Crossover design problems/suggestions

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

i once again need your help/suggestion for the design of a crossover. The frequency response curves of the drivers i want to use are not that flat, and i m not really sure how to solve these issues. The woofer i m using is the Monacor Sp-6/108Pro and for the high frequencies, i chose the Eminence Apt80. The frd measurements i did myself with my Umik-1 in REW, both the drivers were meassured with exactly the same gain/volume settings, but the microphone is not spl calibrated. The Apt80 has a horn, so the source of the sound is about 2 inches behind the woofer. Eminence recommends 3.5khz with a second or third order highpass as the lowest crossover point. My goal is to keep the efficiency as high as possible, as the speaker will be used for a small sort of Pa Setup with one or two 12" Subs. Form my point of view the frequency response of the Atp80 is extremely difficult to flatten out and that's where i need your help.

Thanks in advance.
 

Attachments

  • Files_Speaker.zip
    926.8 KB · Views: 46
  • 4mDUkpY.jpg
    4mDUkpY.jpg
    475.1 KB · Views: 394
Last edited:
Well, you can try this quickly tossed-together one (XSim design file is here: http://www.diyaudio.com/forums/attachment.php?attachmentid=668167&stc=1&d=1520897421).

attachment.php

attachment.php

But I gotta tell you, you really should get your phase response files accurately aligned (about "2 inches" really isn't very accurate, and the placement of the driver magnets doesn't really give the whole information needed). Particularly with such a high crossover frequency (3500Hz has a wavelength of less than 4 inches, so to be within 1/8 wavelength you need to be right to closer than 0.5 inch).

Maybe see Jeff Bagby's how-to on getting measurements for this at http://techtalk.parts-express.com/filedata/fetch?id=1149302
 

Attachments

  • quick ckt try.zip
    945.7 KB · Views: 62
  • quick ckt.png
    quick ckt.png
    10.7 KB · Views: 945
  • quick ckt resp.png
    quick ckt resp.png
    16.5 KB · Views: 904
Hi,

Here's another kick at the can.

attachment.php
attachment.php


The tweeter distance is set to "0".
- If the offset of the tweeter is truly 2" behind the woofer then you would need to flip the polarity of either the woofer or the tweeter.

This design tries to maintain maximum efficiency ( at the expense of the low-end ).

Download the "zipped" dxo file that I have included ( place it in the same folder as your current zma & frd files ) and then play with it.
- Do the same with the one Bill posted.
- Use the best ideas from each & see what you can create on your own.

There are a couple of notch filters in mine ( one in each the top & bottom sections ).
- See if you can eliminate them in your actual implementation .

:)
 

Attachments

  • 6in PA Speaker_Schematic_.png
    6in PA Speaker_Schematic_.png
    9.6 KB · Views: 861
  • 6in PA Speaker_Trace_.png
    6in PA Speaker_Trace_.png
    21.9 KB · Views: 485
  • 6inch PA Speaker_modE.zip
    945.4 KB · Views: 23
Last edited:
In the text from Jeff Bagby's it is mentioned to connect the drivers in parallel while remaining the polarity, but the problem ist my amp is not 2ohm stable, so is it possible to connect a 4ohm Resistor in series to the two drivers connected in parallel to keep the overall impedance of two 4 Ohm drivers above 2 Ohm ?
 
In the text from Jeff Bagby's it is mentioned to connect the drivers in parallel while remaining the polarity, but the problem ist my amp is not 2ohm stable, so is it possible to connect a 4ohm Resistor in series to the two drivers connected in parallel to keep the overall impedance of two 4 Ohm drivers above 2 Ohm ?

That's really a question you could get closer to an answer ( on your own ) by running a Xsim of the 2 drivers ( paralleled ) with a 2 ohm resistor put inline between the source and the load.

Anyways, it's not ideal, nor perfect ( due to how the drivers react to losing amp loading/damping ) but yes it's doable.

2 ohms is the best value ( to minimize response changes & still meet the 4 ohm load criteria ).

See for yourself ( in XSim ) how a 10R resistor changes the relative curves of the individual drivers.

:)
 
Last edited:
Hi sese, you have got the drivers impedance in simulation so put these in parallel and you will find that summed impedance is not a problem, plus if you set the driving voltage to 1Vrms, what's the worse current the amp might send through these, a couple of hundreds of mA's perhaps for as long as it takes the software to measure response which is normally a short period. Yes, no?
 
You can actually drive about any amp into very low impedances provided you don't try to drive hard. Amplifiers are damaged by excess power dissipation, and dissipation is minimum when output is minimum (except SE amps, which disippate most at no signal, unrelated to output level). The amp won't be at its best in terms of distortion (which should be low at low level anyway) and some amplifiers can get twitchy with very low impedances, particularly if low at highest frequencies (capacitive load). But with the vast majority of amps you should be fine putting some milliwatts or even a watt into single ohm type loads. Just be careful with the level control, and remember, if it sounds bad turn it down or off!
 
