mbutzkies said:
They are probably from the room it was measured in, I don't remember the exact conditions for that particular test (it was years ago)
You can't get past a spider resonance on high excursion drivers, they will all have it.
That impedance peak is low, try a Kapton former high Qms driver to really see a peak. No impedance notches are needed in the bass region for most any decent amplifier.
PB2 said:
Pete you sound like someone who knows some things about bass reproduction. Volume displacement (driver diameter times xmax) is key for low bass output, some here don't agree.
The key difference in the IB copper motor is that it can produce the same HF output no matter the cone position which is determined by the lowest bass signal input. You will not hear the intermodulation distortion of cheaper drivers. A good test is mixing a 20Hz signal with a 1Khz signal and listening to this mix vs a clean 1Khz signal with a raw bass driver. Its really good if you can vary the 20Hz level while keeping the 1Khz the same. Lets just say that many drivers don't do this very well.............
Graham Maynard said:
Hi Graham,
Nick and Pete won't appear to make any sense as there are some very important assumptions you have made which are common misconceptions in the audio world. First and foremost is the operation of an acoustically small loudspeaker as it relates to excursion, phase, transient response, and amplitude response.
As Nick stated in noting that loudspeakers are constant acceleration devices, a loudspeaker's cone motion & position *DOES NOT* track the Voltage waveform we visualize as it would be seen on an oscilloscope. This relationship is why a woofer's cone has to move 4x as far to produce the same SPL each time the frequency drops by 1/2. If driver position followed Voltage directly, there would be the same excursion required for 20kHz as 20Hz.
Directly related to this fact is the reality that an acoustically small speaker (most lower frequency speakers), will never have both flat amplitude response and ideal transient or phase response without external processing (DSP, etc). A sealed or IB driver will have a broad band 90 deg phase lag in its flat operating range combined with the effects of either end of the passband following the phase response that corresponds to the high and low frequency roll off of the driver.
As it turns out, a driver with a downward sloping frequency response with lower frequency will appear to have a better phase and transient response. Of course without some fancy DSP, as soon as you correct the magnitude response, the phase response goes back to what we would expect.
Personally I have yet to see any valid argument that a high Q driver is not the preferred solution for an OB system, unless you are planning on the passive crossover having a similar effect of raising the Q of the system at low frequencies. Even in this case, this is just a different means to the same end, and all acoustic behavior will be the same.
Hi PB2,
We cannot convince each other - each must work out for their own.
The best way is by using a microphine to observe the output versus the electrical signal waveform, and to not measure/study a Qts modified steady sine response only.
There is a need to understand how Q affects reproduction, as with one note bass, tonality, phase response etc.
It is !
Yes Qts is important, being a related to both Qms and Qes charateristics. What happens due to mechanical Q must be controlled by the voice coil, and with the likes of the Alpha this control is very weak, giving rise to such a dynamically induced amplitude peak that it can 'flatten' the amplitude response at frequencies around resonance without need for much EQ.
The IB15 resonance is so low (too low for OB ?) that its resonance falls below the AF spectrum, so that the phase response within the audio spectrum is more related to the baffle, EQ and room. With the likes of the high Qts Alpha its phase response is voiced around its own resonance and thus it imparts a tonality upon all LF reproduction which lower Qts and lower Fs drivers will not.
Hi PB2,
We appear to be coming to the same point from different directions. You from the amplitude response, me from the dynamic However merely looking at sine amplitude measured responses in isolation will not give an idea of dynamic behaviour capabilities around resonance (Qts modified reproduction), even with damping rings, phase plug etc.
Separately I would expect that the low resonance IB15 on an OB could easily over-excurse from the LF voltage outputs of some less than perfect CD productions due to automatic ADC/DAC compander circuitry. This would at least be automatically reduced by the suspension of a driver having resonance say 20 to 25Hz.
Hi Nick,
Of course I do agree with you that the phase etc. above resonance is more accurate if the Fs is lowered. I was dealing with reproduction responses at the resonance itself, where the falling baffle response can mask a peak which can appear 'good' on the amplitude response because it has developed due to driver Q, but result in a 'euphonic' tonality which might not be uncomfortable or tiring, but which is not actually good reproduction either.
Hi Mark,
Please do not assume that I have made assumptions !
