resume this old thread, only to post my experience:
Motu UltraLite mk4 with ESS dac → diy NC500MP amplifier → Eighteeensound ND3ST driver with XT1464 horn without passive crossover
same issues like you, except there is even a turnon noise. The noise is clearly audible at the listening position (2,5m).
With an RME Babyface pro only a subtle hiss over amplifier noise
I can live with turnon and turnoff noise, but these crackling noise are very annoying... Maybe power supply issues?
I'll be forced to find an affordable substitute.
Fabio

 

[wadec22]

Sorry to resume this old thread, only to post my experience:
Motu UltraLite mk4 with ESS dac → diy NC500MP amplifier → Eighteeensound ND3ST driver with XT1464 horn without passive crossover
same issues like you, except there is even a turnon noise. The noise is clearly audible at the listening position (2,5m).
With an RME Babyface pro only a subtle hiss over amplifier noise
I can live with turnon and turnoff noise, but these crackling noise are very annoying... Maybe power supply issues?
I'll be forced to find an affordable substitute.
Fabio

What OS?

 

[sfabio]

What OS?

Windows 10 64 bit, but it's an hardware issue, it's still present with soundcard disconnected from pc.

 

[pos]

Sorry to resume this old thread, only to post my experience:
Motu UltraLite mk4 with ESS dac → diy NC500MP amplifier → Eighteeensound ND3ST driver with XT1464 horn without passive crossover
same issues like you, except there is even a turnon noise. The noise is clearly audible at the listening position (2,5m).
With an RME Babyface pro only a subtle hiss over amplifier noise
I can live with turnon and turnoff noise, but these crackling noise are very annoying... Maybe power supply issues?
I'll be forced to find an affordable substitute.
Fabio

Hello

The RME can be configured for a +4dBu max output level, which of course lowers its noise compraed to its +19dBu mode.
IIRC the Motu output level is around +20dBu, and the output elvel can only be adjusted in the digital domain (a big step back compared to previous models, that the better DAC chip does not compensate for).
Did you compare the RME in +19dBu mode with the motu?

I don't know for the UltraLite, but the 24Ao sure seemed more noisy than many other DACs I have tried, compared to its output level (S/N), and what I heard was not uncorrelated random noise...

In any case a +20dBu DAC with a +26dB gain amp and a compression driver is calling for trouble
What you can do is introduce some form of analog volume reduction along the line: either between the amp and dac (a passive reduction like an xlr pad should work given the amp has a buffer), or between the amp and compression driver (lpad, or an autoformer).

I guess you could also try removing the buffer on your nc500 build and see if that works well with your DAC