It would seem that the majority of speaker talking heads agree that a "high" damping factor, maybe 10, is needed throughout the operating frequency range of a driver. Usually that includes full frequency response not in the region of resonance magnification of the driver. Thus, for that majority, passive crossover networks (between the power amplifier and the driver), are seen as problematic. The problem is that near the crossover frequency, the damping factor is reduced by the impedance of low-pass and high-pass filters. For example, see
Active Vs. Passive Crossovers
Vance Dickason in his Cookbook (5th ed.) claims improved transient response by replacing a passive crossover with an active one, citing a paper in JAES, Sept. 1971.
It would seem that Colloms in his book High Performance Loudspeakers, sixth ed. p. 325, takes the view that the damping factor (df) is only relevant to the region of resonance magnification as he discusses how the damping factor affects Q of the driver.
As back EMF as a percent of the driving voltage is very diminished except in the (frequency band) region of resonance magnification, or around the resonant frequency of the driver, I would argue that electrical damping or the df is irrelevant for example at a crossover frequency that is far removed from the resonant frequency of the driver. That is, diminished electrical damping of the driver has no affect on transient response of the driver at frequencies far removed from the resonant frequency of the driver.
Would anyone care to refute this. Not being an engineer, I'm open to being mistaken, but if so, then what is a logical explanation?
Regards,
Pete
Active Vs. Passive Crossovers
Vance Dickason in his Cookbook (5th ed.) claims improved transient response by replacing a passive crossover with an active one, citing a paper in JAES, Sept. 1971.
It would seem that Colloms in his book High Performance Loudspeakers, sixth ed. p. 325, takes the view that the damping factor (df) is only relevant to the region of resonance magnification as he discusses how the damping factor affects Q of the driver.
As back EMF as a percent of the driving voltage is very diminished except in the (frequency band) region of resonance magnification, or around the resonant frequency of the driver, I would argue that electrical damping or the df is irrelevant for example at a crossover frequency that is far removed from the resonant frequency of the driver. That is, diminished electrical damping of the driver has no affect on transient response of the driver at frequencies far removed from the resonant frequency of the driver.
Would anyone care to refute this. Not being an engineer, I'm open to being mistaken, but if so, then what is a logical explanation?
Regards,
Pete
Hi,
The so called "damping factor" is increased by driver impedance peaks
and impedance peaks caused by passive crossover section interaction.
These peaks only cause problems with low damping factors, as they
increase the damping factor causing variation of the voltage response.
(For higher damping factors the DF is still increased but it makes little
difference to the voltage response as long as DF is above about ten.)
Damping factor is relevant, not irrelevant, but not in the way you describe.
rgds, sreten.
The so called "damping factor" is increased by driver impedance peaks
and impedance peaks caused by passive crossover section interaction.
These peaks only cause problems with low damping factors, as they
increase the damping factor causing variation of the voltage response.
(For higher damping factors the DF is still increased but it makes little
difference to the voltage response as long as DF is above about ten.)
Damping factor is relevant, not irrelevant, but not in the way you describe.
rgds, sreten.
This is an interesting question. I'm writing a speaker simulation program and i already had the idea, to display the resulting damping factor of a driver combined with a passive crossover. It is obvious, that even order high- or low-pass filters will perform much better than odd order filters, because they have a shunting component.
Without thinking about it, i did assume, that a high damping factor is preferable in the whole pass-band of a driver. But is that really true?
I would say yes. As you mention, the back EMF induces a voltage in the voice coil. The major part is at the resonance frequency, but there are also other parts. Every irregularity in the impedance plot of a driver is caused by induction. A standing wave at 100 Hz, a surround resonance at 800 Hz or a cone resonance at 5 kHz, they are all visible in the impedance.
The other question is, will electrical damping really reduces these resonances? I guess, the effect is very limited, maybe a reduction of 1-2dB can be achieved.
Without thinking about it, i did assume, that a high damping factor is preferable in the whole pass-band of a driver. But is that really true?
I would say yes. As you mention, the back EMF induces a voltage in the voice coil. The major part is at the resonance frequency, but there are also other parts. Every irregularity in the impedance plot of a driver is caused by induction. A standing wave at 100 Hz, a surround resonance at 800 Hz or a cone resonance at 5 kHz, they are all visible in the impedance.
The other question is, will electrical damping really reduces these resonances? I guess, the effect is very limited, maybe a reduction of 1-2dB can be achieved.
Some simple truths can sum it up. Taken as systems the combination of active crossover/amp/driver and amp/passive crossover/driver represent predominately linear systems. Since linear systems with the same frequency response have the same transient response it follows that if the active and passive crossovers systems are designed to have identical frequency response then they must have the same transient response. Any difference is damping due to series impedance in the passive crossover simply means that the transfer function of the passive crossover will have to be somewhat different that that of the active system to account for the difference in damping.
The real difference between passive and active systems is insertion loss and the way the response to changes in VC temperature.
