You will find that most speaker designers today do not give much emphasis of SYMMETRICAL SLOPES. This at first appears to be the most elegant solution, but it almost always destroys the transient behavior of the speaker system.
The TP crossover is almost faultless, in itself, and if you made a Spice emulation of it you might be VERY surprised how faultless it can be.
It certainly taught me a thing or two, about what phase shift really is, as well as group delay.
The TP crossover is almost faultless, in itself, and if you made a Spice emulation of it you might be VERY surprised how faultless it can be.
It certainly taught me a thing or two, about what phase shift really is, as well as group delay.
This is one of these rare days when I can agree with everybody; gave me a new insight in the matter too.
The problem with substractive filters as you correctly imply is to achieve good enough acoustic addition.
With two drivers mounted with different acoustic centres, this cannot be achieved. The reason being that this misalignment introduces a vector in the acoustic addition, which means that the addition will only be 100% on the mark at one specific measuring angle.
Now, with coaxial drivers, the acoustic centres of the low and high drivers are aligned. In other words, a substractive filter might be a nifty way of creating a linear phase setup for a coax with minimal components.
Next on the list. It might not be easy to find a driver that can do this since the high driver needs to work harder and lower to turn back the phase shift of the bass driver, but who knows.
No J6 and J8 act as folded cascodes to one side of the input stage and the other side drives the gates of J6 and J8 but at a fairly low gain. The result is the differential current in the input quad comes out of the drains of J5 and J7 and essentially all the gain is taken at that node. The usual use of this connection has substantial voltage gain from J1 and J3 and that side does most of the "work".
Just play with it in LTSPICE, it does work as drawn.
Last edited:
This type of xovers are not good at all. It really doesn't matter the order of the proper filter block, the other "half" always has a -6db/oct slope, which is almost useless as an audio's filter. Assuming you use a low-pass filter as the main filter block, if you move the xover's cut-off frequency higher enough to avoid damaging the higher frequency speaker (such as 4kHz) the lower frequency driver becomes highly directional. Yet we haven't tried to correct the phase shift, let alone the peak at the xover frequency.
Perhaps, it probably will be marginally useful if you get the "right" speakers, but just by spending a little money on one opamp and a pair of resistors and capacitors, you can get a fourth order LR xover... and can reuse the time aligment circuit....
Likewise, unless the drivers have infinite bandwidth, the acoustic crossover will not be flat and hence not "transient perfect."
EVERY CROSSOVER HAS PROBLEMS AND TRADEOFFS.
Please don't be so rigid EZA... Many speakers use 6 dB/octave networks successfully, and they are Transient Perfect also.
With every crossover type, you have to note the 'goods' and the 'bads' and trust me, higher order xovers have phase shift and delay that can create all kinds of problems. Of course, you have to trust the human ear to hear the problems easily measured. If you don't believe that some things can be heard, then you can get away with all kinds of tradeoffs, that optimize the frequency response at the expense of transient response.
Please don't be so rigid EZA... Many speakers use 6 dB/octave networks successfully, and they are Transient Perfect also.
With every crossover type, you have to note the 'goods' and the 'bads' and trust me, higher order xovers have phase shift and delay that can create all kinds of problems. Of course, you have to trust the human ear to hear the problems easily measured. If you don't believe that some things can be heard, then you can get away with all kinds of tradeoffs, that optimize the frequency response at the expense of transient response.
that can only be by the sales numbers
but I guess if a speaker manufacturer have managed to impress the buyers once he will possibly get away with almost anything, like 1. order crossovers in speakers
but maybe there are several 'variations' of 1. order filter
and are they actually 1. order when measured ?
electrically 1. order with zobel can be one of the best
probably because its anything but 1.order, acousticly
actually, it would be possible to do acoustic 2. order with filter on tweeter only
though I'm not suggesting it
just because acoustic rolloff is what matters most, it's still not like the electric side(phase ?) won't matter at all
probably because its anything but 1.order, acousticly
actually, it would be possible to do acoustic 2. order with filter on tweeter only
though I'm not suggesting it
just because acoustic rolloff is what matters most, it's still not like the electric side(phase ?) won't matter at all
For their size I have never seen an equal. It's disappointing that Dick has no suggestions for any improvements at all. Just leave well enough alone. At $300 a pair on the used market a bargain.
John where else would you go and hold court with folks that have actually known you for the last 40 years.
Last edited:
- Status
- Not open for further replies.