MiniDSP H-Frame.

rich31td

Member
2011-06-21 5:58 pm
Kent
I have been building an H-frame dipole woofer that is based on the Orion's woofer.

I am using the same Peerless 830452 10" woofer and same dimensions for the H-frame as the Orion but have used a single woofer each side.:(


[IMGHTTPDEAD]http://imageshack.us/a/img715/9403/hframepic.jpg[/IMGHTTPDEAD]

I have been programming the MiniDSP today.
I first applied a LT biquad using the woofers TS specs as a reference and a target of 35hz, Q0.5.
Was I right to use the drivers FS as one of the poles?


[IMGHTTPDEAD]http://imageshack.us/a/img831/8433/hframeexcel.jpg[/IMGHTTPDEAD]
I then tried to add the 6dB/oct dipole compensation, 35hz-305hz, and soon found the MiniDSP was clipping. So I tried splitting the EQ with one low shelf and one high shelf and splitting the difference regarding the Gain. The crossover is a 100hz LR4. Again, is this a good way to do it?


[IMGHTTPDEAD]http://imageshack.us/a/img198/6347/dsphframe.jpg[/IMGHTTPDEAD]
And here is the outcome measured at the opening to the front of the H-frame. Any comments or advice is welcome.



[IMGHTTPDEAD]http://imageshack.us/a/img690/3503/hframe.jpg[/IMGHTTPDEAD]
 
Last edited:
Rich,

If you're measuring at the H-baffle opening and have the 6db/octave dipole correction applied you should see a boosted response which is roughly equivalent to that correction. When you pull your microphone back and take further measurements at a more nominal distance then you would observe the measured response "flattening" to approximate your overall target.

I think I would approach this in two steps. Start by placing your microphone near the dust cap of the woofer and correct its response first to approximately achieve your 35Hz/0.5 target. Then move your microphone away from the enclosure/baffle and apply an estimated dipole correction to the miniDSP and see how the response looks. You can fiddle with the miniDSP settings to fine tune it. You don't need to use the "advanced" features of the miniDSP programming....you can easily achieve your target using standard shelving filters in the "basic" configuration.

The LT spreadsheet doesn't do you that much good in this instance since it's designed for a different application.

Cheers,

Dave.
 
Last edited:

rich31td

Member
2011-06-21 5:58 pm
Kent
That makes sense, thanks Davey, was I right in using SHP and SLP filters together to avoid clipping?

I shall try that today, I would imagine that is the same way to measure the mid range EQ?

And just for the sake of it I attempted to program the Minidsp with the figures that are on SL's ASP diagram;
Driver EQ: 20-110
Dipole EQ:20-305

and this is what I got,

[IMGDEAD]http://img37.imageshack.us/img37/4209/orioneq.jpg[/IMGDEAD]
 
Last edited:

john k...

Member
2004-08-10 2:50 am
US
FWIW: Gradient woofer equalization

I don't particularly care for targeting a 2nd order high pass response for a dipole woofer as it means excursion will keep increasing below the cut off point. 3rd or 4th order limits excursion below cut of, but does increase group delay. However, when you look at the GD it is apparent that the major contribution to the woofer GD above 60 Hz low pass filter between woofer and mains.Most recently I have determined that the easiest way to obtain a 4th order high pass alignment is with a single stage Q boost and a single stage Q cut. For the woofer you are using and a 35 Hz cut off, a boost with Fc = 37 Hz, Q = 0.75 and Gain = 14dB, and a cut with Fc = 325 Hz, Q = 0.45 and gain = -12.5 dB should get you close. Then couple with an LR4 low pass to the midrange. [note: the Q's specified are the analog Q's and are not exactly the same at what miniDSP uses. You will need to fine tune.]
 
Hi Richard, nice speakers ;)

Equalising H-frame (and any subs) is quite easy. What your spreadsheet is missing is the "measured" graph. I've modified that one to include it (will send you when I get home).

Basically:
- Measure at the opening plane of the H-frame
- obtain the .frd and import into the spreadsheet
- 'guess' the Fs and Qts (grey line below matching blue)
- punch in the target Fs and Qts
- copy paste in MiniDSP :)

After the LT is applied, then apply the dipole SLP. Normally this is simply 6db/oct from xo frequency to 20hz.


yx3Qh.png
 
FWIW: Gradient woofer equalization

I don't particularly care for targeting a 2nd order high pass response for a dipole woofer as it means excursion will keep increasing below the cut off point.

Interesting. I've never actually tried a higher order highpass for the subwoofer. Always equalised with target response of 20hz, Q=0.5 which sounds great but can be SPL-limited with certain recording/movies.

So this morning I tried to have 24db/oct highpass at 20hz for the sub. To do this by following the prescription: linkwitzlab.com/filters.htm#2

- First, I equalise the woofer for F=20hz, Q=0.707 using Linkwitz Transform
- Then add another biquad for Highpass filter, F=20hz, Q=0.707



Having it for a few moments before going to work, I heard nothing wrong with this. Perhaps group delay is not that audible. I will observe further and it it's good then more SPL can be obtained !! :cool:
 

john k...

Member
2004-08-10 2:50 am
US
Interesting. I've never actually tried a higher order highpass for the subwoofer. Always equalised with target response of 20hz, Q=0.5 which sounds great but can be SPL-limited with certain recording/movies.

