Compression Driver Compensation

[Zilch]

How are you getting +12 db for Gregs Voltage Drive?? It's about -8db down. That means with 1 watt you have under 250mw at the driver.

You're measuring from the wrong place, Rob, by my view. Unity, or 0 dB is irrelevant to the issue under discussion, significant in an absolute sense only relative to the sensitivity of the particular woofer it's mated with, and how much "headroom" is available for accomplishing the passive compensation. Envision, for example, if you were working with a 103 dB woofer, set at 0 dB. The compensation for the HF would all be +dB.

I'm measuring the relative "boost" from the midband, where, presumably, there is none required. That's the baseline, for the discussion, not some arbitrary relationship to a random woofer, and I've chosen 1.8 kHz, where the Timbers filter begins the HF compensation, as the benchmark. Looking again, the Timbers "boost" is 13 dB to 20 kHz, whereas Guido's is 21 dB, and you earlier stated yours is even more aggressive than that. :dont-know

The major difference between what Timbers and Guido are doing occurs above 10 kHz, where Guido increases the rate of "boost" by three fold over what Timbers uses up there. I'm suggesting that there are likely sonic consequences inherent in doing that; I'd guess Timbers would desire to achieve flat response out to 20 kHz himself, if it were feasible within the performance limitations of the driver.

******

I'm having difficulty finding the response curve of the Timbers system as has been posted on several occasions by the member formerly know as Giskard here. Most notably, on at least one occasion, he stated that he was well satisfied with letting the driver roll off naturally at the high end, thereby avoiding certain (indeterminate) performance compromises. Does anyone have a link to that, or have it saved for reposting here? It'd be worthwhile to illustrate the differences in actual system frequency response which result from the two approaches under consideration here, as companion to the voltage drivers....

[Zilch]

I have a question.

What does all this have to do with a passive, full-range network?

You can't have "boost" of a particular frequency band within a passive network....excluding the use of an equalizer in a full-range, non-biamped system.

Yes, of course, and all of the participants understand the difference, but the net result is the same; the passive approach merely takes advantage of an available headroom differential between the woofer and HF driver to accomplish the task. In this case, the passive filter may be driven actively, as well, and several of us are using it that way. The active controllers do it the same way, when all is said and done, actually.

BUT, unless what is going on is viewed from the perspective of "Boost," it's easy to ignore the consequences, as is apparent in this discussion. Do we begin the midband attenuation at 40 kHz, 20 kHz, or 16 kHz? How rapidly do we have it occur in the last octave? No matter? I doubt it....

[Robh3606]

You're measuring from the wrong place, Rob, by my view. Unity, or 0 dB is irrelevant to the issue under discussion, significant in an absolute sense only relative to the sensitivity of the particular woofer it's mated with, and how much "headroom" is available for accomplishing the passive compensation. Envision, for example, if you were working with a 103 dB woofer, set at 0 dB. The compensation for the HF would all be +dB.

Well no. The amount of attenuation from unity or Zero Db line is what the networks is actually doing to the input signal. What the Zero Db line ends up being based on the woofer sensitivity is irrelevant to what the actual attenuation is. If the attenuation is 10db down from "0" you use 1/10 the applied power at that frequency. This is a passive crossover and in almost all cases the tailoring done is attenuation.

BUT, unless what is going on is viewed from the perspective of "Boost," it's easy to ignore the consequences, as is apparent in this discussion.

Nothing is being ignored. It's a choice that we are all free to make. If you want to run your drivers at -7db down as opposed to -4 go right ahead. If you are concerned that the break up modes could be more audible with the 3db change that's fine too. They may be. What ever is there is less that a 2425 judging from the distortion curves.

If you take Widget's position that you need a tweeter under any circumstance that's fine too.

The reason I posted my THD curves was because there was no indication it was rising or had a step function like the 2425 does in the 4430 crossover.

Rob

[edgewound]

Yes, of course, and all of the participants understand the difference, but the net result is the same; the passive approach merely takes advantage of an available headroom differential between the woofer and HF driver to accomplish the task. In this case, the passive filter may be driven actively, as well, and several of us are using it that way. The active controllers do it the same way, when all is said and done, actually.

..

I respectfully disagree with your point of view Zilch.

In a passive network fed from a full range signal, excluding an equalizer of course, you aren't adding gain to a particular filter section of the network. All you are doing...as Rob said...is taking energy away from the frequency range you are tailoring with a filter....there by giving a relative "boost" illusion which isn't a boost at all. It just seems to sound like one.

As far as woofer sensitivity is concerned, it makes all the difference in the world on compression driver distortion. The lower the sensitivity of the woofer the more you have to pad down the HF driver, which in turn makes gobs of headroom for that loafing driver that is hardly on due to it's gargantuan relative sensitivity/efficiency.

The controllers on the other hand do have some active tonal shaping going on which in turn does add some gain....just as a boost/cut equalizer does.

[Zilch]

Well no. The amount of attenuation from unity or Zero Db line is what the networks is actually doing to the input signal. What the Zero Db line ends up being based on the woofer sensitivity is irrelevant to what the actual attenuation is. If the attenuation is 10db down from "0" you use 1/10 the applied power at that frequency. This is a passive crossover and in almost all cases the tailoring done is attenuation.

Zero dB is established by the woofer sensitivity in Guido's curves. The HF compensation curve will simply move up or down depending upon what that sensitivity actually is, and how hard the HF driver is being driven depends upon the absolute input voltage. dB is a ratio, relative to a reference, in this case, Guido's woofer sensitivity.

In the Timbers curve, Zero dB is where LEAP (or whatever was used) started the highpass slope, somewhere out beyond 200 kHz, according to my sims. There is no woofer, as the filter was designed for active use. It may be used as a passive filter, but the attenuation would have to be independently established relative to the sensitivity of the particular woofer used. It's apples and oranges here, with respect to unity.

Fortunately, however, since dB IS a ratio, the curves may be compared, not with respect to their relative attenuation from unity, but rather, with respect to the differential amount of attenuation, in dB, employed in each case to accomplish the compensation. To make that determination, two reference points are required: a starting point and an ending point. Whether we choose those points to be 10 kHz and 20 kHz, or 1.8 kHz and 20 kHz, it's clear that Guido is employing MORE attenuation between those reference points than Timbers, i.e., from the reciprocal perspective, more "Boost," in order to achieve flat response out to 20 kHz.

Guido acknowledges that, but offers that he does not hear anything untoward in the response. You state that your approach is even more aggressive, and it sounds fine to you, as well. I am merely suggesting (and I trust, demonstrating,) that this comprises a departure from what might be considered prudent practice according to the evidence I find from those who know more about the limitations of these devices than we do.

Everyone is entitled, and welcome, presumably, to do whatever they like in this regard, but I am continuously reminded here of the importance of operating within the limits of the design intent and the performance envelope, as best we are able to ascertain these, as outsiders not privy to them in their entirety, from what information is available to us.

For me personally, of course, there is delicious irony in this, having been the pioneering practitioner of pushing this specific performance envelope. I have since backed off from that somewhat, and am explaining my rationale for doing so. If I find a particular driver plays to 20 kHz on a particular horn or waveguide with compensation conservatively applied, so much the better. If it rolls off above 16 or 18 kHz, however, thats no big whoop, either, necessarily; what most matters is finding and working within the practical limits....

[Zilch]

As far as woofer sensitivity is concerned, it makes all the difference in the world on compression driver distortion. The lower the sensitivity of the woofer the more you have to pad down the HF driver, which in turn makes gobs of headroom for that loafing driver that is hardly on due to it's gargantuan relative sensitivity/efficiency.

Well, sure, but then you crank them BOTH right back up to deliver the requisite SPL to satisfy the use. The HF driver gets the same power independent of the woofer sensitivity.

That's all somewhat irrelevant, however. No matter how we get there, the fundamental issue here is whether it's appropriate to be delivering proportionately more power to the driver in breakup mode for the purpose of extending the frequency response flat to 20 kHz.

******

Headroom is headroom, and ultimately, as you say, you can't make something out of nothing, either passively OR actively. When you're out of gain, you're out of gain, in both cases, and it's merely sematics in this context, whether you call it gain or, reciprocally, attenuation, boost or cut....

[Robh3606]

In the Timbers curve, Zero dB is where LEAP (or whatever was used) started the high pass slope, somewhere out beyond 200 kHz, according to my sims. There is no woofer, as the filter was designed for active use. It may be used as a passive filter, but the attenuation would have to be independently established relative to the sensitivity of the particular woofer used. It's apples and oranges here, with respect to unity.

That is incorrect. The 0db line is simply where the woofer curve normally lies for the simple fact that you don't attenuate woofers. They have the lowest sensitivity and with the exception of a Zobel only have a low-pass filter to roll the response off at the crossover point. Here's a case in point for the L250Ti Jubilee Crossover.

That is my LEAP curve plot of Greg's crossover.

Guido acknowledges that, but offers that he does not hear anything untoward in the response. You state that your approach is even more aggressive, and it sounds fine to you, as well. I am merely suggesting (and I trust, demonstrating,) that this comprises a departure from what might be considered prudent practice according to the evidence I find from those who know more about the limitations of these devices than we do.