You are going on to things I have not mentioned, which does not mean that I am not aware of them. In return I am not assuming anything about what you think.
>> You wrote;-
As it turns out, a driver with a downward sloping frequency response with lower frequency will appear to have a better phase and transient response. Of course without some fancy DSP, as soon as you correct the magnitude response, the phase response goes back to what we would expect.
Personally I have yet to see any valid argument that a high Q driver is not the preferred solution for an OB system, unless you are planning on the passive crossover having a similar effect of raising the Q of the system at low frequencies. Even in this case, this is just a different means to the same end, and all acoustic behavior will be the same. <<
Mostly I agree, and of course it is cone displacement/microphone measurement which must be considered, however EQ cannot correct a response if the driver does not have a low Qes.
Where the Qes is high then the voice coil's transduction capabilities can be inadequate to control mechanical cone motion, this being where resonantly induced tonality arises. EQ cannot prevent resonant tonalities arising, especially at Fs, and these remain audible even when less powerfully energised.
By the same token, EQ cannot pre compensate the drive during only the first 90 degrees of a LF waveform to counter any Qes related loss of impulse transduction.
There are many factors relating to reproduction quality, dynamics and tonality, thus looking at sine measured/simulated frequency responses in time isolation fails to provide adequate information.
Our Qes figure can be entered into simulation programs in order to generate *sine* amplitude resonses, and Qes always tends to be mentioned in relation both steady sine amplitudes and optimally 'flat' responses, but we must not lose sight of dynamic response capabilities when it comes to reproducing music, and must not assume that EQ can compensate for driver Qes related inadequacies.
Cheers ......... Graham.
We cannot convince each other - each must work out for their own.
The best way is by using a microphine to observe the output versus the electrical signal waveform, and to not measure/study a Qts modified steady sine response only.
There is a need to understand how Q affects reproduction, as with one note bass, tonality, phase response etc.
tinitus said:
It is !
Yes Qts is important, being a related to both Qms and Qes charateristics. What happens due to mechanical Q must be controlled by the voice coil, and with the likes of the Alpha this control is very weak, giving rise to such a dynamically induced amplitude peak that it can 'flatten' the amplitude response at frequencies around resonance without need for much EQ.
The IB15 resonance is so low (too low for OB ?) that its resonance falls below the AF spectrum, so that the phase response within the audio spectrum is more related to the baffle, EQ and room. With the likes of the high Qts Alpha its phase response is voiced around its own resonance and thus it imparts a tonality upon all LF reproduction which lower Qts and lower Fs drivers will not.
Hi PB2,
We appear to be coming to the same point from different directions. You from the amplitude response, me from the dynamic However merely looking at sine amplitude measured responses in isolation will not give an idea of dynamic behaviour capabilities around resonance (Qts modified reproduction), even with damping rings, phase plug etc.
Separately I would expect that the low resonance IB15 on an OB could easily over-excurse from the LF voltage outputs of some less than perfect CD productions due to automatic ADC/DAC compander circuitry. This would at least be automatically reduced by the suspension of a driver having resonance say 20 to 25Hz.
Hi Nick,
Of course I do agree with you that the phase etc. above resonance is more accurate if the Fs is lowered. I was dealing with reproduction responses at the resonance itself, where the falling baffle response can mask a peak which can appear 'good' on the amplitude response because it has developed due to driver Q, but result in a 'euphonic' tonality which might not be uncomfortable or tiring, but which is not actually good reproduction either.
Hi Mark,
Please do not assume that I have made assumptions !
You are going on to things I have not mentioned, which does not mean that I am not aware of them. In return I am not assuming anything about what you think.
>> You wrote;-
As it turns out, a driver with a downward sloping frequency response with lower frequency will appear to have a better phase and transient response. Of course without some fancy DSP, as soon as you correct the magnitude response, the phase response goes back to what we would expect.
Personally I have yet to see any valid argument that a high Q driver is not the preferred solution for an OB system, unless you are planning on the passive crossover having a similar effect of raising the Q of the system at low frequencies. Even in this case, this is just a different means to the same end, and all acoustic behavior will be the same. <<
Mostly I agree, and of course it is cone displacement/microphone measurement which must be considered, however EQ cannot correct a response if the driver does not have a low Qes.
