Milos DIY Electrostats (rev 1)

Discussion in 'Headphone Measurements' started by ultrabike, Jan 18, 2016.

  1. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    These to me were a highlight of our recent Anaheim meet. I've read about chinsettawong's DIY electrostats. Well, now we have a SoCal native giving stats a shot. @n3rdling!

    Awesome sauce pictures brought to you by The Merv:

    [​IMG]

    These were driven by the Aristaeus, which is one of my favorite electrostatic amps.

    IMO there was a little bit of a channel imbalance, where the right channel (w/o the hole in the middle) was a little brighter than the left channel. Though it could be that it was just the right channel was a little louder over all. I did not try all kinds of music nor did I spend countless hours with these. This is not possible during meets. But I felt these sounded great.

    I know plenty of folks are not happy with Sennheiser, HFM and many others seemingly arbitrary and meaningless pricing practices. So I talked to Milos, and told him I hoped his efforts would result eventually in an electrostat product that would be at the reach of us penniless commoners. We shall see. In any case, kudos to Milos!

    Here are some meet measurements (disregard absolute SPL level):

    Frequency Response

    Milos_cans_fr.png

    CSD Right

    Milos_cans_csd_right.png

    CSD Left

    Milos_cans_csd_left.png
     
  2. Merrick

    Merrick A lidless ear

    Pyrate
    Joined:
    Jan 6, 2016
    Likes Received:
    12,569
    Trophy Points:
    113
    Location:
    Portland, OR
    I would dearly love to hear these.
     
  3. n3rdling

    n3rdling Friend

    Pyrate
    Joined:
    Sep 26, 2015
    Likes Received:
    929
    Trophy Points:
    93
    Location:
    Los Angeles
    Neat! People seemed to enjoy them at the meet so that was nice. I pretty much use these exclusively nowadays.

    Me and ultra did some listening to specific tone sweep windows (like 3k-6k and then 4.5k-6k etc) to see if those dips were artifacts since I don't really hear them that way. We both agreed that the dips were somewhat accurate in placement, but I personally thought there was no way the dips were that severe (not sure if ultra agrees with me there). The larger dips sounded more like 3-5 dB down to me. I realized today that since we didn't set the dB calibration up, the measurements were shown in REW as 130 dB even though it was actually probably in the 80s. I wonder if that throws off the scale and stretches things vertically?

    Regarding the slight imbalance I'm not sure if the hole I put in the center of the left driver resulted in a weaker field or if that side wasn't as fully charged. Both scenarios should be easy enough for me to test though.

    I'm working on a DIY stat IEM that's almost complete and a couple other projects, so I should have some new toys at the next meet. Maybe when these projects are complete I'll send them to you or Tyll to play around with. :)
     
  4. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    That looks awesome! Nice job! Would love to hear these myself.

    I wouldn't worry about the dips. Measurements tend to exacerbate dips and the like vs subjective tests. The types of dips in those measurements probably are not very audible during listening.
     
  5. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    I know using tone sweeps can be dangerous when checking things out. Specially considering the equal-loudness Fletcher-Munson curves.

    Say we were listening around 80 dB. Between 4.5k to 6k one needs to put out more power going up in frequency to perceive same loudness (about 5 dB). The most silent tone should have been 6k (by like 5 dB). However, we perceived a minimum at around 5.5k. This may indicate that 5.5 might have been a little more than 5 dB down.

    Here is the curve for convinience:

    [​IMG]

    Tones may be useful when looking for deep notches. But otherwise, it can be hard to make definitive assesments. It may be useful to use flat speakers and do relative comparisons on narrow frequency ranges.

    With music, things may be a little off sometimes, but it can be hard to pin point what. And so we make up all these terms to try to convey what we think, which in many cases may be confusing in it's own right.

    I agree with Hans also in that narrow notches are usually not that obvious. Omisions tend to be more benign than peaks (heard that before and I agree). It could be frequency masking and other stuff. But that's not something I'm a foremost expert on.

    As far as the y-axis, things should be ok as far as scale. It is the reference which is wrong, so plots would have to be shifted down quite a bit. But again, relative (scale) should be correct.

    Would love to check out IEMs. I don't have a stat amp to test them though.

    It is very nice and encouraging to see Milos efforts. Indeed these cans sound great IMO.
     
    Last edited: Jan 18, 2016
  6. donjoe0

    donjoe0 most likely deaf

    Banned
    Joined:
    Jun 19, 2018
    Likes Received:
    5
    Trophy Points:
    18
    Location:
    Romania
    That looks more like the Robinson-Dadson set of curves. The original Fletcher-Munson set had the 100 phon curve flatten out across the mids and bass from 1k all the way down. And both are superseded by newer measurements that produced the ISO 226:2003 equal-loudness curves:
    [​IMG]
     
  7. Ardacer

    Ardacer Friend

    Pyrate Contributor
    Joined:
    Oct 15, 2017
    Likes Received:
    986
    Trophy Points:
    93
    Location:
    Croatia
    Better dips than peaks, definitively. Also they appear to be very narrow, which is great. What I'd say from this is that average trebble energy is a tad bit lower then the very nice and flat rest of the spectrum, without much instant red flags. For a diy phone, and electrostat to boot, amazing results.
     
  8. n3rdling

    n3rdling Friend

    Pyrate
    Joined:
    Sep 26, 2015
    Likes Received:
    929
    Trophy Points:
    93
    Location:
    Los Angeles
    FWIW the measurements in this thread were from an early iteration that had some channel imbalance which has since been fixed. @ultrabike took measurements of the updated version at a later meet:

    [​IMG]

    Impressions of said prototype iteration along with more measurements can be found in my thread here:
    https://www.superbestaudiofriends.o...g-he90-orpheus-clone-loaner-impressions.3569/

    Getting the cable made has been a slow and frustrating process, but the earpads are essentially finalized and ready for a production order.
     

Share This Page