Very preliminary tempest measurements

Discussion in 'Archived Threads 2001-2004' started by Mark Krawiec, Apr 28, 2002.

  1. Mark Krawiec

    Mark Krawiec Agent

    Joined:
    Jan 9, 2002
    Messages:
    49
    Likes Received:
    0
    Check them out and tell me what you think. I'm still trying to figure the cliowin system out. It's likely going to take a couple of weeks to months to really get good with this, but any comments about my measurements would be appreciated.
    Sealed tempest measurements
     
  2. Mark Krawiec

    Mark Krawiec Agent

    Joined:
    Jan 9, 2002
    Messages:
    49
    Likes Received:
    0
    The more I think about it, the more i suspect that my MLS frequency response measurements are faulty.

    Oh well, a bit more reading and some more testing in a couple of weeks.

    The impedance, rta and spl are ok, i think.

    mark k
     
  3. Jon Hancock

    Jon Hancock Stunt Coordinator

    Joined:
    Mar 18, 1999
    Messages:
    67
    Likes Received:
    0
    Hi Mark,

    Your CLIO WIN can do most of what you want, but it sort of assumes you know everything about setting it up- otherwise, you get weird results.

    For example, in setting up the FFT mode for RTA, be sure to use at least a 16K point sample size, and always set sampling frequency to the max. You seem to have done that, looking at the headings on your chart. CLIO will do 1/12 octave FFT, but without the sample size and sampling rate being set correctly (which, really, the software ought to be smart enough to do for you), you wind up with "holes" in the FFT plot because there's no data to go in those frequency bins.

    Same thing applies to MLS for amplitude or impedance. Sampling rate and sample size must be up there- the only advantage to the smaller sizes is computation speed, and that's offset by how dirty the data is.

    Also, my personal preference is to use the swept sine for impedance; gated and at slow speed. Then I get pretty accurate results, if I remember right, using 1/48th octave. Your impedance sweep looks pretty rough- I'd suspect the measurement. The MLS for impedance is quite quick, but it has poor resolution in the lower frequency range, and is not as good at identifying driver resonance issues. It's annoying, because it does take a while.

    If you're curious, I'll try to fire up the rig on the weekend and check all my settings in case that will give you some useful baselines to work from.

    Otherwise, you look to be on the right track, and that's pretty much what the SHIVA's look like- what distance are you actually measuring as "nearfield"? If you get up within an inch or so, you can mic different parts of the cone, and store and overlay the sweeps in CLIOWIN, which gives you a very clear idea (besides the impedance plot) where you're leaving pistonic behavior.

    I just got in a pair of 11" (called 10's) Titanic II's to check out for the Arvo Part project; they look very promising for the dipole woofers (their inductance results in a 6 dB/octave roll off above 75 Hz; "built in dipole EQ"). Problem, as usual, is long work hours lately. I doubt I'll get to look at them before I go to ThomasW's on the 7th.

    Regards,

    Jon
     
  4. Mark Krawiec

    Mark Krawiec Agent

    Joined:
    Jan 9, 2002
    Messages:
    49
    Likes Received:
    0
    thanks jon,

    it's actually cliowin lite-so i'm limited to 16k sample size and sine wave for impedance. octave filtering is limited to 1/3 for rta measurements in the lite version. the waterfall plot feature is disabled. i may upgrade one of these days, but 16k is at least acceptable and i can always download praxis for up to 1/48 octave filtering rta in the demo mode. that means that basically i'd have to spend $395 for the waterfall feature and higher sample sizes. nice, but maybe not that nice. the lite version was ~400 less anyway, so, it's not like you're saving any money by purchasing the full version.

    the impedance sweep was 1/48 octave. not suprisingly, looked smoother doing it with 1/24.

    what was puzzling was the hump 40-80hz with the nearfield mls (1/4" mic distance). I went back and looked at the tempest white paper and lo and behold, there is also the same rise in adire's published curves. this was termed an artifact of nearfield measurement? i sent dan wiggins an email this morning to see what they meant about this.

    turns out my measurement might not be that bad. though, in retrospect i should have also done a sinewave nearfield measurement and compared the two.

    i'm going to try and test some more speakers next weekend as well as getting the MMII preamp assembled.

    thanks jon
     

Share This Page