So this morning I tried to have 24db/oct highpass at 20hz for the sub. To do this by following the prescription: linkwitzlab.com/filters.htm#2

- First, I equalise the woofer for F=20hz, Q=0.707 using Linkwitz Transform
- Then add another biquad for Highpass filter, F=20hz, Q=0.707



Having it for a few moments before going to work, I heard nothing wrong with this. Perhaps group delay is not that audible. I will observe further and it it's good then more SPL can be obtained !! :cool:

Try playing vinyl on a dipole woofer system with 2nd order alignment. :)
 
John K.. you wrote "Then couple with an LR4 low pass to the midrange". Martin King utilizes a LR2 filter to flatten the rising response of the H-(or U-) frame. Do you prefer the LR4 for a particular reason?

Since I haven't understood most of the discussion can you suggest a good primer for a noob on digital crossovers?

Thanks in advance!
 

ttan98

Member
2006-04-04 11:24 am
Melb
FWIW: Gradient woofer equalization

I don't particularly care for targeting a 2nd order high pass response for a dipole woofer as it means excursion will keep increasing below the cut off point. 3rd or 4th order limits excursion below cut of, but does increase group delay. However, when you look at the GD it is apparent that the major contribution to the woofer GD above 60 Hz low pass filter between woofer and mains.Most recently I have determined that the easiest way to obtain a 4th order high pass alignment is with a single stage Q boost and a single stage Q cut. For the woofer you are using and a 35 Hz cut off, a boost with Fc = 37 Hz, Q = 0.75 and Gain = 14dB, and a cut with Fc = 325 Hz, Q = 0.45 and gain = -12.5 dB should get you close. Then couple with an LR4 low pass to the midrange. [note: the Q's specified are the analog Q's and are not exactly the same at what miniDSP uses. You will need to fine tune.]


John,

Pardon my ignorance why do you need the cut at 325 Hz, Q=0.45 and gain=-12.5db? What happen if the OB between the woofer and mid-woofer is crossed around 300Hz?
I having some difficulty in crossing my woofer and mid woofer accurately(currently around 300Hz) and also boosting around 35-40Hz and how to go about it. I am using the Behringer DCX2496. FYI my woofers(twin) is mounted in a U-frame.

An explanation would help and some guidelines to better interface the woofer to mid woofer, any inputs from anyone is encouraged. Thanks.
 
The MiniDSP does not have a very high max output level. You can easily cause internal clipping when using boosting functions like the Linkwitz Transform or other EQ. One way to avoid this is to use the Active Crossover Designer to create advanced biquad coefficients for your speaker. In the stage the will be boosting the level you simultaneously specify gain reduction equal to the maximum boost, essentially creating an analogous "cut" filter. This will prevent internal clipping, however you will need to make up the lost gain elsewhere, e.g. at the amplifier or with an external analog gain stage.

Get ACD here:
the Active Crossover Designer web page

-Charlie
 

john k...

Member
2004-08-10 2:50 am
US
John,

Pardon my ignorance why do you need the cut at 325 Hz, Q=0.45 and gain=-12.5db? What happen if the OB between the woofer and mid-woofer is crossed around 300Hz?
I having some difficulty in crossing my woofer and mid woofer accurately(currently around 300Hz) and also boosting around 35-40Hz and how to go about it. I am using the Behringer DCX2496. FYI my woofers(twin) is mounted in a U-frame.

An explanation would help and some guidelines to better interface the woofer to mid woofer, any inputs from anyone is encouraged. Thanks.

The eq I suggested is correct for the XLS woofer in a 18" H frame. It should be close for H frames within a couple of inches either way. The two stages serve to shape the response to a 35 Hz cut off, correct for the dipole roll off and bring the sensitivity to that of the woofer operating as a monopole in 2Pi space (the manufactures specified sensitivity) at the dipole = monopole frequency. This is simpler that using a couple 1st order shelving filters and a Q boost to raise the system Q to other than 0.5 and then attenuating to correct the sensitivity.

I would expect in your case you would have trouble if the U-frame is also on the order of 18" as the on axis response peak (cardioid peak) would be at about 185 Hz. If the U-frame is correctly damped and on that order in length, it is only suitable for use to about 125Hz.

My recommendation has always been to low pass the woofer at least 1/2 octave below the "dipole peak" frequency and to place the low frequency cut off so that there is no more that 20dB gain max at the cut off frequency. If you look at an 18" dipole the dipole = monopole frequency is about 125 Hz. 35 Hz is about 2 octaves below that requiring 12dB gain at 35 Hz for the dipole roll off and some additional to bring the driver Q to the desired level netting the +14dB. The cut at higher frequency is close to the dipole peak frequency (376 Hz ideally) and the cut is to compensate for the continued rise above the dipole = monopole frequency to the dipole peak. -12 db turned out to be what was required. The Q of each stage was what was required to make the response basically "flat" above the woofer cut off. Then the LP filter is for the crossover to the midrange.
 
That's five. Maybe six.

:)

But seriously, I would have thought EQing an H-Frame sub is one of the easier DIY dipole tasks, especially compared to EQing the midrange which is still bugging me after ten years.

If the H-frame is the same width and depth as the Orions, with the same driver, it should require the same EQ.

As for a high pass, I've spent more time than I will ever get back trying and measuring different high passes. Some to a monopole sub below 40Hz. Some at 20Hz or so to protect the drivers. There is certainly a measurable increase in Group Delay with some, but I would be lying if I said it was audible once you correct for phase and response.

Just last night I wasted time checking the difference between my prepros 12db and 24 db sub crossover. Big difference at first. Until I remembered to flip phase.
 
Last edited: