Motu M4 - Tear down, bit of internals analysis and few in-house measurements

[mightycicadalord]

M4 has no sub-out, so I'm not sure I understand how a monitor can have it's driver moving forward while the attached sub to move its driver backwards. Also, RCA is connected to V+ and GND, so there's nothing inverted there.

I'll wait for @Daverz to reply here on how he can perceive this inverted polarity, so I can check it myself on my M4 as well. Thanks!

I'm running rca outs to subs and 1/4" outs to monitors, have to flip phase on subs.

 

[trl]

I'm running the same config as yours, so both monitors and sub are in-phase because they are paralleled, otherwise sub's controls are inverting sub's phase with 180 degrees, but not the studio interface.

 

[Daverz]

How did you verified this, please? Have you used a dual-spot scope or something else?

By looking at the impulse response of my speakers in REW. They have a positive going impulse with my other DACs and a negative going impulse with the Motu. The speakers are connected (via my amp) to balanced outputs 1 & 2. The subwoofer is connected to unbalanced outputs 3 & 4.

I've measured the impulse response with both the Motu's microphone input and a Behringer UMC202, so I don't think it's just the input that's inverting.

This guy did a loopback test:

https://www.reddit.com/r/audioengineering/comments/jf2dt3

You're on firmware 2.0? Hmmm, I went to update and see no firmware listed.

Yeah, I noticed that 1.02 is the latest firmware on the web. But 2.00 is what is reported by the Motu driver panel in windows. It's a later revision of the M4.

I know the rca outs are opposite polarity of the balanced outs, idk which is correct polarity. I had to invert the subs I ran off the rca outs. Idk what the signal analysis says above but it's pretty obvious when your monitors are moving forward on a kick and the sub is moving backwards.

Ah, OK, thanks for the confirmation. I can invert the sub and then invert the input signal to get both outputs non-inverted. It's not absolutely necessary as digital room correction can sort it all out anyway, but it's just the sort of inconsistency that drives me nuts.

M4 has no sub-out, so I'm not sure I understand how a monitor can have it's driver moving forward while the attached sub to move its driver backwards. Also, RCA is connected to V+ and GND, so there's nothing inverted there.

I'll wait for @Daverz to reply here on how he can perceive this inverted polarity, so I can check it myself on my M4 as well. Thanks!

I do the crossover in software with CamillaDSP. The sub is connected to RCA outputs 3 & 4. The amp is connected to TRS outputs 1 & 2. I'm guessing the inversion is in the balanced output circuit. I tested with 2 different TRS to XLR cables from different manufacturers, though I suppose it's possible both have hot and cold pins reversed.

 

[trl]

As long as both speakers are sharing the same phase there should be no audible differences. However, if you're saying that RCA is having an inverted sound vs. the TRS plugs, and you're using the RCA for the subwoofer, then you should definitely change the polarity on the sub with 180 degrees from sub's DSP.

https://www.audiocheck.net/audiotests_polaritycheck.php might help in checking if the monitors and subwoofers are properly in-phase. I personally have my sub facing the monitors, so I setup the sub with 180 degrees phase shift.

 

[trl]

Seems that @Daverz is right after all, thanks! Top half is the recorded signal, while bottom half is the original taken from audiocheck.net.

Recording the TRS output of M4 - both signals are in-phase​

Recording the RCA output of M4 - signals are 180-degrees out of phase
​

Not quite an issue, but if someone is connecting subwoofers on the RCA-output and studio monitors on the TRS-output then indeed the sub will need to be adjusted from it's pahse-adjust knob.

 

[mightycicadalord]

Wonder if this is intentional, a limitation, or just overlooked. I wanna ask motu but their support link is broken.

 

[trl]

There are many amplifiers out there having inverted outputs; human's hearing cannot perceive this, unless the connected speakers are out-of-phase like in the scenario from one of the above posts.

For some reasons (or perhaps by ignorance/overlooked) Motu decided to tie the output buffers from the RCA plugs to the inverted pin of DAC's output (or maybe output buffer inverted input was used).

However, as long as we're all aware of this particularity I think we're fine; after all, most folks are not using studio interface's RCA-out for subwoofers and if they do then they'll need to switch sub's 180-degrees knob. I personally use the most common setup out there: TRS-out connected to subwoofer, then from subwoofer I get the monitors connected.

 

[mightycicadalord]

Well another scenario I'll have to account for is running a little 4 channel headphone amp to some dt770's. We have three and it's just for my little trio. We all noticed something was kind of off so next time we meet I'll have to switch polarity in the DAW and see what happens. Getting monitored drum signal out of phase with some acoustic drum noise bleed into the headphones is strange indeed.

 

[Daverz]

Seems that @Daverz is right after all, thanks! Top half is the recorded signal, while bottom half is the original taken from audiocheck.net.

I was measuring with a microphone into XLR input 1 and seeing the TRS outputs inverted and the RCA outputs non-inverted, so perhaps the mic input is also inverting. I'll try similar tests.

EDIT: well, I just managed to get myself confused, and it's nearly 1 AM. I see both RCA and TRS being inverted into inputs 3 & 4, but neither inverted into inputs 1 & 2. Test signal is one second of silence followed by a sine wave. This allows you to see the leading edge of the recorded sine wave.

Each recording is TRS output on left channel and RCA on right channel. I use a TS (that is only tip and sleeve) to RCA cable to connect RCA to input. First screenshot is output 1 to inputs 3 & 4. Second screenshot is output 1 to inputs 1 & 2.