Where the Qes is high then the voice coil's transduction capabilities can be inadequate to control mechanical cone motion, this being where resonantly induced tonality arises. EQ cannot prevent resonant tonalities arising, especially at Fs, and these remain audible even when less powerfully energised.
By the same token, EQ cannot pre compensate the drive during only the first 90 degrees of a LF waveform to counter any Qes related loss of impulse transduction.
There are many factors relating to reproduction quality, dynamics and tonality, thus looking at sine measured/simulated frequency responses in time isolation fails to provide adequate information.
Our Qes figure can be entered into simulation programs in order to generate *sine* amplitude resonses, and Qes always tends to be mentioned in relation both steady sine amplitudes and optimally 'flat' responses, but we must not lose sight of dynamic response capabilities when it comes to reproducing music, and must not assume that EQ can compensate for driver Qes related inadequacies.
Cheers ......... Graham.
Hi Michael,
Enough to reduce the Fs to say 2/3 of its original value might give some idea of differences ?
I very much look forwards to reading your results.
Would you be able to determine any T/S VAS and Q figures to examine the driver responses within XlBaffle too.
(To see how the dynamic response balances out against the Qes/VAS augmented amplitude response.)
I knew I had seen this somewhere on a .co.uk website last year, so it is available in the UK, though not cheap.
http://www.volumeonline.co.uk/product_info.php?cPath=23_43_83&products_id=218
This leads me towards thinking about using two different 15"ers on the same baffle, this one close to the floor and running lower only.
Cheers ....... Graham.
Enough to reduce the Fs to say 2/3 of its original value might give some idea of differences ?
I very much look forwards to reading your results.
Would you be able to determine any T/S VAS and Q figures to examine the driver responses within XlBaffle too.
(To see how the dynamic response balances out against the Qes/VAS augmented amplitude response.)
I knew I had seen this somewhere on a .co.uk website last year, so it is available in the UK, though not cheap.
http://www.volumeonline.co.uk/product_info.php?cPath=23_43_83&products_id=218
This leads me towards thinking about using two different 15"ers on the same baffle, this one close to the floor and running lower only.
Cheers ....... Graham.
Interresting read
John Kreskovsky has a interresting discussion on Q here: http://www.geocities.com/kreskovs/Box-Q.html and here: http://www.geocities.com/kreskovs/Box-Qa.html
(The paper is on sealed system bass but should be applicable to other resonant systems as well)
John Kreskovsky has a interresting discussion on Q here: http://www.geocities.com/kreskovs/Box-Q.html and here: http://www.geocities.com/kreskovs/Box-Qa.html
(The paper is on sealed system bass but should be applicable to other resonant systems as well)
Hi Nick,
I feel that the difference between these two arises due to the dynamic loss, storage and re-release of resonant energy lies within the 'listening' bandwidth with the Alpha; whereas the IB15 resonance is below the normal AF bandwidth and thus the within-band dynamics and coherence are minimally disturbed.
The Qts of the Alpha-15A is deliberately being used to flatten a naturally sloping response, but there is no free lunch !
I am already aware of those other aspects you mention, and hence a need to be careful before allowing too high a Qes to develop, especially if Fs arises within the listening bandwidth.
Thanks for the link Ion.
There is another relevent section within Seigfreid Linkwitz's writing, but I cannot find it at the moment. It relates to using EQ to compensate for driver characteristics, but it also illustrates how, when set to equalise the steady sine response, the driver amplitude error arising during the first cycle (dynamic response) is not corrected.
Cheers ....... Graham.
nickmckinney said:
I feel that the difference between these two arises due to the dynamic loss, storage and re-release of resonant energy lies within the 'listening' bandwidth with the Alpha; whereas the IB15 resonance is below the normal AF bandwidth and thus the within-band dynamics and coherence are minimally disturbed.
The Qts of the Alpha-15A is deliberately being used to flatten a naturally sloping response, but there is no free lunch !
I am already aware of those other aspects you mention, and hence a need to be careful before allowing too high a Qes to develop, especially if Fs arises within the listening bandwidth.
Thanks for the link Ion.
There is another relevent section within Seigfreid Linkwitz's writing, but I cannot find it at the moment. It relates to using EQ to compensate for driver characteristics, but it also illustrates how, when set to equalise the steady sine response, the driver amplitude error arising during the first cycle (dynamic response) is not corrected.
