• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

desk tower 3way WIP.. (BlieSMa t25b + ETON 3-212 + Seas l26roy) + Hypex + PE boxes

OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
Awaiting the end result with excitement!
Cheers! It will be a week+.

Do you have idea on what frequency can be attenuate by felt only and what frequency need to use MPP? I am thinking is whether 2000Hz a good number there or we can go lower and leave the higher frequency to felt.

Splitting the baffle can attenuate multiple frequencies, but does it make the effect lower because you have less area per frequency range?

I expect the felt to run out of steam around 1-2khz. I went up to 3khz because I wanted to overlap the tweeter's range. Where each section is tuned should shift down with each layer of felt that I add.

I hope to measure over each area of the baffle and see if there is anything to see. Again, my baffle is tiny so this could all be a waste.

https://www.researchgate.net/figure...lt-and-wool-felt-with-membrane_fig4_316773713

That is one view of wool felt. The MPP paper used a plant-based jute felt.

more:

https://www.scirp.org/pdf/OJA_2017061616554618.pdf
https://www.researchgate.net/public...ials_with_Improved_Sound-Absorbing_Properties
https://www.researchgate.net/public...xtile-Based_Waste_for_the_Automobile_Industry

The felt I am using is listed as between f1 and f5. This is the more dense stuff so it should be good for what I need it to do.
http://www.usfelt.com/sae_felt_specs.html

The pics are ripped from the papers. They mostly show different densities of felt and different ways of making it. WHXXX, WM, and CML are all felt.
 

Attachments

  • felt3.jpg
    felt3.jpg
    42.9 KB · Views: 128
  • felt2.jpg
    felt2.jpg
    141.6 KB · Views: 154
  • felt.jpg
    felt.jpg
    92.8 KB · Views: 156
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
I did not want to solder the tweeter out of fear of cooking it. I also have the need to change out the baffles. Crimping it is! The BlieSMa has gold connectors and I wanted to show it some non-corrosive love. I had to buy all of the gold "Female Disconnect" sizes since none of my drivers are the same. (Thanks, manufacturers!)

I had to use a tiny screwdriver to open up the fem. disconnect. You will go mad if you just try the fem. disconnects out of the package on a driver. Dry fit FTW. The wires are soldered to a SpeakON connector.

https://www.parts-express.com/Gold-...2yHENYAIuG5UM0kc5Uv2_vq0bKy9JYqsaAvBHEALw_wcB
 

Attachments

  • Female_Disconnect_done.jpg
    Female_Disconnect_done.jpg
    376.2 KB · Views: 169
  • Female_Disconnect_dryfit.jpg
    Female_Disconnect_dryfit.jpg
    588.1 KB · Views: 163
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
Woot! I got my parts today for the final round of my little experiments. I did a dry fit and things look great.

The scientific paper for the perforated metal that I based lots of this off of is kinda old. There are newer papers that cover what is happening inside the chambers I am making. Looks like a larger hole size could let me control the phase.

I hope to be listening and measuring tomorrow.
 

Attachments

  • round3felt_stainlessbaffle.jpg
    round3felt_stainlessbaffle.jpg
    743.7 KB · Views: 156
  • Unit-cells-of-metasurface-used-for-acoustic-surface-engineering-a-A.jpg
    Unit-cells-of-metasurface-used-for-acoustic-surface-engineering-a-A.jpg
    69.7 KB · Views: 179
Last edited:
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
Some pics/measurements of the new baffle....

"dry_fit_2layerfelt_mpp_noMELAMINEvsMDF.png " Shows what would happen if you had a mic 6" above the new baffle on a table while you played 500-5000hz pink noise from a speaker in the center of the room 2m away. My aim was to measure some off-axis room soup off the surface. Each measurement is made from 10 RTA averages. I would place a 1/2" thick mdf board on top for the mdf measurements. There is no melamine in the new baffle for this measurement, just 2 layers of felt on top and the steel plate. This was my sneak peek before I glued it up. So far so good!

The unexpected thing is I both hear and measure more bass. Steel seems to really start reflecting around 600hz.

I also need to take a higher SPL distortion sweep.

The decay looks great to me. None of my felt-alone measurements look this good.

Back to measuring....
 

