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

[KSTR]

When I measure with an empty TRS jack, it says: 10nF.

Unlikely that 10nF is real. Did you measure with the unit powered on? If not, you should do so.

 

[dolfff]

When switched on I can't get a measurement, meter is showing random values (just tried). Switched off it shows a rock solid 10.2 nF (I just tried it again and on both inputs) and again: I checked my meter. The only correct method would be measuring input frequency characteristic using different impedances but I don't have the right equipment for that.

The reason I asked: I have a high impedance output antique r2r tape deck for digitizing tapes. I am not even using the regular output but I created some sort of "source-direct" in order to skip a lot of junk electronics and switches. Two problems: signal is too strong for M2 at this point and maximum load is about 200kΩ. And I don't want to use an extra buffer amp. I only want to use a simple voltage divider. So, in order to calculate my divider, I need to know the decoupling C. I agree that 10nF is too much to be real. But I know that there is some decoupling going on because high tones disappear when I use too high values for my divider. And it's not caused by the cable because it's only 20cm long and the divider is directly at Motu's input.

I should rephrase my question: what is maximum allowed output impedance of the source at which a flat frequency response until 20kHz will be obtained? (Motu doesn't answer my decoupling-C question, I hope they will answer this one. If they do, I 'll post it here. If not I hope someone can tell me the correct answer).

 

[trl]

OK, so you're talking about a possible 10nF non-polarised cap paralleled with the output of the TRS front input plugs (Low-Pass Filter)? I remember measuring only regular audio sources like Topping D90 MQA, ASUS Xonar Essence One, Burson Conductor V2/V3, Motu M4 (loopback) etc., all of then having an output impedance of around100 Ohms or so.

I would rely on Wikipedia regarding the line-output impedance of audio sources and ignore those input caps for now, unless you can find an audio source having a very high impedance output that causes issues in measuring flat on Motu interfaces. AFAIK all preamps are using such input caps to prevent getting oscillations coming from audio sources or any kind of unwanted audio frequencies.

 

[dolfff]

@trl Sorry to say but you really seem to not read my posts (entirely) (or is my English so bad?): I DO have such a problematic source as I explained. Solving at the source would be the right way of course but I don't want that for several reasons.
Thanks for acknowledging the anti-hf input caps on pre amps. The question remains: how big is it in the M2? (And how on earth can my intact Fluke see 10nF?).

I decided to go measure it the right way myself (frequency response while using increasing series resistor). I wrote that I don't have the right equipment but I can use other means by being creative.

 

[Blumlein 88]

@trl Sorry to say but you really seem to not read my posts (entirely) (or is my English so bad?): I DO have such a problematic source as I explained. Solving at the source would be the right way of course but I don't want that for several reasons.
Thanks for acknowledging the anti-hf input caps on pre amps. The question remains: how big is it in the M2? (And how on earth can my intact Fluke see 10nF?).

I decided to go measure it the right way myself (frequency response while using increasing series resistor). I wrote that I don't have the right equipment but I can use other means by being creative.

Dude trl and KSTR are good guys and trying to help. Calm down plenty of us will help if we can. Don't be stingy with details of what you are using.

 

[Blumlein 88]

When switched on I can't get a measurement, meter is showing random values (just tried). Switched off it shows a rock solid 10.2 nF (I just tried it again and on both inputs) and again: I checked my meter. The only correct method would be measuring input frequency characteristic using different impedances but I don't have the right equipment for that.

The reason I asked: I have a high impedance output antique r2r tape deck for digitizing tapes. I am not even using the regular output but I created some sort of "source-direct" in order to skip a lot of junk electronics and switches. Two problems: signal is too strong for M2 at this point and maximum load is about 200kΩ. And I don't want to use an extra buffer amp. I only want to use a simple voltage divider. So, in order to calculate my divider, I need to know the decoupling C. I agree that 10nF is too much to be real. But I know that there is some decoupling going on because high tones disappear when I use too high values for my divider. And it's not caused by the cable because it's only 20cm long and the divider is directly at Motu's input.

I should rephrase my question: what is maximum allowed output impedance of the source at which a flat frequency response until 20kHz will be obtained? (Motu doesn't answer my decoupling-C question, I hope they will answer this one. If they do, I 'll post it here. If not I hope someone can tell me the correct answer).

What exactly do you mean by maximum load is 200k?

 

[dolfff]

Sorry for sounding "stingy". Yes you are all good guys, I am convinced of that.

Max load 200kΩ has to do with my source. I could explain but it would become quite off topic. The main point is that I have a problematic source and I don't want to use a buffer amp after that. So I need to know the value of the anti-hf input cap of the Motu in order to make correct calculations and prevent too low rolloff frequency.

 

[Blumlein 88]

Sorry for sounding "stingy". Yes you are all good guys, I am convinced of that.

Max load 200kΩ has to do with my source. I could explain but it would become quite off topic. The main point is that I have a problematic source and I don't want to use a buffer amp after that. So I need to know the value of the anti-hf input cap of the Motu in order to make correct calculations and prevent too low rolloff frequency.

Go ahead and explain. Sounds like the output of your rtr is 20 kohm or more.

 

[KSTR]

When switched on I can't get a measurement, meter is showing random values (just tried).

That's almost what I would expect. There is negligible input capacitcance and your meter may not be able to measure it.

The TRS input is terminated 1Meg Ohm (per leg) to be suitable to guitars etc which have output impedances of 100k Ohms and more, in certain settings, and still shall be able to capture 10kHz. Therefore it is unlikely there is more than 100pF installed.

You could measure the M4 in cable loopback (use the excellent free software REW), one time directly and another time with 100k Ohms inserted in series on the center conductor (assuming you're going unbalanced). Or test with you divider in place (what's your required attenuation?), and fed through the estimated output resisistance (lower divider leg something like 10k...20k and you should be fine). Is the RTR (make and model?) really maxing out the +16dBu (5Vrms) clipping level of the TRS at minumum gain?

 

[trl]

I decided to go measure it the right way myself (frequency response while using increasing series resistor). I wrote that I don't have the right equipment but I can use other means by being creative.

Just use your existing M2 in loopback and use STEPS or ARTA software to see freq. response for any audio source that allows an analogue input.

Use http://sim.okawa-denshi.jp/en/CRlowkeisan.htm to do the math needed, although I think you're overthinking this more than needed. Just use a dedicated pre-amp if you can't connect directly your R2R tape deck.

 

[dolfff]

@Blumlein 88

I feel a bit bothered to pollute this great topic with my source issues. Shouldn't it be about Motu? Well, here we go anyway. A mod can put it in a separate topic if needed.

Regular line-out of my r2r is indeed 20kΩ but as said: I don't use that. I created some sort of "source-direct" in order to skip a lot of junk electronics and switches. Two problems at that point in the circuit: signal is too strong for M2 and maximum load is about 200kΩ. With that I mean: the output of that amp (playback head amp) must be closed with 200kΩ or more. Otherwise, the frequency correction network in that amp's feedback could be influenced. So normally, one would need a buffer amp after that. But I am not normal so I don't want that (for now). I just want to try a voltage divider in order to bring the signal down to acceptable level. I will put that divider IN the TRS connector going into the Motu (I should say unbalanced TS to be precise) so cable capacitance is no issue (which is extremely short anyways). This divider (R1 --- R2) should total at 200kΩ. I have to bring the signal down at least 40% so a divider of let's say 100kΩ --- 100kΩ would do it. However: this would cause an effective source output impedance of 50kΩ which is quite high. Unbalanced input impedance of Motu is 1MΩ so at first I thought: easy peasy. But I didn't account for that anti-hf input cap! I started of with an even much higher value divider of 1MΩ --- 220kΩ. That caused a very audible rolloff which started me thinking. I didn't try 100kΩ --- 100kΩ yet. It think I won't hear a rolloff with it but my ears are old so I want to make a calculation of the crossover frequency. For that I need to know the value of the anti-hf input cap of the Motu's preamp.

@KSTR : I have probably panicked a bit after measuring 10nF so I am glad to hear that it is probably a lot less. (I am still puzzled though how my trustworthy meter shows a stable 10.2nF). So my divider of 100kΩ --- 100kΩ should work flawlessly. Nevertheless, now that I started this investigation, I want to be absolutely sure.
Thanks for your measuring suggestions. I am going to do it a bit different but with same results. I 'll let you know the result.
Yes, the output signal of my source-direct is really clipping out the M2! I was surprised too. The regular line-out is lower though. +16dBu corresponds to 13.8 Vpp but for unbalanced it's half: 6.9 Vpp. My source-direct creates 10.5 Vpp.

@trl Thanks for your suggestions. I 'll post results later on.

 

[Blumlein 88]

@Blumlein 88

I feel a bit bothered to pollute this great topic with my source issues. Shouldn't it be about Motu? Well, here we go anyway. A mod can put it in a separate topic if needed.

Regular line-out of my r2r is indeed 20kΩ but as said: I don't use that. I created some sort of "source-direct" in order to skip a lot of junk electronics and switches. Two problems at that point in the circuit: signal is too strong for M2 and maximum load is about 200kΩ. With that I mean: the output of that amp (playback head amp) must be closed with 200kΩ or more. Otherwise, the frequency correction network in that amp's feedback could be influenced. So normally, one would need a buffer amp after that. But I am not normal so I don't want that (for now). I just want to try a voltage divider in order to bring the signal down to acceptable level. I will put that divider IN the TRS connector going into the Motu (I should say unbalanced TS to be precise) so cable capacitance is no issue (which is extremely short anyways). This divider (R1 --- R2) should total at 200kΩ. I have to bring the signal down at least 40% so a divider of let's say 100kΩ --- 100kΩ would do it. However: this would cause an effective source output impedance of 50kΩ which is quite high. Unbalanced input impedance of Motu is 1MΩ so at first I thought: easy peasy. But I didn't account for that anti-hf input cap! I started of with an even much higher value divider of 1MΩ --- 220kΩ. That caused a very audible rolloff which started me thinking. I didn't try 100kΩ --- 100kΩ yet. It think I won't hear a rolloff with it but my ears are old so I want to make a calculation of the crossover frequency. For that I need to know the value of the anti-hf input cap of the Motu's preamp.

