Stacking H-Frames in 2 Way?


2011-08-20 6:02 am

I have recently built four 18" HFrames for my system, and am utilizing them with a pair of Visaton B200.

In trying various configurations, it was asked what my thoughts were on trying an MTM type setup, where a single full range driver sits between a pair of HFrames, all stacked vertically. I am interested in the idea but have a few questions.

This won't be a traditional MTM and more a W(FR)W... So what would the fast and loose rules be on spacing? If I go as close as possible, the FR driver loses height and sound will suffer. I was thinking maybe 18" or so of a baffle, with the driver centred in height and width.

Would moving a driver towards the inside edge yield predictable results?

Or is stacking these just stupid?

I believe you want to cross over at 200-300 Hz. This will be ok with the geometry of three 18" baffles stacked one upon the other. You will even be able to move the B200 up on its baffle to 75-80 cm (including the height of the H-frame below). You want the plane of the B200 baffle in line with the front of the H frames. This would still be ok imho.



2011-08-20 6:02 am
Of course... I keep forgetting that bit. Still not sure how it works figuring electric/acoustic crossovers. I am not as well versed as you, but come to learn, my ears are suspect due to lots of playing in bands, and lack measurement tools... sometimes ignorance is bliss. until someone comes over to hear themselves haha. Youre far more technical than i and likely have higher standards, I wish you lived up the street... I only pay in beer though. Your choice of brand :D

Then again, your high standards may preclude you from socializing with me ;)
The acoustic crossover is what you get when you combine the frequency response of your driver and cabinet combination, with some kind of electrical filter.

For example, if I've got this driver (scroll down for frequency response) mounted in an infinite baffle (for simplicity)...

If I add a crossover of, say, 1st or 2nd order at 1kHz, there's a peak above 1kHz that won't be attenuated much. So the woofer might to go 1.5kHz before rolling off.
Electrical XO = 1kHz, acoustic = 1.5kHz.

A very steep electric crossover will have an acoustic crossover similar in frequency.
Lets say I go a bit OTT and put an 8th order crossover (48dB/octave) at 1kHz on that driver, the peak is going to be attenuated a lot, so that the magitude of the peak becomes tiny compared to the attenuation I'm putting on it.
The result is electric crossover = 1kHz, acoustic crossover = ~1kHz.

Steep filters have their own set of problems, though, and shouldn't be used without understanding their problems and deciding a suitable compromise.



2011-08-20 6:02 am
Right I get slopes... So is electric a crossover that assumes a flat driver response?

So the divers peak at a certain frequency will reshape a curve such that acoustically it performs at a different level, because the electrical is not necessarily taking the frequency response into consideration?
I haven't seen any data or theoretical argument that supports any benefit from a W-FR-W arrangement. I think stacking H frames on their own and putting FR on a separate narrower baffle of its own would be the best solution.
Is your crossover active? Are you using any EQ on the woofers?
What size room?
If you are passive and no EQ: you might be okay with single H frame per side and you can place your FR on top without much ill effect.
If you are active with EQ capability and very keen on W-FR-W, I would suggest getting rid of H frames and putting all drivers on a flat baffle with very strong vertical bracing (like Jamo).
Then again, your high standards may preclude you from socializing with me ;)
If we are talking about having a beer together, that could be true, yes.
List of countries by beer consumption per capita - Wikipedia, the free encyclopedia
I wouldn't want to die of dehydration while waiting for you to empty your glass. :p:D
But seriously:
If you could give me the rough dimensions of your H baffle and the TSP of the driver, I could probably show you how the driver curve and the xover slope work together. Is your 2nd order low pass an electronic device or just a coil and cap?

Hi mortron,

I see that you built the ubiquitous MJK Goldwood frame. If you look at its response without a xover, it is the red line in the diagramm:
H frame1.gif
The same driver in an infinite baffle would have a rather linear horizontal response line at ~94 dB in this diagram. If you put a driver in a OB, it will develop a dipol peak at 6dB above the infinite baffle response and a 6dB/octave roll-off below that peak.
The dipole peak is at 200-300 Hz and the roll-off below 200 Hz. The blue line marks a 6 dB/oct roll-off. I put it below the actual response, because I didn't want to clutter things. The Goldwood in its frame follows it down to ~25 Hz. Why not lower?

Look at the impedance curves for the Goldwood in infinite baffle (dashed red line) and in the H frame (full red line):
H frame2.gif
Both impedance peaks are at the resonance frequency of the driver. For the IB it is the advertised 30 Hz, the H frame lowers it to about 23 Hz. You don't want to EQ a driver in a dipole below its resonance frequency. So this is the lower response limit.

The green line in the upper diagram shows the application of a second order low pass filter at 60 Hz. MJK recommended a 100 Hz xover frequency to have a linear response over most of the range. 60 Hz is a bit too much, but this is only a simulation. Notice how you dont get the full 12 dB/oct of a 2nd order filter. You compensate the 6 dB roll-off, but don't have an additional full 6 dB - it's more like a 9 dB/oct filter.

Above simulation does not yet include the floor influence. I come up with that and the added Visaton later.


In regards to beer consumption: I admire every citizen who does everything possible to raise the status of his country in the world. I'm drinking in the same spirit ;). Cheers ...
Since we are talking about such a dry topic, I try to keep the discussion wet :D.

How does the response of the H frame change, if we place it on the floor? Red is the response without floor, green is with floor:
H frame3.gif
We get a gain of 6 dB at the lowest frequencies, which dwindles with rising frequency. May be we like the bass boost - may be not. In the last case we would have been better off with the 100 Hz low pass.
Now to the Visaton in a 20x20 inch OB:
H frame4.gif
Continuing from the previous post ...
Red is the B 200 response with the 250 Hz highpass. It is 10-12 dB louder than the H frame on the floor (green) and needs to be attenuated - like in the blue line.

For the next diagram I moved to 3 m distance from the loudspeaker. Blue is the lower H frame, red a second H frame above the B 200 and green the B 200 itself:
H frame5.gif
Floor reflections are better visible now: For the upper H frame at 300 Hz and for the B 200 at 400 Hz and upwards. We need to add up both H frame SPLs for a combined response, that will be 3-6 dB above the single lines. It will probably be better matched to the B 200 in level.