Attachments

  • melamine_baffle.jpg
    melamine_baffle.jpg
    535.6 KB · Views: 262
  • 1khz_plate.jpg
    1khz_plate.jpg
    406.7 KB · Views: 255
  • plate_felt_layers.jpg
    plate_felt_layers.jpg
    474.4 KB · Views: 231
  • clarity_1m_2layerFelt_1khzPLATE.png
    clarity_1m_2layerFelt_1khzPLATE.png
    43.2 KB · Views: 211
  • diist_1m_2layerFelt_1khzPLATE.png
    diist_1m_2layerFelt_1khzPLATE.png
    106.9 KB · Views: 218
  • decaysm_1meter_2layerFelt_1khzPLATE.png
    decaysm_1meter_2layerFelt_1khzPLATE.png
    89 KB · Views: 184
  • decay_1meter_2layerFelt_1khzPLATE.png
    decay_1meter_2layerFelt_1khzPLATE.png
    137.5 KB · Views: 184
  • rt60_1ft_2layerFelt_1khzPLATE.png
    rt60_1ft_2layerFelt_1khzPLATE.png
    238.1 KB · Views: 173
  • rt60_1meter_2layerFelt_1khzPLATE.png
    rt60_1meter_2layerFelt_1khzPLATE.png
    361.6 KB · Views: 167
  • dry_fit_2layerfelt_mpp_noMELAMINEvsMDF.png
    dry_fit_2layerfelt_mpp_noMELAMINEvsMDF.png
    58 KB · Views: 174
  • spectrogram_1m_2layerFelt_1khzPLATE.jpg
    spectrogram_1m_2layerFelt_1khzPLATE.jpg
    542 KB · Views: 165
  • spectrogram_1ft_2layerFelt_1khzPLATE.jpg
    spectrogram_1ft_2layerFelt_1khzPLATE.jpg
    428.8 KB · Views: 219
Last edited:
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
I have some slapdash data to compare a hardboard baffle with no felt VS the new baffle. I did not think of comparing until after measuring for different things. My plan is to do a better job later by using the same drivers for all the measurements. So keep in mind the drivers are my other set and the eq is meant for the new baffle. The spl's were also a bit different for each sweep. A laser measure was used to keep the mic in the same spot.

1m loud = new baffle (2 layers of felt + perforrated steel + airgap (50% air + 50%melamine))
1m wood = hardboard baffle w/o felt

To my newbie eyes, the new baffle is cleaning up 1khz-2khz and a bunch of other nice things. The XO is at 1.7khz. I find it interesting that the XO null moved and there is more power at the XO. I wonder how many of these things can be cleaned up with a better EQ for each baffle? We'll see.

My ears hear a difference. I hope to test my perception a bit when I can do faster speaker swaps.
 

Attachments

  • cal_rt60_1m_hardboard_diffDrivers.jpg
    cal_rt60_1m_hardboard_diffDrivers.jpg
    436.8 KB · Views: 99
  • cal_rt60_1m_2layerFelt_1khzPLATE.jpg
    cal_rt60_1m_2layerFelt_1khzPLATE.jpg
    416.4 KB · Views: 98
  • cal_spectrogram_1m_2layerFelt_1khzPLATE2.jpg
    cal_spectrogram_1m_2layerFelt_1khzPLATE2.jpg
    481.9 KB · Views: 104
  • cal_spectrogram_1m_hardboard_diffDrivers2.jpg
    cal_spectrogram_1m_hardboard_diffDrivers2.jpg
    478.6 KB · Views: 109
  • cal_spectrogram_1m_hardboard_diffDrivers.jpg
    cal_spectrogram_1m_hardboard_diffDrivers.jpg
    541.4 KB · Views: 111
  • cal_spectrogram_1m_2layerFelt_1khzPLATE.jpg
    cal_spectrogram_1m_2layerFelt_1khzPLATE.jpg
    567.3 KB · Views: 99
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
Some new images from the last week of tinkering.

MPP... the new perf baffle design. In the first round, I guessed at what freqs. to target. Now I have measurements and made chambers to target real resonances. The resonances show up on both baffles, wood, or whatever. I believe the baffle shape causes some frequencies to ring on the corners. I also think some of these issues are tiny speaker problems.

on_axis_1ft... This is an on-axis look at each of the steel plates with a baffle that is 2ft wide (next pic). The all 1khz plate VS the 600ish-3khz plate battle using the same drivers, amps, etc... everything was the same except the plates. I did have to take apart the speaker between measurements. I love how 2.5khz decay gets hammered down by the multi freq plate.