@KSTR : I have probably panicked a bit after measuring 10nF so I am glad to hear that it is probably a lot less. (I am still puzzled though how my trustworthy meter shows a stable 10.2nF). So my divider of 100kΩ --- 100kΩ should work flawlessly. Nevertheless, now that I started this investigation, I want to be absolutely sure.
Thanks for your measuring suggestions. I am going to do it a bit different but with same results. I 'll let you know the result.
Yes, the output signal of my source-direct is really clipping out the M2! I was surprised too. The regular line-out is lower though. +16dBu corresponds to 13.8 Vpp but for unbalanced it's half: 6.9 Vpp. My source-direct creates 10.5 Vpp.

@trl Thanks for your suggestions. I 'll post results later on.

The guitar input on the M2/4 is 1 meg ohm single ended. Feed your recorder to a 200k ohm/20 kohm divider. This will drop the level and give an output impedance as seen by the M2 of 18.2 kohm. Put it near the end of the cable closest to the M2 and it should work just fine. You may have to use a bit of gain on the M2 then, but it should work without altering response.

Also, without knowing the machine and such, bypassing the things you did may not be the best idea.

 

[dolfff]

Yup I agree. I too figured that attenuating the signal more than needed is no problem because M2 has plenty of gain. So I will do that if a resulting impedance of 50kΩ appears to be too much. First I will measure the M2's anti-hf input C. If a resulting 50kΩ impedance is no problem, I prefer the 100kΩ --- 100kΩ solution.

 

[dolfff]

Okay, here we go:

For measurement I used a very low impedance source. I measured everything unbalanced. First I measured without series resistor. Response is flat between 20Hz - 20kHz as expected.

Next I added 150kΩ series resistor close to M2 input. -3dB point is at 4550Hz. Which means Motu M2 has an anti-HF capacitor of close to 230pF at the input. I am really glad to see confirmed that it's not 10nF like my meter was telling (still don't understand why though).

So.... for -3dB point at 20kHz, source should have an output impedance of 34kΩ or less.

For my proposed divider of 100kΩ --- 100kΩ the -3dB point is at about 14kHz. I don't like that number (although you probably need good ears to distinguish this from flat response). So I 'll go for 180kΩ --- 33kΩ. This weakens the signal too much which is no problem because Motu has enough gain. @Blumlein 88predicted this solution perfectly.

@KSTR Are there really guitars of 100kΩ and more? My results show that these WILL have a problem! At 100kΩ, the Motu will have it's -3dB point at 5kHz which is quite unacceptable I presume.

Edit: in a previous version of this post the numbers were a bit different. I made a little measuring error (I had a Fluke at the input as well, adding a bit of extra capacitance) and corrected that.

 

[dolfff]

Thanks everyone for the help and suggestions.

 

[KSTR]

@KSTR Are there really guitars of 100kΩ and more? My results show that these WILL have a problem! At 100kΩ, the Motu will have it's -3dB point at 5kHz which is quite unacceptable I presume.

For example, the classic Les Paul has 500kOhms volume pots, which gives us no less than 125kOhms output impedance when the wiper is at the electrical center position. Obviously, cable capacitance alone already plays against us here.

And true, from your results it seems the High-Z input of the Motu is not really nice for High-Z sources with its additional 300pF of loading.

 

[dolfff]

Wow, the Gibson Les Paul guitar player needs to carry his own tube pre-amp on his back. No really, I don't see how this could ever work because typically you need a long cable. Or could it be that rolloff attributed to it's famous warm sound??

Back to my r2r - Motu adaptation. My r2r is a Philips N4506 by the way. I created a source-direct directly from the head amp's output. As stated before: a voltage divider of 180kΩ --- 33kkΩ should satisfy all the needs with respect to load, impedance, attenuation and frequency response. I soldered it into the TS jack connector (at the end of an only 20cm long cable) going into the Motu. Result: perfect! Same quality as my earlier samples without the divider but this time without clipping. Mission accomplished.

Thanks everyone.

 

[dolfff]

In a previous post I said Motu has 5kHz -3dB point with a 100kΩ source. Sorry, that is wrong, the correct number is: 6.9kHz. Sorry for that.

 

[mightycicadalord]

500k pots are standard on passive humbucker guitars. 250k for single coils. Can't say I've ever had a problem running both directly into the m4 pre's.

 

[mightycicadalord]

Anyone ever hear an update on the out of phase rca outputs?