Announcement

Collapse
No announcement yet.

Phantom oddity

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Phantom oddity

    Hello there,

    I have been doing some precise measurements of the inputs, outputs and recorded files of the Maxx lately.
    (The results are quite good and quite interesting!)

    While doing so i discovered a strange thing: the phantom power is producing quite a significant amount of noise when turned on in some situations.
    Here is what i got, phantom on, gain at max:

    +4 dB of noise on an open input
    +10 dB on a close circuit input
    +10 dB on a 150Ω input

    Then i tried to see if there would be any difference on an actual microphone, and it seemed that no noise was added when turning the phantom power on a dynamic mike, which is fortunate and means that the use of static microphones would not come with extra noise generated by the phantom power.

    Still, i would like to understand why i'm getting this extra noise (which i never got on any other preamp) and to confirm that it stays "clean" when using a phantom-powered mike.

    Thanks a lot,
    pali

  • #2
    Pali, never measure noise on an unterminated input. this will always give you false readings. You need to put the (recommended) dummy load on, as is for mic and line impedances accordingly. And dynamic mics don't take P48, even if they don't blow up, it's not what you're supposed to do with them. Maybe you want to retry your tests with that in mind

    Comment


    • #3
      Can you elaborate on the recommended dummy load in that case?

      I'm not familiar with the english terms, but the close circuit and 150Ω termination xlrs i used are made for measurements.

      Of course i know that the open input measure was not supposed to be relevant, and of course-of course i know that 48v is not for dynamic mics, did it just out of curiosity.

      For those interested in such figures, the Maxx measures are very close to what i got from a SD 744T in terms of signal-to-noise ratio, dynamics and EIN, with 9 dB more gain and 2 dBU more before clipping the input for the Maxx…

      Comment


      • #4
        huh sorry Pali I'm not good with electronics, but I'm guessing the mixer would be looking at something like 150-600 Ohms for a mic, and around 2k-10k for a Line input? anyone pls confirm?
        ..and sorry about the P48 comment, no offence. Did you also measure SNR utilizing 24Bit ISO with extra headroom?

        Comment


        • #5
          I would have say something like that as well. Anyway i'd love to have a few words from the makers of the maxx about that phantom noise…

          Measures on recorded files were made on 24 bits - one from the mix at max gain (trim+fader fully up (with +10dB input gain option activated)), the other on ISO track with attenuated gain (it was attenuated so my -60 dBU input signal would be recorded at -20dBFS, so the gain attenuation was -14 dB for the Maxx and -4 dB for the 744…)

          The two measures were consistent, with a noise floor at -72 dBFS in the first case and -86 dBFS in the second, both giving an EIN of -126 dB.

          Comment


          • #6
            Pali, you can use up to -24dB for ISO attenuation, selectable. How much did you use for your test? and what's your signal source?

            Comment


            • #7
              Well, in fact the gain was lowered at the input. I didn't thought ISO attenuation would have any effect on the relative level of noise. But i'll try it next time to be sure.

              The input source was -60 dBU from an NTI signal generator…

              Comment


              • #8
                And the next step will be to test the never-clip with stronger input signals and ISO attenuation…

                Comment


                • #9
                  Originally posted by grafpez View Post
                  huh sorry Pali I'm not good with electronics, but I'm guessing the mixer would be looking at something like 150-600 Ohms for a mic, and around 2k-10k for a Line input? anyone pls confirm?
                  ..and sorry about the P48 comment, no offence. Did you also measure SNR utilizing 24Bit ISO with extra headroom?
                  If you're referring to the characteristics of the circuit feeding the mixer, these numbers are definitely wrong.

                  Comment


                  • #10
                    To be honest, cheating, I got these values from the recent SoundDevices 633 manual, p.64/65 [input type], but maybe these guys are wrong? John, could you please enlighten so we can share your knowledge?
                    Pali, which NTI did you use? the analogue one (Minirator) can well only have <96dB THD+N themselves...
                    anyway, that's the nice thing about Maxx: chances are, your input is noisier than your mixer. grand.
                    I personally use an 8060 on the pole, and I had no issue with noise. ever. just people, aircon, planes..the usual. but no noise.
                    So what exactly are you looking for or are you just curious? I've seen several T12/or old P48 mics starting to 'boil' at the end of their day..is it that you're experiencing?
                    Last edited by grafpez; 02-25-2014, 04:43 AM.