1khz pink... This is the foam sheet around the baffle.

between_ 1ft.. This is a comparison of the 2' wide baffle vs the normal one with the 44mm edge radius. The mic was between the speakers so this is an off-axis look using the same speaker for both measurements. Notice how the decay shifts up with a baffle size change.

baffle sim... I've was going mad guessing at what was happening with the new baffle. I went back to 2D to help see what was going on since my brain stinks at simulating how acoustics work. The sim shows a power increase on-axis and the improved decay- things I see with my real measurements. It also shows that a plain baffle messes with the decay no matter what kind of rounding you use.

https://www.falstad.com/ripple/Ripple.html

Code:
$ 3 1023 127 35 0 1199 0.048828125
202 0 390 553 548 648 0
203 0 421 532 471 532 1 2 8
b 0 470 532 545 582 0
c 0 417 532 471 553 0 10000
e 0 415 563 457 605 0
b 0 397 532 425 571 0
b 0 214 0 467 1018 0
b 0 467 0 719 1018 0
P 1 399 287
P 1 536 287
s 2 463 529 2 2.566663 0 38.274470661478325 147.2095025441474 1 0
s 2 471 529 2 2.566663 0 38.274470661478325 147.2095025441474 1 0
e 0 358 532 431 605 0
P 1 508 486
P 1 427 485
 

Attachments

  • MPP_multi2.png
    MPP_multi2.png
    29.6 KB · Views: 133
  • on_axis_1ft_MultiPlate.jpg
    on_axis_1ft_MultiPlate.jpg
    466.4 KB · Views: 111
  • on_axis_1ft_1khzPlate.jpg
    on_axis_1ft_1khzPlate.jpg
    481.9 KB · Views: 143
  • 1khz_pinkbaffle.jpg
    1khz_pinkbaffle.jpg
    442.8 KB · Views: 149
  • between_1ft_44mmEdge_1khz_1feltLayer2.jpg
    between_1ft_44mmEdge_1khz_1feltLayer2.jpg
    467.7 KB · Views: 113
  • between_1ft_pinksheet_1khz_1feltLayer2.jpg
    between_1ft_pinksheet_1khz_1feltLayer2.jpg
    491.3 KB · Views: 118
  • baffle_sim_perf_VS_square.jpg
    baffle_sim_perf_VS_square.jpg
    343.8 KB · Views: 124
Last edited:
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
After way too long, I now have the seas l26roy in the final spot. Piano sounds so nice with the bass so close to the mid. Before the l26roy was about 11" away and pointed up. The change from a speaker stand to the wide sub box also changed the performance of the mid and tweeter. I had to bring both of them down in SPL's.

I compared different amounts of stuffing inside the sub-box. If I put in too much, it messed with the off-axis bass- I was seeing a dip at 150hz. I ended up with 2" of melamine spaced 2" from the back wall of the cabinet.

I am also doing a second round of perforated plates for the baffle. This time with specific targets and a nomogram that is calibrated for the spacing+melamine based on measurements of the round 1 baffle. The distribution should be better for round 2. I included an image from round 1 with labels.
 

Attachments

  • inside-l26roy2.jpg
    inside-l26roy2.jpg
    256 KB · Views: 155
  • inside-l26roy.jpg
    inside-l26roy.jpg
    332.6 KB · Views: 168
  • l26ROY.jpg
    l26ROY.jpg
    291.2 KB · Views: 166
  • MPP_multi2_bigmix.png
    MPP_multi2_bigmix.png
    45.7 KB · Views: 168
  • MPP_multi_labels.png
    MPP_multi_labels.png
    35.4 KB · Views: 142
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
My newbie journey continues...

I'm been doing lots of rounds of measurements and eq fixes. I am using FDW in REW to smooth things out. My main interest is finding the spots that need a filter and making sure it does not mess with the off-axis.

For all of the eq rounds, I moved the mic, not the speaker. For rounds 2-4, I used a laser measure and tape on the tweeter to get the distance right. Between all of the rounds, I added/changed filters.

First, the very rough sim made from traced measurements and run through vcad's baffle sim. I also use the baffle sim to generate the way off-axis data. I am shooting for this shape. The only filters I am using from my sim are for the XO. What is cool is that the sim predicts a dip off-axis for the XO and reality shows the same (xo at 1700hz).
eton3__t25b (hor).pngeton3__t25b plus seas XO-schema-1.png

Round 1: According to kimmo, a 0.5x wl xo has a bump after the XO. I think this is what I am seeing in these measurements at 2.4khz. Around 6-8khz is also a bit much. The mic was moved at random distances (this is why the off-axis is louder than the on-axis) and I am spitballing the angles.
Mixp_1ft_0degree_tweeter_3ms_20_width_in_cycles.jpgMixp_1ft_0degree_tweeter_3ms_no_window.jpgMixp_1ft_0degree_tweeter_3ms_6_width_in_cycles.jpg

