Trinium, my latest build

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi All,

I like to share my latest build. It's a active 2 way based on the Hypex PSC2.400d modules.

The PSC2.400d is a plate amp loaded with very nice features; 2x UCD400 amplifiers, a SMPS and a powerful DSP.

The drivers are 2 Seas W15LY001 midwoofers and a Fountek NeoCD2.0 ribbon tweeter in a MTM. The crossover is LR8 @ 3.5KHz time aligned.

The cabinet is a closed design with a Q of 0.71. The front is made of 36mm MDF (2x 18mm glued), the rest is 18mm MDF. Internally the cabinets are damped with
"loodbitumen" (2mm) this stuff is also used in car doors, pritex (4cm) and angel hair (100g).

The "filter" thumbnail is not really a filter but more of a safety measure for the tweeter. It's a 50uF capacitor followed by 2 5.6Ohm resistors in parallel. This provides protection against DC and some attenaution and does not influence the filter inside the DSP.
 

Attachments

  • DSC_0003.JPG
    DSC_0003.JPG
    612.6 KB · Views: 1,792
  • DSC_0050.JPG
    DSC_0050.JPG
    461.6 KB · Views: 451
  • DSC_0048.JPG
    DSC_0048.JPG
    714.2 KB · Views: 372
  • DSC_0042.JPG
    DSC_0042.JPG
    614.6 KB · Views: 361
  • DSC_0039.JPG
    DSC_0039.JPG
    766.7 KB · Views: 550
  • DSC_0034.JPG
    DSC_0034.JPG
    811.9 KB · Views: 624
  • DSC_0023.JPG
    DSC_0023.JPG
    564.4 KB · Views: 1,378
  • DSC_0020.JPG
    DSC_0020.JPG
    601.2 KB · Views: 1,395
  • DSC_0014.JPG
    DSC_0014.JPG
    575.6 KB · Views: 1,478
  • DSC_0005.JPG
    DSC_0005.JPG
    678.3 KB · Views: 1,598
Last edited:
More photo's.

I live in a really small apartment so my living room doubled up as a workspace for building these speakers... I don't think i've ever vacuum cleaned my living room as much as I did while building these speakers :p


On a side note: I would like to give Hypex a big :up: . I thought I had a problem with 1 of my modules, I contacted them and got a reply the same day to send the unit back for testing, I did and only 2 days later I had the unit back. They checked it out and said it was oke and the problem was most likely due to something going wrong during the upload of the filters to the DSP. And indeed when I installed it it worked. Very good service!
 

Attachments

  • DSC_0060.JPG
    DSC_0060.JPG
    717.8 KB · Views: 566
  • DSC_0055.JPG
    DSC_0055.JPG
    680.5 KB · Views: 563
  • DSC_0057.JPG
    DSC_0057.JPG
    504.9 KB · Views: 464
  • DSC_0029.JPG
    DSC_0029.JPG
    767.6 KB · Views: 482
  • DSC_0062.JPG
    DSC_0062.JPG
    781.3 KB · Views: 625
Last edited:
Thank you all,

Hi Piersma,

Yes indeed I wired the 2 midwoofers in parallel for a 4 ohm load. I just finished the last speaker today so no detailed measurements on SPL capability yet. In theory the pair should be capable of ~116dB at 1M.

Plenty of power in the amps to try this someday, but for now I am tweaking the filters.

The speakerstands are very HiFi, It really takes the sharp edge of the tweeters :D
 
Last edited:
Kind of like bringing a knife to a gun fight isn't it? Impressive on it's own but not intended for that.

You're right, they are not designed for very high SPL capabilities. This is the number that rolled out of the sims I did.

They are designed for listening to music and the occasional movie, and with a narrow vertical polar response to suite my room (which is only 20m2).

For the bottom end they have support of a 15'' subwoofer crossed over @80Hz LR4


Edit: just did a very quick measurement which peaked at 102dBc, sounded the same as at 85dBc just a lot louder. Don't want to go much louder than that, and my neighbors agree with that :)
 
Last edited:
Been tweaking the filters the last couple of months, tried different slopes and frequencies. EQ-ing the drivers flat before filtering, time alignment, and a lot of other stuff. I really like the DSP in this sytem now :) The "sientific" aproach rather then trusting my ears. Only when I was finished measuring and setting filters I really took the time to listen.