 

[trl]

@mightycicadalord I am not sure I fully understand your headamp scenario, but I know for sure that there are several headamps out there with phase inverted and they measure very well. Actually, some headamp designer are inverting the phase on purpose, because some folks say that it 'sounds better". However, https://www.elprocus.com/difference-between-inverting-and-non-inverting-amplifier/ explains the difference between inverting and non-inverting amplifiers.

@Daverz, I observed that only RCA-outputs are inverting the sound, this isn't related to the four inputs. However, this isn't a bug, but more like a particularity of M4's design that we may need to take it into account more or less.

 

[mightycicadalord]

@mightycicadalord I am not sure I fully understand your headamp scenario, but I know for sure that there are several headamps out there with phase inverted and they measure very well. Actually, some headamp designer are inverting the phase on purpose, because some folks say that it 'sounds better". However, https://www.elprocus.com/difference-between-inverting-and-non-inverting-amplifier/ explains the difference between inverting and non-inverting amplifiers.

@Daverz, I observed that only RCA-outputs are inverting the sound, this isn't related to the four inputs. However, this isn't a bug, but more like a particularity of M4's design that we may need to take it into account more or less.

The issue with the phones is I'm taking signals created in the room, like drums and guitar amps, pa speakers, what have you and send them to people in headphones with the phase inverted. There's a good bit of bleed that gets through the phones, it doesn't mix with the headphone signal as well when inverted. I'm not really worried about phase when dealing with just one signal and one thing reproducing it, if I have an issue it's because I have two or more signals not in phase with each other.

Like If I'm tracking drums, it's super duper obvious on the snare, inverting the phase of the output in my daw gives me "in phase" signal with the signal created in the room. Snare sounds totally different really. My headphone amps coming out of the rca's (have to unplug subs but thats ok).

 

[trl]

Not sure if it will help, but I guess you could DIY an TRS to TRS+RCA cable splitter and fed the devices you need this way. Best would probably be to purchase an external mixer that can deal with all the inputs/outputs you may need.

I personally don't think Motu can resolve this with a firmware update, but let's wait and see what the next firmware upgrade will give us.

 

[trl]

MOTU support recommended me to file an issue to their techlink portal, so: "MTL169142 - Playback from the RCA outputs of M4 inverts the signals with 180-degrees (audio is out of phase)" got created.

 

[dshreter]

I have an M2 and just recently set it up with powered monitors off the TRS and subwoofer off RCA. EQing my subwoofer, I found that I needed the phase set at 180 degrees for the best response…. I thought it was odd but didn’t think about it any further since the response looked good.

Now I know why!

 

[umbral]

New revision of Motu M4 (maybe the new M2 too) use NXP USB controller, not Xmos, and currently only has Win10 / Mac drivers.
Unfortunately, support for Win7 (and Android, with driver from USB audio player / recorder?) still missing.

And DAC ESS9026pro instead of ESS9016s in prev. revision .

Dear all, wanted to confirm that MOTU obviously now ships the new revision of the m4, as first made public by Evgeniy in the post above. have inspected a newly bought unit, seems the same as above.
win10 drivers (asio and mme/ks/dx) seem not to play so well on a ryzen mobile 4800u apu notebook, will have to do some many more tests. a rme fireface uc plays stable as a rock on the same notebook.
Evgeniy: did you encounter driver problems (=cracks and pops) with this new revision of the m4, if tested on win10?
Cheerio!

How can i tell which version do i have ?

Which one is better the old one or the new one ?

DAC ESS9026 pro vs ESS9016s ?

https://www.esstech.com/wp-content/uploads/2021/02/ES9026PRO_Datasheet_v3.4.pdf

https://www.esstech.com/wp-content/uploads/2021/02/ES9016S_Datasheet_v1.91.pdf

NXP USB controller vs not Xmos

 

[AnalogSteph]

How can i tell which version do i have ?

Check the installed firmware version. V1.xx should be old, v2.xx should be new.

 

[trl]

Which one is better the old one or the new one ?

DAC ESS9026 pro vs ESS9016s ?

Based on pure specs ES9026 is better than ES9016s on the bitrate playback:
- PCM 768 kHz vs. PCM 384 kHz
- DSD 1024 (DSD 45.2MHz) vs. DSD 256 (DSD 11.2MHz).

If the above will be the only change, then I see no reason to upgrade from M4 v1 to M4 v2.

However, in the end, it's all about the implementation, so there may be few slight measurable differences between the two versions of the Motu M4, but let's wait and see for the for the first M4 v2 to get measured.

 

[dolfff]

I am trying to find out if there are any decoupling capacitors at the TRS inputs at the front (1, 2). When I measure with an empty TRS jack, it says: 10nF. Which is a lot! That would mean you need a source with an output impedance no higher than 1kΩ in order to get a good frequency curve. Normal modern line outputs will qualify easily. But hi-Z outputs can be more than 10kΩ. Moto M2/M4 has no switch for line/inst, it has a fits-all solution. So I don't understand my measurement of 10nF. (Of course I did check my meter, it's accurate). With a 10kΩ output that means a rolloff at 1.6 kHz which would be unacceptable. I haven't done any frequency response measurements unfortunately (I lack the equipment). I did however open my M2 and tried to find those 10nF and I think I succeeded: C111A and C111B.

So what am I overlooking? Does anyone made some frequency response measurements at several hi-Z values at the TRS inputs?

Btw, not sure if "decoupling C" is the right term here. What I mean is a small C from input to mass in order to keep unwanted HF out.

 

[trl]

There are no 10nF caps in signal path. However, if some manufacturer is willing to add caps in signal path will add few uF caps, not few nF.

 

[dolfff]

I hope you understand I am not talking about coupling capacitor? Please read last sentence of my post...