John,
You make a very good point and one that argues strongly against the use of high output impedance amps and low order crossovers. I have never and would never use either of those, so I don't think that I'll worry too much about it.
By the way, you were right about the thermal modulation of a driver. I tried and tried to measure it but was never able to, so either my tests were not careful enough or the effect does not exist. I still wonder what is behind the subjective impression of "dynamics". It appears to be a valid impression, but one that seems to evade objective detection.
You make a very good point and one that argues strongly against the use of high output impedance amps and low order crossovers. I have never and would never use either of those, so I don't think that I'll worry too much about it.
By the way, you were right about the thermal modulation of a driver. I tried and tried to measure it but was never able to, so either my tests were not careful enough or the effect does not exist. I still wonder what is behind the subjective impression of "dynamics". It appears to be a valid impression, but one that seems to evade objective detection.
John K,
I have a question concerning an observation made during the 'old school' days of testing and designing crossovers years ago, before the days of LspCAD, SoundEasy and other computer software programs.
Years ago, I used a sine wave generator, a decent dual trace Hitachi oscilloscope, a good amplifier and the recommended resistive circuit to measure voltage response and impedance of the drivers. While testing tweeters, including ScanSpeak tweeters, an interesting observation was that while the image on the oscilloscope for the input across the amplifier was 20 mV and a clearly defined sine wave image, the measured response across the tweeter was 22 mV with a fuzzy sine wave image. The voltage actually increased on the output, across the tweeter terminals. As if some type of resonance was occurring. Installing a resistive L-pad across the tweeter input terminals reduce the voltage to 20mV or below and cleaned up the previous fuzzy lines of the sine wave image on the oscilloscope.
Is this related to your discussion? I have never seen this phenomena mentioned or discussed. As a result of this observation, I have always use an L-pad on tweeter circuits, with a resistor across the terminals of the tweeter in all my passive crossover designs. My intension of the parallel resistor was to act as a damping component for any back EMF voltage from the tweeter voice coil.
I have a question concerning an observation made during the 'old school' days of testing and designing crossovers years ago, before the days of LspCAD, SoundEasy and other computer software programs.
Years ago, I used a sine wave generator, a decent dual trace Hitachi oscilloscope, a good amplifier and the recommended resistive circuit to measure voltage response and impedance of the drivers. While testing tweeters, including ScanSpeak tweeters, an interesting observation was that while the image on the oscilloscope for the input across the amplifier was 20 mV and a clearly defined sine wave image, the measured response across the tweeter was 22 mV with a fuzzy sine wave image. The voltage actually increased on the output, across the tweeter terminals. As if some type of resonance was occurring. Installing a resistive L-pad across the tweeter input terminals reduce the voltage to 20mV or below and cleaned up the previous fuzzy lines of the sine wave image on the oscilloscope.
Is this related to your discussion? I have never seen this phenomena mentioned or discussed. As a result of this observation, I have always use an L-pad on tweeter circuits, with a resistor across the terminals of the tweeter in all my passive crossover designs. My intension of the parallel resistor was to act as a damping component for any back EMF voltage from the tweeter voice coil.
Last edited:
I agree, as long as the filter functions are the same. The problem is, they most often are not. And DSP is the big game changer there. It is much easier and faster for most people to build a good crossover with DSP than it is with passive parts. So the active - especially DSP - crossovers tend to be better optimized.
IME, when someone says that active sounds better than passive, they are not comparing the same crossovers. The active will be much more dialed in, simply because it's so much faster and easier to do so. If one compares the same filter functions between active and passive, the difference isn't enough to worry about.
I agree. A highly refined, high quality passive is perfectly capable, but for experimenters/diyers, active is worth the investment to avoid stocking a wide variety of components of good quality and taking the time to do good passive XOs.
Sometimes, but finding the targets can be more involved than implementing them so I can't agree with the logic of this.
A 20 to 22 mV peak-to-peak input to the tweeter terminals is well within the acceptable operating range. That's 0.020 to 0.022 Volts peak-to-peak.
A don't have an answer for why you would see a different voltage. Perhaps their was some type interaction of the tweeter VC inductance with some capacitive element causing a resonance peak. The Lpad might have damped the resonance but this it totally speculation.
As for thermal compression, yes, I fond it not to be a dynamic effect. Rather it appears to be a quasi steady state phenomena where the VC temp heats up/cools down slowly due to the time averaged application of and dissipation of power. The thermal time scales seem to be much longer than those associated with audio frequencies and dynamic transients.
But, this is another region where active crossovers are superior. With a passive crossover VC heating will alter the driver's Z and, thus, alter the transfer function (TF) of the crossover filter. This doesn't happen with an active crossover. Of course, while the active crossover is immune to changes in TF due to VC heating, they are still audible effects due to the change in VC impedance. But then there are also effects due to changes in VC inductance with excursion which can alter a passive crossover's TF depending on how loud the speaker is played. It's amazing dynamic speakers work as well as they do.
Unfortunately, i think you didnt try other ways to solve the issue. I see that a notch filter on tweeter resonance, even if the frequency seems to be outside of the pass band, always have positive effect. No side effect, as long as the position and shape of the notch is very precise. Problem is, only slight miss will bring damaging effect to sound.