Cheers ....... Graham.
Graham Maynard said:
An important thing to consider is that mass does not only come from the cone, so comparing Mms alone is not that easy. Say the Mms is 75grams in a driver. There can be a big difference between a 25gram cone with 50gram coil and 50gram cone with 25 gram coil. This is because of the effect the coil has in modulating the flux field as it moves through the gap. The more turns of wire in the gap, the greater the effect of the modulation of the field. In a driver without a full copper sleeve on the pole, or other proper shorting ring to reduce this modulation, you could expect a difference between the two different VC drivers. The difference in this case comes down to not only the difference in inductance, but the effects the coil creates on the flux in the gap. High inductance does not only affect things at higher frequencies.
mbutzkies said:
After spending nearly 3 months in trying to get consistent repeatable impedance curves, I can probably give a valid guess on this. We often see these blips in drivers close in this range due to motion of the driver while running the curve. We built a 300lb steel structure to securely clamp drivers down in the test boxes. One is fully open baffle and the other is sealed to do a delta compliance measurement on each driver. We still have some work to do to further secure drivers when testing. Yes, even at 1w inputs, the drivers create a lot of force. Often times now, one of us has to physically climb up and sit on top of the test enclosure to get a clean impedance curve without these two peaks. The steel tube structure can resonate as well as the baltic birch enclosures, which believe it or not will show up in the impedance peak. We'll be filling the whole steel structure with 5lb structural foam and adding another section on top of the structure to physically clamp down on the enclosure as well.
John
panomaniac said:
A few things you need to keep in mind with this. It's along the lines of measuring T/S parameters and the reason that delta/mass method is not good. One, you cannot have the driver facing upwards to do this. The suspension will already sag, moving it to a point of slightly lower Bl and stiffer compliance. This will affect any tests. So, you do need to have it mounted horizontally. The big thing is applying the mass without any other effects. As Nick stated, gluing to the VC form is probably the best place. That means removing a dustcap. If you add mass anywhere on the cone, dustcap, etc you add other issues. For example if you added mass along the outside edge of the cone, you would be increasing the possibility of cone flex. The same goes for adding it to the front of the dustcap as the dustcap is not intended to support a high amount of mass.
The other thing is simply that the mass cannot move. I've seen people try to use modeling clay, hot glue on nickels, things like that, and none of them will get the rigid bond you need.
John
Beau said:
Hello Beau,
The IB15 was designed really to be a budget driver for infinite baffle alignments. If you look at the price per liter of displacement, there are very few drivers comparable. Then when you figure in the full copper shorting ring, the extended bandwidth, the low distortion, etc, there are really no drivers comparable at all.
That said, a driver with similar sized motor with a phase plug could be done, however it wouldn't be able to be at the same price. We're really approaching the cost of one of the TD woofers at this point though. The motor on the IB15 is a little smaller so saves some cost, but not a whole lot. I'd say that the 15" with a phase plug on the smaller IB motor would come in at right around $200 vs the $279 for a standard TD driver of the Dipole15.
John
I have to agree with MJK above with respect to acoustical music ! Regarding my 'Volks-OB' see:
http://www.audiocircle.com/circles/index.php?topic=46951.0 I have yet to see something more than 1-2 mm cone motion how heavy bass I play. For instance The Blind Boys of Alabama's ATOM BOMB 1st and 3rd track played at over 95 dB just move the cone about 1-2 mm with its heavy bass. Likewise Manger's test record 'Musik von einem anderen Stern' the last Percussion Group at similar SPLs does not indicate any heavier cone motions. Conemotions certainly not over 2 mm, hardly visible.
/Erling
http://www.audiocircle.com/circles/index.php?topic=46951.0 I have yet to see something more than 1-2 mm cone motion how heavy bass I play. For instance The Blind Boys of Alabama's ATOM BOMB 1st and 3rd track played at over 95 dB just move the cone about 1-2 mm with its heavy bass. Likewise Manger's test record 'Musik von einem anderen Stern' the last Percussion Group at similar SPLs does not indicate any heavier cone motions. Conemotions certainly not over 2 mm, hardly visible.
/Erling
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.