audio-talk

Useful input chaps, thanks.

The DIY Audio Pass-design crossover is something I've been following and is top of my list for the 'end-result crossover' because of it's versatility and quality. First though I need to determine the target configuration and I think getting hold of a DSP unit is the way to do that. If possible I want to run the full range unit 'sans crossover' by bandwidth limiting the MoFo amp I'm planning to use with it so the next step is to build a suitably configured MoFo - the parts shipped yesterday but probably won't arrive until next week now.

First though I need to determine the target configuration and I think getting hold of a DSP unit is the way to do that. If possible I want to run the full range unit 'sans crossover' by bandwidth limiting the MoFo amp I'm planning to use with it so the next step is to build a suitably configured MoFo - the parts shipped yesterday but probably won't arrive until next week now.

Ray,
analog filters have phase shift. If you plan to use DSP you should be aware of the following:
https://www.minidsp.com/applications/ds ... -filtering

You can avoid problems if you use the Behringer Ultradrive for instance because you can set filters and phase independently.

Although this is about adding a sub it is helpful information regarding phase and how to adjust:
http://www.rythmikaudio.com/phase2.html

In order to be absolute sure I would always measure the results which will show you what's going on without racking your brain.

Thanks Wolfgang. I don't have any DSP kit at all, nor do I know that much about this stuff so it's all a learning opportunity.

In order to be absolute sure I would always measure the results which will show you what's going on without racking your brain.

Though measuring phase without the right kit is very difficult.

You can avoid problems if you use the Behringer Ultradrive for instance because you can set filters and phase independently.

I don't think the Ultradrive offers FIR filters so you still have the same frequency dependent phase issues (if they are an issue for you).

If you want to control phase with FIR’s in the sub 100Hz you will need a serious DSP with 20k taps or more or do it via a PC using software. MiniDSP only has 2000 taps per channel for their ASharc. I would question if you would hear any differences with phase below the Schroeder frequency of you listening room which will be above 100hz unless you have a very large room.

Though measuring phase without the right kit is very difficult.

Carma 4 for example shows magnitude and phase together. But even just by looking at the magnitude response one would see if there is a dip around the x-over point which would indicate a problem.

I don't think the Ultradrive offers FIR filters so you still have the same frequency dependent phase issues (if they are an issue for you).

According to Behringer the filters are linear. The phase can be adjusted in 5 degr. increments.

MiniDSP only has 2000 taps per channel for their ASharc. I would question if you would hear any differences with phase below the Schroeder frequency of you listening room which will be above 100hz unless you have a very large room

The mini SHARC can switch from IIR to FIR (unfortunately not back again). I tried it many times when I was still using it in my system and could hardly hear any difference at any freqeuncy.

I forgot to mention that I used the miniSHARC PEQs only for adjusting the frequency response of the room. I never tried the x-over option which would be a different thing regarding phase shift and filter types.

Wolfgang wrote: ↑Thu Dec 31, 2020 6:12 pm
According to Behringer the filters are linear. The phase can be adjusted in 5 degr. increments.

AFAIK, no Behringer equipment has FIR capabilities. I own the Ultradrive (DCX2496) and the Ultracurve (DEQ2496), and neither have any FIR stuff. Plenty of standard (IIR) EQ options in those, though.

Chris

It looks like you are right! I did some more research and found this:
“The phase control setting is a bit confusing and the operation is not explained well in the owners manual. The DCX uses the LP filter frequency setting (even if it's "Off") as the reference point for the phase adjustment. In the example I set the frequency to 499Hz and adjusted the phase control to 180 degrees which yielded the all-pass function you see with the 180 degree "turn"over. (…)The DCX phase response of an LR24 crossover looks exactly like an analog version would. The response across the entire band would vary 360 degrees and at 2khz the response would be 180 degrees out-of-phase with the input.”

I was hoping that the filter could be a made "linear" by adjusting the phase. That obviously also doesn’t work. Somebody had already the same idea:

“I wonder if it would be possible to use the phase adjustment to compensate for the filters phase? Since you can set the "turn point" of the phase adjustment to match the x-over point it should be possible to exactly cancel the phase shift of the filter? At least in theory for 2nd order filters? "

Answer (most likely correct):
“The 2nd-order all-pass function of the phase control is equivalent to a 4th-order LR phase shift......meaning that it's shifting the same direction. So, if you were to line up the turnover point right on top of the x-over point you'd only succeed in INCREASING the phase-shift to an equivalent 8th-order.”
Bummer.

Having used a DCX2496 in anger for a number of years, I'd be really surprised if anything FIR could be done. Thanks for confirming.

FWIW, a typical laptop running Equaliser APO has the option for a LOT of FIR taps. A basic 3rd gen i5 laptop (which I use as a HTPC) maxed out at 100,000. Couldn't hack 1,000,000. I suspect a higher-end PC would manage more, but 100,000 taps is plenty for most applications.

Chris

Back to my project...

I have in hand a build of a bandwidth limited MoFo amplifier to test with the full-range drive units sans active or conventional crossover - maybe get the test amp built over the weekend. Not sure if I'll be able to get a piece of chipboard or plywood for a test baffle though given the more restictive Covid restrictions that we'll be facing - we'll see.

Let's assume that the bandwidth limited MoFo driving a full-range drive unit on the baffle gives a good result - time will tell.

Is it a reasonable expectation to achieve a complete crossover by doing the reverse, ie, lo-pass bandwidth limiting, with the bass amplifier? In my simple world, conceptually this is just moving the passive crossover elements 'upstream'.

Here is the schematic of the chip amp I plan to use for the bass driver unit;
as I understand it, R4 and C10 are a low pass filter (currently at several 100KHz IIRC) so I'm wondering if I could reconfigure it to give a lo-pass filter with a corner point at 180Hz. I appreciate some measurement/tweaking/listening might be required to get it right, but ignoring that aspect for now, the result would be a crossover constituted of two first-order filters at about 180Hz - are there implications to that approach that my limited understanding is missing - I'm aware I may well have too simplistic a view.

BTW, I appreciate there will need to be a volume control in front of the chip amp to be able to set the relative levels of the full-range/bass sections - the chip amp will have more gain than the full-range amplification arrangement.

Yep, that would be a simple place to do it.

I appreciate some measurement/tweaking/listening might be required to get it right, but ignoring that aspect for now, the result would be a crossover constituted of two first-order filters at about 180Hz

Ray, FWIW, and to some this may be considered not much, but I have yet to hear a speaker using conventional coned drivers that can do proper bass and have full control and grip of bass notes with a simple 1st order filter.

As I say just my opinion from all the multi-way speakers I have heard and built that use 1st order filters.

But he is using two of them with the same corner frequency, so will get the roll off of a second order filter.