On a 6kHz Helmholtz Resonator for the Sennheiser HD800

Discussion in 'Modifications and Tweaks' started by yaluen, Mar 28, 2016.

  1. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,955
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Gotta run now, 3 measurements ...
    SD comparo.png
    probably not too accurate...
    no mods, old pads

    EDIT: On my phone now
    What is the resonant frequency of an open tube with those dimensions? Looks like the mesh might not be as permeable as we think (purrin), even at these high frequencies.
    I wouldn't worry about the drop after 10k with SD upside down, simple positioning error.

    I'm really not sure how accurate these measurements are, the SD measurements I got before were a lot better. Will be able to verify in a couple hours.


    Also: the calculations you posted before look very different from the resonator we are using right now, I have my doubts if they are as effective but it would sure be interesting to try. Does someone have a super accurate 3D printer, an HD800 and measurement equipment?
     
    Last edited: Mar 29, 2016
  2. yaluen

    yaluen Acquaintance

    Contributor
    Joined:
    Sep 29, 2015
    Likes Received:
    82
    Trophy Points:
    18
    Location:
    Metro Van
    Hmmmm, looks interesting to say the least. SD upside down is it? What to make of this I wonder...
     
  3. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,955
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Yeah, me too. Keep in mind my coupler shows less of a peak and higher in frequency. It is closer to purrin's v1 coupler (but denser). Essentially this is a pretty dense open cell foam (afaik).
    It seem that the SD resonant frequency is probably too low, but keep in mind that the measurements with the mic in my ear looked pretty good (didn't post these yet and don't have them with me rn). I would love to measure a HD800S.

    EDIT:
    It would really seem like it was a bad idea to leave the hole open like with the original HD800. I think the SD resonator can be improved upon for sure.
    Away now for a while.
     
    Last edited: Mar 29, 2016
  4. sorrodje

    sorrodje Carla Bruni's other lover - Friend

    Staff Member Friend MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    2,786
    Trophy Points:
    113
    Location:
    Dijon / Burgundy / France / EU
    @Serious : You results with the SD resonator are not fully in line with best results measured elsewhere. Did you take the time to fit it properly ? the creato ring should be flush to the top and the disc of paper should be put at the bottom of the hole. if the SD mod is squeashed in the hole, it does not work properly.

    See Stratocaster measurements : suboptimal positionning : http://www.superbestaudiofriends.or...superdupont-resonator.1147/page-24#post-39945 ... Good positionning :http://www.superbestaudiofriends.or...superdupont-resonator.1147/page-24#post-39959

    My experience told me that whatever we put in the hole it brings some improvements on the treble peak. but the best I achieved was always with something that behaves (at least partly) as a resonator with an acoustic chamber that "trap" the appropriate frequencies. when I worked on the hole in the creatology ( tried 4mm , 6mm , 7mm and 8mm) ,I was able to see difference : affected frequencies was not the same and I clearly saw the 6khz appears or disappear depending on the specs of what I was using.

    IMO , the first mod with the tip was very close from what we could call a fully filled hole. I tried to compress a full shure tip (without to cut it) in the hole and the result was not good ( still the -2b in the treble but nothin as good and specific than the current resonator)

    My 2 cents.
     
  5. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,955
    Trophy Points:
    113
    Location:
    near Munich, Germany
    I agree, the SD measurements look much worse than even my prior measurements. But even then I have never done measuremente without the Anax 2.0 kind of mods. The old pads also change the treble measurements. When I am back home I will reverify. I did, however, make sure that the paper was on the bottom and that the foam was flush to the top. Besides, the resonant frequency should decrease as the volume gets larger.

    EDIT: Back home
    Dug through my old measurements. They look quite similar actually. It was still the most effective resonator on my measurement rig, so I think it is safe to say that the differences in the measurements are related to the coupler (it measures less peaky on the more leaky baffle of my rig to begin with).

    Measurements:
    With all these measurements I used the same mods which consisted of:
    • Thin 2mm felt ring (1mm wider than the metal ring that sits around the driver) on the metal ring
    • Foam trapezoid
    • Weird Anax 2.0 style liner, but maybe ever so slightly thinner and black
    • Pads (not sure if the same on all 3 measurements): New pads with a thinner than stock plastic ring underneath. (Over half a year back I used a dremel to cut a large portion off the plastic ring. The size of the plastic ring was reduced to the bare minimum to still hold the pads in place)
    • No metal grill on the outside. I think it sounds more open and smoother. No measurable difference

    #1:
    SD vs no resonator 2.png
    This is an older measurement and I made sure to match them at 4kHz which emphasizes the difference at 6kHz. It still looks as though the resonant frequency of the resonator might be slightly too low.

    #2:
    SD vs no resonator 3.png
    I'm not sure if this is as accurate but it shows the difference at 4kHz more clearly. I'm thinking that there was some glue residue left on the bottom of the hole which blocked the hole. This is why there is such a big dip at 9kHz with this one.