...Download the "zipped" dxo file that I have included ( place it in the same folder as your current zma & frd files ) and then play with it.

EarlK,

-- the FRD files and ZMA files get embedded into the dxo files, so there is no need to find them again when you get an XSim design file. Should be able to just unzip the dxo file and start it up from any folder. I had hoped that diyAudio would allow dxo files to be directly attached (wouldn't it be cool to just double-click on a dxo file in a post and the design simulation pops up on-screen?), but it's not a 'supported file type', so it has to be zipped before attaching.

Bill
 
In the text from Jeff Bagby's it is mentioned to connect the drivers in parallel while remaining the polarity, but the problem ist my amp is not 2ohm stable, so is it possible to connect a 4ohm Resistor in series to the two drivers connected in parallel to keep the overall impedance of two 4 Ohm drivers above 2 Ohm ?

Wouldn't it be a solution if you connect the tweeter to the right output, mid-woofer to the left output, and put the same signal on both inputs? A "mono" button would be useful, if you have one in your amp.
 
I already did the measurement, but i totaly screwed up. I did the measurement of the tweeter on axis with the tweeter, the measurement of the woofer on axis with the woofer, and placed the mic between the drivers, to measure both of the drivers connected in parallel. I overread that all of the measurements need to be taken without moving the mic, so my measurements are not usable for calculating the offset. The amp did just fine with both of the drivers connected in parallel, but the idea to use both channels is also worth a try. Thanks for your help, i will post new and hopefully correct measurements tomorrow.
 
My Advice?

To confirm that 1.5 inch ( offset ) figure;

- Create a new network prediction in your design software utilizing only LCR parts that you currently have on hand ( the test network can really be anything that doesn't smoke your amplifier ).
- Wire up your "test network" to your drivers.
- Measure the resulting curve of one speaker.
- Compare "software prediction" to "reality" .

- If you make these measurements from the same exact location that you used to make the new files ( used to figure out the "offset" ), then the curves should to be the same.

:)
 
Last edited:
Hi,

Here's my final effort for you ( using your 1.5" offset figure ).

HF has been increased ( above 10K > though usually that's not that critical for a PA speaker ).

attachment.php


- The red trace uses "Auditory ERB" smoothing which offers a good representation of human hearing ( resolution wise ).

- I've zipped and included the XSim file for you to play with.
( the woofer section has 2 notch filters > see if you can find a way to eliminate one or both of them ).

:)
 

Attachments

  • 6in PA Speaker_Trace_modL.png
    6in PA Speaker_Trace_modL.png
    15.9 KB · Views: 314
  • 6inch PA Speaker_modL.zip
    945.4 KB · Views: 25
Last edited:
So i tried some things in Xsim, and thats what i came up with :

x97nRmU.jpg


U2DudK9.jpg



Hi,

Here's my final effort for you ( using your 1.5" offset figure ).

I overlayed them both, and then the superiority of your crossover design gets really obvious. (the red one is my design)

UKT2v3l.jpg


Why should i try to eliminate the notch filters, do they have any bad impact on the sound ?

Thank you so much for your effort in helping me, i m just amazed how great and helpful this community is. Thank you.
 
You're very welcome !

Since you had frd & zma files, I was glad to help out.

Notch Filters ?
There's nothing much controversial about the selective use of notch filters ( apart from the fact they use more components ).

attachment.php


The two here ( in the woofer section ) are located out in the stop-band helping to steepen the slope of the low-pass filter.

:)
 

Attachments

  • 6in PA Speaker_Schematic modM_.png
    6in PA Speaker_Schematic modM_.png
    10.8 KB · Views: 297
Yet another "90dB" approach. What I see as an advantage is that the circuit has only 3 coils, so there should be no problems with separations, directions etc. The elements are easy to acquire: the coils are from JantzenAudio list, the rest of the elements have standard values.
 

Attachments

  • mo_em_1.5i_4r_3L_5C.jpg
    mo_em_1.5i_4r_3L_5C.jpg
    342.3 KB · Views: 248
  • mo-em-1.5inch_ph_meas.zip
    945.1 KB · Views: 19
Maybe another tweeter

I tried to confirm my offset, and i tried 3 times and got three different results all the way from 1.4" to 2.1". I m thinking of replacing the Eminence atp80 with another tweeter that has a "simpler" response curve and not so many dips. As Peerless by Tymphany drivers are now finaly avaliable from sound imports, it is much easier and cheaper to buy them in germany, and i think the DA25BG08-06 would be the best choice. Additionaly i could use a much lower crossover point, which would simplify the crossover section of the woofer.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.