Round2: For this round, I placed the mic 0.6meters from the tweeter. It looks like I got rid of the 2.4khz issue. I would say this covers 0-70 degrees with a big jump at the end because I was getting bored. I forgot to turn off filters for rounding off 10khz and above- this is why it rounds off for these measurements.
psy + no FDW w/ average MIXP_0_600meter_roughPOLAR_psyAVG.jpgpsy + no FDWMIXP_0_600meter_roughPOLAR_psy.jpg1/24th + no FDWMIXP_0_600meter_roughPOLAR.jpg1/24 + FDWMIXP_0_600meter_roughPOLAR_FDW.jpg

Round3: I measured 1 meter from the tweeter. This shows 0,45, and 90 degrees---ish! The mic stand was actually up against the right speaker to take a measurement of the left one. The tweeter is still too hot and I think I see a baffle step issue.
PSYMIXP_1meter_roughPOLAR_psy.jpgFDWMIXP_1meter_roughPOLAR_FDW.jpg

Round4: On-axis only. I started messing with the mid and woofer. I don't have any filters on the l26roy to boost bass.
1/24thmixp_eq6_24th.jpgFDWmixp_eq6_fdw.jpgdecaymixp_eq6_decay.jpg

I will do a few more rounds, but things are looking and sounding nice!
 
Last edited:
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
The new speakers are meant for the desk but I thought it would be good to compare them in another room to my JBL 580's.

The topping 10s outputs to spdif and analog at the same time so I was able to hook it to my receiver and fa123's. This allowed for instant swaps between speakers. For movies, I had the toppings 10s connected my nvidia shield via usb.

I also had hifijim from diyaudio over for a listen a few weeks back. I have never met anyone that listens to music like hifijim. He has an amazing ear that reminds me of: https://www.wnycstudios.org/podcasts/radiolab/segments/301427-head-full-symphonies

We listened to some tracks that I thought showed off the plusses of the system and some tracks that I have seen folks use to test speakers.
https://open.spotify.com/playlist/2x8t73JDQPNKW4yTLbPaO7?si=154d8f0340904cde

hifijim noticed my tweeter level was down. I smiled because I did this on purpose for my own ears and it was only a little down. He is damn impressive!

So between comparing the new speakers to the 580's and getting another person's opinion, I have more confidence in the sound. Now I just need to glue everything up and paint.
LR_jbl580vs.jpg
 
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
Now that I am getting close to being done w/ speaker 1, speakers 2-3 are in the works. One is a coaxial and the other a small full-range. I am thinking of a SEAS for the coaxial and the scanspeak 10f for the small full range. Jury is still out through. They will just sit on the L26ROY for bass.

The first graph is a comparison of data from SEAS and audioxcite for a tweeter on a coaxial. I am running both through the baffle sim in VCAD. It looks like I can trust 1-2khz of my sim which is fine for my XO.
kingL26roy_t Diffraction_traceCOMP.png

I like to ask VCAD questions like "is it worth it to add an edge radius to a coaxial?". Solid is 0 degrees, dotted is 30,60,and 90 degrees. The baffle sim says: "yes if I don't want to fuck up 1-3 khz, a 44mm edge radius will make things lovely." This data is just a sim based on the baffle size and SD of the driver. No trace data.
kingL26roy_t Diffraction_noedge.pngkingL26roy_t Diffraction_44mmedge.png

So I tested out a quick sim from the trace data using the baffle sim. I want a coax that falls off differently than my bliesma/eton 3way and this looks like it. Some more hacky sims should tell me which way to go.

seas3_r1SEAS Directivity (hor).png
 
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
3way_dist_omni_byhifijim.png


hifijim from diyaudio came over for a listen and to help confirm some measurements with his setup.

This is an omni mic distortion measurement of the 3way at about 0.6m. Looks good to me! I do expect some distortion around the tweeter's XO due to how low I am crossing if things were a bit higher. This is about as loud as I listen so I am good.

Thanks again, hifijim!!!
 

bigjacko

Addicted to Fun and Learning
Joined
Sep 18, 2019
Messages
721
Likes
359
View attachment 164116

hifijim from diyaudio came over for a listen and to help confirm some measurements with his setup.

This is an omni mic distortion measurement of the 3way at about 0.6m. Looks good to me! I do expect some distortion around the tweeter's XO due to how low I am crossing if things were a bit higher. This is about as loud as I listen so I am good.

Thanks again, hifijim!!!
Is this distortion graph correct? All distortions are cramped at the bottom. Or maybe the scale is too wide to see the small difference between distortions.
 
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
All distortions are cramped at the bottom.
The scale is not the norm for sure.

When he did the measurement, hifijim mentioned that he did not like how omni-mic generates reports. So this is a pic of the screen from the app. The key takeaway for me is the distortion is over -40db down at 80db.