I ended up with a LR8 @ 3.1KHz, a notch on the tweeter @16.5KHz (Q2.9, G-2.2dB) a notch on the woofers @ 5KHz (Q2.9, G-3.0dB)and a baffle step compensation of 5.5dB.
 
The drivers are 2 Seas W15LY001 midwoofers and a Fountek NeoCD2.0 ribbon tweeter in a MTM. The crossover is LR8 @ 3.5KHz time aligned..

Hi Mark,

Excellent work with top level drivers and the latest amp_DSP technology.

You could contribute to the great "ribbon tweeters in MTM" design debate! Do you plan any off-axis measurements?

1) side one: only very short ribbons crossed at low'ish frequencies to very small diameter midranges can meet the M-T and M-M spacing math for good lobing control.
2) side two: the ribbon M-T spacing should not be measured from the center of the ribbon, but rather near the edges. Adequate lobing control can be achieved with long'ish ribbons and small diameter midranges.

The math behind reducing lobing in a MTM topology requires very close M-M and M-T spacing. Crossover_wavelength/4 is the ideal for M-T spacing and Crossover_wavelength/2 is the limit for M-T and M-M spacing.

If you believe the math, even for ribbon tweeters, then only very short ribbons like the RAAL 70-20, crossed at low'ish frequencies like 1.4-1.6Khz to small 4"-5" midbass can maintain adequate lobing control.


One decent reference:
Vertically Symmetric Two-Way Loudspeaker Arrays Reconsidered
MITHAT F. KONAR, AES Member
Biro Technology
 
Hi LineSource,

Thank you!

Yes I know vertical lobing is a problem with such a setup. Did a lot of sims with WinISD, Boxsim, Edge, Xdir and AADA before starting this build. In the end I just wanted to try a big ribbon to see what they can do with good amps and DSP filtering.

With these drivers you need a 3KHz minimum XO to avoid distorsion problems with the tweeter. (hence the Nextel woofers in stead of Magnesium) Which of course translates to vertical lobing problems and comb filtering effects. I am playing with the idea of a NTM type crossover to push the XO frequency a bit lower.

In my setup I don't really care about the vertical lobing because I just sit down in 1 place to listen. Heck I even use a modded DEQ2496 for some room correction... attached an early Boxsim sim (no DSP data included), it gives a rough idea of the polar response and comb filtering effects.

Edit: because of the vertical response pattern of this tweeter you will always have lobing, no matter what type of driver arrangement you build!
 

Attachments

  • vertical horizontal response.JPG
    vertical horizontal response.JPG
    266.7 KB · Views: 613
Last edited:
And the response plots of the planned system when completely done (active 3 way with 2X Dayton Audio RSS265HF-4 per side added)
 

Attachments

  • 2pi response.JPG
    2pi response.JPG
    199.1 KB · Views: 617
  • vertical horizontal response full.JPG
    vertical horizontal response full.JPG
    268.4 KB · Views: 598
  • Filter response.JPG
    Filter response.JPG
    192.1 KB · Views: 577
Last edited:
diyAudio Moderator
Joined 2008
Paid Member
I like the overall response in the 200-600Hz region for reducing floor and ceiling reflections down where they're harder to deal with.

But looking at the horizontal response I notice also that between 200-2k, the zero degree axis is higher than the overall response. So this might benefit from the right toe-in.
 
Ended up pushing the XO frequency between the mid and tweeter a little higher. It's now at 2.7KHz. The tweeter likes this a lot better, distorsion wise.

The 4 notches stayed, 1 on the woofer to compensate for the rising response and 1 one the tweeter to tame a resonance around 1200Hz. The other 2 notches on the tweeter are there to flatten the response up to 22KHz.

The XO itself is a LR8, and the baffle step compensation is 5.5dB. The (sub)woofer section is still a sim only but it looks like that when I take into account the room gain I get a frequency response that covers ~ 20Hz - 22KHz.
 

Attachments

  • Trinium filter 17-07-2012.JPG
    Trinium filter 17-07-2012.JPG
    214 KB · Views: 138
  • Trinium filter response 17-07-2012.JPG
    Trinium filter response 17-07-2012.JPG
    207 KB · Views: 134
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.