OTOH, lpad will always have negative effect. I found that in all my cases, the problem solved by lpad can be solved without lpad. And i found out that the primary audible strength of active crossover is similar to what has been mentioned by Linkwitz: the direct and stronger control of the voice coil by the amplifier. The lpad is just another form of this weakness of passive crossover.
Sure...but it is a very confusing and tiring process. Get a fixed passive crossover and enjoy music! 😀
This slight increase in voltage at the tweeter terminal occurred throughout the pass band, from about 2kHz to 20kHz.
As far has applying proper notch filters, the current crossover design and testing software packages, such as LspCAD, make that achievable.
It can be, yes. With the hardware and software available to today's DIY'er, building good crossovers is easier and easier. DSP still makes it very fast to experiment, find good filters, test, measure and listen to them. Doing that with passive crossovers is more time consuming.
Not comparing like with like? So what? Why would I deliberately hobble the final performance of a DSP design by limiting the xover implementation to the equivalent passive?
DSP is faster to get a basic set up dialled in, but can take quite some time and many measurements to extract the best out of the design.
It is almost impossible to make it apple-to-apple. But it doesn't have to be apple-to-apple actually. Active can be preferred because of its perceived dynamics due to having 'easier' load (voice coil) for the amplifier. With passive, amplifier must be of much higher quality.
But it is never apple-to-apple. It is simply about what we like, what we don't like, what we can perceive, what we cannot perceive... (If we were deaf, everything sounds the same, all are apples).
Things that dsp can not do vs passive xo and vise versa
I'm a newbie regarding xo design, but i want to jump into this topic to learn on how designing my own crossover. i only understand basic knowledge about RC, RL, RLC, zobel, bsc and etc.
looking at the value choices on passive which will require massive inventory, i think playing with dsp product like minidsp 2x8 will only need 1 time investment.
Considering I already have several amplifiers, can anyone teach me on what kind of things that dsp can not do compared to passive xo? my short knowledge says : can i flatten impedance load on dsp? or do you need to worry about load impedance where you use dsp? and other hints for newbie.
thanks
I'm a newbie regarding xo design, but i want to jump into this topic to learn on how designing my own crossover. i only understand basic knowledge about RC, RL, RLC, zobel, bsc and etc.
looking at the value choices on passive which will require massive inventory, i think playing with dsp product like minidsp 2x8 will only need 1 time investment.
Considering I already have several amplifiers, can anyone teach me on what kind of things that dsp can not do compared to passive xo? my short knowledge says : can i flatten impedance load on dsp? or do you need to worry about load impedance where you use dsp? and other hints for newbie.
thanks
The impedance of the speaker is a matter for the power amplifier because it acts as a buffer to the pre stages such as the processor.
I built a DIY three-way active system based on an active analogue X-over and found the active X-over far more versatile than an equivalent passive one. Never used a miniDSP but have read up on them and they are clearly so much more versatile than my old analogue X-over and a zillion times more versatile than a passive X-over.
Suggest you read up on the miniDSP rather than me list a whole string of things the miniDSP can do that a passive crossover can't. As far as I'm aware, there's nothing a passive crossover can do that a miniDSP can't do better/more accurately:
DSP basics
The miniDSP outputs feed into the power amps so, as AllenB said, drive unit load impedance isn't an issue for the miniDSP. Any half-decent power amp will cope with variation of drive unit impedance.
Well, there is that whole trick of powering a multi-way speaker with a single amp. But gadut did indicate having extra amps.
Much changed since post #1 in 2011. Esp...
1. room acoustics including recognition of importance and ability to use REW and cheap mics to assess performance at your chair
2. price of gear moving amps and DSP relatively cheaper
After a few years with the excellent Behringer DCX2496, totally inconceivable to me how anybody could claim to be aiming for good DIY sound without powers, ease, and flexibility you get with DSP; it just makes no sense.
Manufacturers who must sell simple intact boxes rather than systems have no choice. But some manufacturers provide alternate bi-amp connections today. But then, you need prayer and luck when you buy (or build) an intact system and hope it plays well in your room.... or at least tests well anywhere besides one-meter when you are done.
B.
1. room acoustics including recognition of importance and ability to use REW and cheap mics to assess performance at your chair
2. price of gear moving amps and DSP relatively cheaper
After a few years with the excellent Behringer DCX2496, totally inconceivable to me how anybody could claim to be aiming for good DIY sound without powers, ease, and flexibility you get with DSP; it just makes no sense.
Manufacturers who must sell simple intact boxes rather than systems have no choice. But some manufacturers provide alternate bi-amp connections today. But then, you need prayer and luck when you buy (or build) an intact system and hope it plays well in your room.... or at least tests well anywhere besides one-meter when you are done.
B.
Last edited:
I have always wanted to export filter response from a simulator and there is some software doing some number crunching that turns it into biquads that can be fed into DSP.
Then viola, the passive design can be auditioned as easily as "print preview" before print.
Does that number crunching software exist?
Then viola, the passive design can be auditioned as easily as "print preview" before print.
Does that number crunching software exist?
- Status
- Not open for further replies.