I will take a REW measurement so you can see the distortion better and try to play it louder to get more above the noise floor.

..

My little speaker is changing! Well, not much. I am just moving it out a bit so that the mid has nothing to bounce off of at sensitive frequencies. The L26ROY has a big rubber surround and 1/4 WL spacing is not a good mix. After the change, I could hear greater clarity and sense of space with the mid. Listening to the small SB A coax speaker helped me track this down. The coax sounded bigger floor to ceiling until I made the change. The nail in the coffin was hifijim hearing issues as well. After he visited. I made the change the next day. It will be neat to see what changed in the measurements.

3way_cfg2_front.jpg3way_cfg2_back.jpg3way_cfg2_to_the_floor.jpg

Now I need to cut some edges to make it look pretty.
 
Last edited:

dualazmak

Major Contributor
Forum Donor
Joined
Feb 29, 2020
Messages
2,820
Likes
2,950
Location
Ichihara City, Chiba Prefecture, Japan
Hello @headshake,

This is really exciting and interesting project! Your objective approach towards the "ideal SPL for your ears" from zero basis (I mean without reference sound system) is very much impressive for me, since in my similar multichannel multi-amplifier project, I always have/preserve a "reference sound system" driven by single amplifier plus LC-network with Yamaha NS-1000 + subwoofer Yamaha YST NS-1000 + super-tweeter Fostex T925A.

Although your wonderful approach is quite different from mine, I really surprised and impressed looking at your latest whole SPL curve (at 0.6 m) is quite similar to my SPL measured at my listening position.

Just for your reference, I use a HiFi (budget) integrated amplifier directly driving super-tweeters so that the SPL over 7 kHz can be easily flexibly adjusted for age-dependent hearing decline in high Fq over ca. 6 kHz; very convenient and nice for listening sessions inviting friends. I usually use rather quiet violin solo music for their preferred SPL over 6 kHz at the beginning of our listening session.

In any way, much congratulations for your exciting project, and I am looking forward to seeing your further steady progress. I hope we may share and exchange invaluable know-how and information in our two project threads.

all the best

dualazmak
 
Last edited:
OP
H

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
238
Hello @headshake,

This is really exciting and interesting project! Your objective approach towards the "ideal SPL for your ears" from zero basis (I mean without reference sound system) is very much impressive for me, since in my similar multichannel multi-amplifier project, I always have/preserve a "reference sound system" driven by single amplifier plus LC-network with Yamaha NS-1000 + subwoofer Yamaha YST NS-1000 + super-tweeter Fostex T925A.

Although your wonderful approach is quite different from mine, I really surprised and impressed looking at your latest whole SPL curve (at 0.6 m) is quite similar to my SPL measured at my listening position.

Just for your reference, I use a HiFi (budget) integrated amplifier directly driving super-tweeters so that the SPL over 7 kHz can be easily flexibly adjusted for age-dependent hearing decline in high Fq over ca. 6 kHz; very convenient and nice for listening sessions inviting friends. I usually use rather quiet violin solo music for their preferred SPL over 6 kHz at the beginning of our listening session.

In any way, much congratulations for your exciting project, and I am looking forward to seeing your further steady progress. I hope we may share and exchange invaluable know-how and information in our two project threads.

all the best

dualazmak
Hello, dualazmak! Thanks for popping by my newbie project. Your system looks so amazing and flexible. It inspired me for sure!

re:"your latest whole SPL curve (at 0.6 m) is quite similar": That is super cool to read! That yamaha dome is amazing though. If you ever want to trade :p


...

A little update.

Nobody else will ever see this so I thought I'd post a bunch of pics of dots.

Round 2 of the perforated plate baffle is back on now that have measured the speaker in (what I hope) is the final shape. My plan is to make 3 of these plates to test out. The plates try to zap 1-3 frequencies in a few ways.
MPP_2200_2600_3000hz.pngMPP_half_1200_2700hz.pngMPP_hearing_2,6.pngMPP_hearing_2,8_and_4khz.pngMPP_hearing_2khz_to_4khz.png

Also, inspired by the yamaha dome, I am thinking about the next speaker based on what I learned from this project. It is the same thing but with a dome mid. I will also change the baffle shape a little since it wont need a box.

This is what I am thinking of so far:
tri_baf_384.pngtribaff_384_adj Diffraction.png
The baffle sim seems to like the driver playing 1-2khz right in the middle The tweeter would be crossed at 3656hz. The shape does not show the 44mm edge that I would add just like the speaker in this thread.

bliemsa_mid_xdir.png
Top: the dome mid sim
Bottom: bliemsa/eton sim
 
Last edited:
Top Bottom