low frequency, high pass filter

[memphissound]

Newbie here.

I'm looking at some speaker box designs and noticed that a lot of them recommend using a steep high pass filter set to a really low frequency -- something like 50Hz. I know it's to protect against overexcursion in a vented box.

But is this just a typical 4th order filter?
Should it go before or after the x-over?
Does it effect other elements of the design, and if so how do you compensate for it?

Thanks.

peace,
memphissound <><

[richie00boy]

It could be anything from 2nd- to 6th-order depending on the specific application.

It will have to go before the crossover if a passive crossover is employed for the speaker. If an active crossover is employed it could go before or after the low-pass filter, but as it will slightly affect the whole signal it's usually best practice to put it before everything to ensure all drive units see the effect and thus are 'aligned'.

[infinia]

Over excursion on vented designs.
Is it a problem you've experienced? or just a potential worry.

It's only really a problem when trying to push an under designed system with more SPL's. The tradeoff while adding an additional highpass filter to optimize for max SPL's is the limiting of LF extension. Theory states the steeper the HP filter, potentially more of original LF extension can be preserved. Beware, by adding steep filters can also change the characteristc bass sound as well.

Is it a real potential problem- Use Unibox program to examine your particular vented systems excursion vs. frequency at various power levels. You can import a HP filter via FRD files to see the possible fixes.
Sometimes an easy solution is judiciously changing the built-in DC blocking caps on the preamp and poweramp inputs to get at least a 2nd order HP function.

[facundonu]

i was about to say what you just said at last.
i've had this problem with a small subwoofer pushed really hard for the kind of driver. it reached excursion limits very often.
i just changed some caps values in the amp and preamp and the problem was solved. DC blocking in the output of the amp, ac coupling caps in the amp and preamp feedback, etc. all values were figured to match a 40Hz pole.

[HiFiNutNut]

Richie00Boy,

It will have to go before the crossover if a passive crossover is employed for the speaker. If an active crossover is employed it could go before or after the low-pass filter, but as it will slightly affect the whole signal it's usually best practice to put it before everything to ensure all drive units see the effect and thus are 'aligned'.

I wonder why in an active circuit it appears that we can add a filter either before the XO or after the XO. Why is it different from the passive circuit? I know for sure in a passive XO we can't.

I have observed this from the implementations of others, including John K and S.L.'s. In the Pheonix circuit, S.L. put various filters into the "branches" after the XO. So as John K's NaO for the woofer EQ. If the masters can do it that way, I guess I can too. Just make sure that I indeed understand them correctly. I also observed S.L. put two all-pass circuits into the tweeter of the Pheonix as delay circuits.

My preference is that less components in the signal path the better, especially in the HP section where human ears are more sensitive. In an active circuit, I would have at least (1) a woofer EQ circuit; (2) woofer cone breakup notch; and (3) baffle step compensation. If (1) and (2) can go after the XO, the tweeter and the midrange will have less opamps in the path and will sound better! As we understand, this won't work in a passive XO unless the XO is redesigned. But why it is OK in an active XO? Would (1), (2) and (3) introduce big phase shift? just checking from LTSpice, my active woofer EQ introduces a 18 degree phase shift. I can understand that in the active circuit, the impedance of the drivers is out of the question. But is this the only reason?

You said "slightly". How slightly?

Please explain a bit more so that non EEs like us pure hobbists can learn a bit more.

Many thanks,
Bill

[Dave Bullet]

Hi memphissound,

Vance Dickason has a sample passive high pass network in his loudspeaker design cookbook. Not sure how steep the filter, filter Q nor cutoff frequency (obviously load dependent) . It probably violates copyright if I post it here. Best you loan / borrow the book.

The other option... run a vintage amp with a subsonic filter. My Pioneer SX-1280 / SX-1050 each have a 15Hz and 30Hz respective highpass filters that have saved my sub on nasty bass music.

Cheers,
David.

[richie00boy]

Originally posted by HiFiNutNut
I wonder why in an active circuit it appears that we can add a filter either before the XO or after the XO. Why is it different from the passive circuit? I know for sure in a passive XO we can't.

Simply because with an active crossover there is no problem with having to consider the various impedances.

Originally posted by HiFiNutNut
You said "slightly". How slightly?

Please explain a bit more so that non EEs like us pure hobbists can learn a bit more.

Well as the high-pass sub-sonic filter will introduce a phase shift, it's best to hit the whole system with the phase shift so it's all aligned the same. If you just did it to the woofer then the woofer would be slightly phase shifted.

It really depends on how close the filter runs to the next driver up. e.g. On a sub crossed at 80Hz, if you just apply the sub-sonic filter to the sub, there might be a small amount of phase shift near the main speaker passband, so the sub would not be totally in phase with the main speakers.

That is often not really all that bad as it seems though, my sub for example runs with it's own sub-sonic filter and sounds great. There will be phase shift introduced by the crossover anyway, and the wavelengths are generally quite long.