Considering that in actual use they use +12db @16KHZ of active equalization when running the Vertec system there is no question in my mind that these are extremely robust drivers and no passive EQ is going to force them to misbehave at the power levels we use them at.

I also see no sign of increased distortion in my sine measurements.


Rob

[Robh3606]

Well, sure, but then you crank them BOTH right back up to deliver the requisite SPL to satisfy the use. The HF driver gets the same power independent of the woofer sensitivity.

No you don't the attenuation of a passive network doesn't just disappear. You are still putting much less power into the compression driver. If the attenuation is 13db on the compression driver to match the woofer sensitivity the power ratio is 20:1 in favor of the compression driver. If you dropped 200 watts of broadband noise into the system you would have only 10 watts into the compression driver. The rest of the power is dissipated in the network as heat.

Rob

[Zilch]

That is my LEAP curve plot of Greg's crossover.

I'm seeing a familiar 4-way there, Rob.

[Not that there's anything inherently wrong with that, of course.... ]

[Robh3606]

I'm seeing a familiar 4-way there, Rob.

Well let's see how long this 4-way takes to happen. I have had the drivers for over a year and, started building the crossovers, and my surround receiver just died on me last night

I am looking at options that have a real analog pass through for both CD and on the 5.1 input for SACD and DVD like on my older receiver . Really bad timing. Think I will start a new thread or revive an old one.

Rob

[Guido]

I planned to stay away from this discussion, but as I find my name here so often I should say something useful.

The reason for implementing a HF bump was simply a result of a 2 week listening session to the GT network. FOR ME (only my opinion) the speaker didn't sound fullrange. I didn't want to add a tweeter cause I believe in 2 way designs. So I did start modifications.

The sensitivity of my system is somewhere between 95dB and 96dB. So the 2435 is padded down very well. Even at the voltage drive peak it operates 4 dB lower than sensitivity.
You won't be able to squeeze out noticable distortion out of this speaker at medium to high home levels. The 2435 in this speaker sounds SO relaxed that I can't believe it.
I respect the listening impressions that you report and I respect that you hear "problems" of this driver above 10kHz or so. I just can't reproduce it.
Try different capacitors, try a nice amp or whatever. It will sound good....

[toddalin]

No you don't the attenuation of a passive network doesn't just disappear. You are still putting much less power into the compression driver. If the attenuation is 13db on the compression driver to match the woofer sensitivity the power ratio is 20:1 in favor of the compression driver. If you dropped 200 watts of broadband noise into the system you would have only 10 watts into the compression driver. The rest of the power is dissipated in the network as heat.

Rob

I have to agree with Zilch on this one. While the attenuation of the network doesn't disappear, one ends up turning up the volume control to compensate for the reduced compression driver output thereby pushing it just as hard.

If the compression driver puts out 100 dB at 1 watt at 1 meter, and you listen at 100 dB at 1 meter, you turn the amp up to where the compression drive is putting out 1 watt. If the network attenuates the driver by 10 dB, you turn the amp up to 10 watts to get the 1 watt. If the network attenuates the driver by 20 dB, you turn the amp up to 100 watts to get the 1 watt. But either way, you still end up with 1 watt at the driver.

[Robh3606]

I have to agree with Zilch on this one. While the attenuation of the network doesn't disappear, one ends up turning up the volume control to compensate for the reduced compression driver output thereby pushing it just as hard.

??????

The compression driver is reduced by an attenuation network to match the woofer. It doesn't matter how much you turn it up the volume the attenuation in the network remains the same. With a passive network you don't turn the compression driver up seperately. They use the same amp source. You can't turn one driver up without doing both.

Rob

[Zilch]

I have to agree with Zilch on this one. While the attenuation of the network doesn't disappear, one ends up turning up the volume control to compensate for the reduced compression driver output thereby pushing it just as hard.

Well, harder, actually, at 20 kHz, where we've bumped the drive up an addtional 6 - 8 dB over the Timbers filter at the same time.

I think Rob and Guido's view is, "No matter, these drivers can easily handle it."

I don't have any Aquaplas-damped drivers here, and there is some possibility the "damps spurious anomalies" thing is at play, but once those arrive, I'll certainly be breaking out the AM filters again and checking it out....

[Zilch]

The compression driver is reduced by an attenuation network to match the woofer. It doesn't matter how much you turn it up the volume the attenuation in the network remains the same. With a passive network you don't turn the compression driver up seperately. They use the same amp source. You can't turn one driver up without doing both.

Help me, now. Where's Guido getting the 25 dB of headroom shown in his voltage drives?

What IS the sensitivity of 2435HPL on these horns? 120 dB? :dont-know