    #3 (In Ear Measurement):
    Inear SD vs no resonator.png
    This is a measurement with the microphone in my ear canal. The microphone was placed in a silicone IEM eartip and then placed so that it was probably flush with the opening of my ear canal. I really didn't want to post this for two reasons:
    1. I have no reference as to how it should measure. Measuring speakers like this resulted in a FR similar to what you would assume that this should look like (big 5-10db emphasis between 2-5kHz) but EQing headphones to measure like that made it sound like shit. The HD600 measures not too different from the modded HD800SD. It looks like a flat FR at the ear canal opening probably sounds pretty good.
    2. The HUGE bass roll off you see in these measurements. This is mostly caused because the HD800 driver in my left earcup is not secured by the screws anymore and with this measurement I didn't make sure to push it completely back in. But on the other hand I also do get a bass roll off starting at >100Hz even on the right channel which is caused mostly because the mods push the region above 100Hz up. The cork ring even more so than the felt ring, which is why I prefer to use a thinner ring with my liner material. The SBAF absorber probably doesn't need the cork ring. >Measurements on my head show even less bass than on my leaky coupler.
    With this one we see a lot more energy between 2 and 3kHz. I also consistently get more energy at 10kHz which seems to always be the case when the hole is blocked off. Even now the resonant frequency of the resonator (which seems to be slightly below 6kHz) seems just slightly low. I might request a new resonator pair and try to widen the hole slightly on mine to see if that improves it. But then again if the area from 6-8kHz was just 1db lower it would look pretty perfect.

    I ordered 10 original Senn screws that hold the driver in place. I think @sorrodje mentioned that his friend is also missing 2 of these screws, so I might be able to send him some if he wants that.
    I think I forgot something but this is already way too long anyways. On to getting more consistent measurements of the same thing that I posted above.

    Edit: Found out how the spoiler function works. Much better now
     
    Last edited: Mar 29, 2016
  6. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,955
    Trophy Points:
    113
    Location:
    near Munich, Germany
    You are right, my cut-to-shape Comply T400 and the SD upside down measure virtually identical:
    SD upside down vs Comply T400.png
    It is interesting that the response is really smooth and the CSDs do indeed look very good.

    I will update this post with more accurate measurements of the SD resonator in a stock HD800 with old pads later.

    Update:
    I really think that the differences are caused by differences in the couplers and in the pads. You wouldn't think how much of a difference the pads make. I doubt the huge differences in Tyll's measurements between S/N 15001 and 457 were caused because the headphones were different.
    It might still be a good idea to make a resonator with a higher resonant frequency.

    Here are more accurate measurements with new pads. The difference in FR is still roughly the same - the SD resonator works, it's just that my measurements show the peak higher in frequency, so it doesn't look as effective.
    new pads - SD vs SD upside down.png

    I will try to get some in ear measurements of the same thing as I find them to match what I hear the closest. I believe the peak is higher in frequency with the in ear measurements, at around 6.5kHz (as seen in the above post), like with Tyll's measurements.
     
    Last edited: Mar 29, 2016
  7. yaluen

    yaluen Acquaintance

    Contributor
    Joined:
    Sep 29, 2015
    Likes Received:
    82
    Trophy Points:
    18
    Location:
    Metro Van
    Well of course this would be the case. If you take away the felt ring, you're increasing the cavity volume, about 1.8x assuming for now felt behaves as rigid walls. It's hard to figure out at a glance how to vary the height of the foam ring because of end correction, but holding heights constant, we'll need to significantly increase the cross-sectional area of the neck to get the same resonant frequency. Since the dimensions we're working at is so small, even a little change will significantly effect the resonant frequency. Again, the resonant frequency in those calculations can be off by as much as 1kHz because the end correction coefficient is inaccurate.

    Thanks for taking the time to take and post all the measurements, still find it curious how different a stock HD800 measures compared to one with hole covered from the inside end.

    For sorrodje: Do you have CSDs from when you were experimenting with the hole diameter in the creatology? It maaaay be useful in coming up with a better end correction coefficient.
     
  8. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,955
    Trophy Points:
    113
    Location:
    near Munich, Germany
    I doubt that felt behaves as a rigid wall. Would be very interesting to see @sorrodje 's measurements of felt vs no felt.

    Yes indeed. HD800 with hole covered actually looks pretty good

    Also: Figured this would be very interesting:
    diff SD vs SD upside down.jpg
    This is the difference between SD upside down and SD. It looks like the resonant frequency is just slightly lower than 5kHz but I would like to take some time to see if I can get it to increase simply by putting the resonator in there in a different way. Maybe more squished is a good thing.
     
  9. sorrodje

    sorrodje Carla Bruni's other lover - Friend

    Staff Member Friend MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    2,786
    Trophy Points:
    113
    Location:
    Dijon / Burgundy / France / EU
    Unfortunately , I didn't keep measurements for unsuccessful attempts :/
     
  10. yaluen

    yaluen Acquaintance

    Contributor
    Joined:
    Sep 29, 2015
    Likes Received:
    82
    Trophy Points:
    18
    Location:
    Metro Van
    Ah, well that's too bad. Still, no big loss since trying to find accurate end correction is probably the least efficient approach to getting a suitable Helmholtz resonator. Experimentation is still easiest.

    I just wanted to clarify the principle at work with the Helmholtz resonator. The HR isn't like a pipe or enclosure that have standing wave resonances and the HR cavity isn't "trapping" acoustic energy. Recall how the dimensions of the HR is required to be much smaller than the wavelength of the frequency of interest. Rather, sound pressure is converted into vibration of the air mass in the neck of the HR at its resonant frequency and energy is absorbed that way.
     
  11. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,955
    Trophy Points:
    113
    Location:
    near Munich, Germany
    I was able to get measurements of an HD800S today. These were taken with my laptop in a dealer showroom.
    HD800S vs HD800SD.png
    HD800S in pink, HD800SD in yellow
    I got consistently more bass extension out of the HD800S. These measurements were both taken with my new pads and without the dust cover. The microphone input is really crappy which explains the huge drop off in level after 7kHz. Disregard the SPL in db, it was not calibrated at all. The resonator in the HD800S is definitely tuned to a higher frequency, which was already evident from @sorrodje's measurements.

    I listened to it out of my Geek Out 450 (without the Leckerton). The HD800S sounded much smoother but to me the higher bass distortion was obvious. Almost sounded like HD600 bass quality. Much muddier than the HD800.

    I also attached a crappy phone pic where I tried to see how deep the resonator was. You can see a slight silver line in the lower corner. It looks like this is less thin than the SD resonator (to make the resonant frequency higher). I guess the material used to block off the lower end matters a lot as @sorrodje mentioned getting worse results with a foam disk instead of the paper.

    EDIT: I should probably also mention that this particular HD800S already had the paint chipped. Truly horrible paint quality. It looks even worse on the HD800S because the contrast between chipped and black paint is so much bigger.

    I also attached a pic of with vs without the dust cover. It looks like the dust cover was already accounted for and it might even sound better with the dust cover. I only listened to both with the dust cover.
     

    Attached Files:

    Last edited: Mar 30, 2016
  12. logscool

    logscool Friend

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    64
    Trophy Points:
    28
    Dude, get in the habit of putting a legend on your graph or saying what color is what. In this case it's obvious but maybe not to everyone.

    I think the importance of using cardboard in the bottom vs. foam has more to do with the foam taking up more room. Because while I noticed a big difference between the materials when placed at the bottom of the hole, if I place the foam on the backside (as I prefer) I get essentially no difference between foam and cardboard.

    I wonder how paper would do as a material either for the top or bottom sections of the resonator?
     
  13. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,955
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Updated.

    Don't we want to increase the resonant frequncy? I thought that this meant either decreasing the volume of the cavity or increasing the area of the hole. The HD800S resonator looked less deep to me. I guess paper would be a pretty good material. I'm thinking that something reflective at the frequencies we're working at would work best.

    Obviously we don't want to make the cavity too small or it will likely be less effective. I'm wondering if the geometry matters if we keep the volume the same.
     
  14. logscool

    logscool Friend

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    64
    Trophy Points:
    28
    Yes you do want less volume to increase the frequency, but I can tell you I have gotten better results by doing things that would seem to increase the volume.
     
  15. yaluen

    yaluen Acquaintance

    Contributor
    Joined:
    Sep 29, 2015
    Likes Received:
    82
    Trophy Points:
    18
    Location:
    Metro Van
    Thanks for the reportage Serious. I would caution against using the HD800S resonator as a frame of reference until we know more about it, characteristics of the cavity backing, neck ring material, etc etc. Also, geometry of the neck does affect end correction.
     
  16. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,955
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Yes, I only used it to confirm my suspicion that the resonator in the HD800S is tuned to a higher frequency.

    The resonator @sorrodje made is already very good and I think he and @logscool know better than me how hard it is to make a resonator that is effective and works at such a high frequency. I will definitely do my own experiments and share future measurements and document the effects changing one factor has to help developing an even more effective helmholtz resonator for the HD800.

    Who knows, maybe with certain mods a resonator isn't even needed. I think that none of the measurement couplers seem to properly reflect the frequency response of the HD800 for me. Only Tyll's dummy head and my own in ear measurements get close. This isn't the case with other headphones such as the HD6X0.
     
  17. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,955
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Sorry for reviving this thread, but I feel that I've not given @sorrodje enough credit.

    I refitted the resonator, so that the inner cavity is slightly smaller and this got rid of the 4kHz dip that I got before. Before the inner part was slightly higher than the outer part, so that while the outer part was flush, the inner part (where the hole is, at the top) was actually slightly higher than the driver cover. Now it is flush all the way around (like you'd get with the screen still intact).

    This is what it looks like now (all measurements with the other mods in place)
    SD before vs after 2.png
    the "SD after 2" was slightly higher and thus had a slightly bigger inner volume (as is evident from the lower resonant frequency). All in all there are now no drawbacks over the SD upside down anymore.

    I also wanted to add that I have done my own resonator experiments and what I've found is that if you manage to create a resonator with a higher resonant frequency, it will also be less effective. Not even once have I gotten a result only nearly as good as the SD resonator.
     
    Last edited: May 2, 2016

Share This Page