                    Comment


                    • #11
                      With modern electronics (the last 50 years, or so) most outputs are of a low impedance and typically feed a bridging input (i.e. higher impedance, usually at least ten times that of the feeding circuit).

                      Comment


                      • #12
                        Originally posted by John Blankenship View Post
                        With modern electronics (the last 50 years, or so) most outputs are of a low impedance and typically feed a bridging input (i.e. higher impedance, usually at least ten times that of the feeding circuit).
                        That is indeed the rule for analog audio: Zout < 10 x Zin
                        If i remember right it is possible to measure the input impedance by sending a signal through a variable impedance circuit, when the signal loss is exactly -6dB that means that Zout = Zin

                        Grafpez: yes, i did that mostly out of curiosity and to understand how it all works a bit better.
                        I did use the Minirator. You're saying it has a noise of its own... Of course, but that does not come into play: i'm sending a -60 dBu from it through the Maxx preamps, which is recorded at -6 dBFS at full gain, then i unplug the the generator and put the 150 ohms dummy load on the input and the noise floor reads -72 dBFS.
                        So the minirator noise is not measured, and the dynamics on the mixdown at full gain is 72 dB.

                        Then i lowered the gain so the recorded signal would read -20dBFS and the noise floor is at -86dBFS, which makes a 86 dB dynamic range on the mixdown.
                        The EIN is calculated by substracting the difference between the recorded signal and the noise floor (66 dB) to the initial -60 dBU signal, which makes -126 dBU

                        You are welcome to comment and correct my method! In any case i'll have to repeat the measure on attenuated ISO tracks to see how the never clip system handles it.

                        Comment


                        • #13
                          thanks John, this is what I remember, roughly a 10:1 ratio, in order to maintain same voltage differential to multiple receiver loads.; however, for line and mic signals these R values are in a different category I believe;
                          And Pali, great workshopping; please remember, the dynamic range is the whole range from noise floor (which you have identified) to the clipping point. Now, in analog systems this could have been anywhere up to +26dBu; with Maxx, being digital AND offering Neverclip, this is different. With this in mind, what results do you get?
                          Last edited by grafpez; 02-26-2014, 03:48 AM.

                          Comment


                          • #14
                            So i've been doing some more measures on ISO tracks...

                            At maximum prefader gain (trim: 30 dB / extra input gain +10dB) on a 24 dB attenuated ISO track, the noise floor is already below -100 dBFS (probably around 108 dB of dynamics in that case), and with 20dB and no extra gain on the input, it extends by … 20 dB, of course.
                            I must say that all the math in the Maxx is perfectly correct, when you go up 20 is reads up 20 in the measure and not a half-dB more.
                            That seems obvious for a Dsp-based machine but it's good to know anyway.

                            My measuring tools are getting a little confused below -100 dBFS so my numbers are not very relevant, probably!
                            I had to add gain in protools to be able to read something an then substract it from my reading. The measure was made on a NTI digilyser, from protools, through the spdif out of my RME.

                            Of course, that's probably too many cables and operations not to be off a few dBs, even if the all process was completely digital.
                            I tried to get to the bottom of it anyway, and with a minimum trim i got a noise floor at -142 dBFS. Even better than what Zaxcom advertises!
                            But that was probably the full extent of the Never clip and/or 24 bits signal dynamics, because if the math had kept steady it would have been -148…

                            Not anything i would be able to hear, anyway!

                            Comment


                            • #15
                              Thanks for doing the tests. I think you are off a few dB but hey at these levels the point is made. Please tell the sound community there is a new sherif in town and his name is Maxx.

                              Glenn

                              Comment

                              Working...
                              X