The real difference between passive and active systems is insertion loss and the way the response to changes in VC temperature.
Also, it should be understood that damping falls off at a rate of 6dB/octave to both sides of fs. In the flat band region of a driver, above fs, damping plays essentially no roll in determining the response of the driver to a given input signal which is why this region is referred to as the mass controlled region.
So you would agree that the presence of a passive crossover network (between power amp and driver) doesn't degrade transient response of two drivers attached to the network, assuming the cross-over frequency in the flat band region of the drivers?
The damping factor tells you what the internal resistance of the output stage of a power amplifier is. For the vast majority of solid state power amplifiers in existence now, that internal resistance is usually so low that it doesn't factor into Q of the speaker system connected to it. You are correct in saying that df strictly only applies to the amplifier. So but the electrical damping and thereby also Q of the driver certainly are affected by any substantial resistance in-series with the driver. There is the equation,
Qes' = Qes*[(Rg + Rx + Re)/ Re]
where
Qes' = modified electrical Q
Qes = electrical Q of the driver (specified Qes)
Rg = amplifier source resistance
Rx = resistance in-series with the driver
Re = voice coil resistance
If Rg is comparable to Re, then the damping factor also affects Q of the driver.
Regards,
Pete
I may be ignorant, but I think some of these seemingly sophisticated analyses are out to lunch. The issue isn't how the system works right but how it works wrong.
As the OP asks, when and for what ever reason there are false perturbations in the driver (as reflected in the voice coil), will the hook-up damp them? And the OP is right in the way he is looking at it. A good reason (among many other good reasons) never to consider passive crossovers unless you are a manufacturer who needs to sell integrated speaker systems in a box.
Frankly, for REAL damping, you want a negative output impedance amp AKA motional feedback.
Ben
As the OP asks, when and for what ever reason there are false perturbations in the driver (as reflected in the voice coil), will the hook-up damp them? And the OP is right in the way he is looking at it. A good reason (among many other good reasons) never to consider passive crossovers unless you are a manufacturer who needs to sell integrated speaker systems in a box.
Frankly, for REAL damping, you want a negative output impedance amp AKA motional feedback.
Ben
Actually, Q of the driver never changes. What changes is Q of the system composed of the driver, any series resistance, and Rg of the amp.
And as has been shown many times to be not that important. As you determine Qes', you will find any DF over about 20 is just not important. The tube camp seems to think 5 is about right as do the ultra low global feedback camp.
Passive crossovers do not degrade transient response any more than active do. Crossovers are FILTERS. Each driver does what it is told based on the filter transfer function and it's basic capabilities. Of course you are preventing the midrange from moving fast. That is what it does. "Speed" and "transient response" are totally bogus magazine reviewer audiophile BS. If your speaker lacks "impact" then it was designed poorly.
Now, active filters have the advantage of letting the amp deal with the ugly impedance problems of the drivers so they do not effect the crossover transfer function. They also have the advantage of allowing very complex equalization like a little bump, which is not easy in a passive. Don't forget to take into account any phase shift from the CD blocking caps that are still mandatory with a active crossover.
Passive crossovers do not degrade transient response any more than active do. Crossovers are FILTERS. Each driver does what it is told based on the filter transfer function and it's basic capabilities. Of course you are preventing the midrange from moving fast. That is what it does. "Speed" and "transient response" are totally bogus magazine reviewer audiophile BS. If your speaker lacks "impact" then it was designed poorly.
Now, active filters have the advantage of letting the amp deal with the ugly impedance problems of the drivers so they do not effect the crossover transfer function. They also have the advantage of allowing very complex equalization like a little bump, which is not easy in a passive. Don't forget to take into account any phase shift from the CD blocking caps that are still mandatory with a active crossover.
For the overly simplistic equation presented here, treat the impedance as resistance. The complex function is way the heck out of scope here. Over in the lounge are a few folks you know who are prepared to talk at that level. Of course they will tell you it does not matter anyway.
I read all the time that active filter are better because the transfer function doesn't change with changes in the driver's Z due to a variety of dynamic effects. Again, that is not an accurate statement. The thing that is overlooked is that this is limited to the voltage transfer function of the filter. But in reality, what determines the acoustic output of a driver is the current flow through the VC. Even in that case where the amplifiers output Z is zero applying a fixed voltage transfer function across the driver terminals does not mean the acoustic output remains constant under dynamic conditions. The driver VC Z changes with time (heating) and excursion and these factors will affect the current and therefor the radiated SPL. Active can make a lot of things easier, and I and a fan of actives. But many proponents of active make arguments for them which are overly simplistic and ignore many of the real issues resulting form the dynamically changing VC Z.
That's understood. However if I wanted to convey that perspective, then I'd have to come up with a new symbol like Qe-sys and I'd rather stick to the original T/S terminology as much as possible.
I am glad you understand that. Most people don't. The lack of considering the entire system, whether active or passive, is typically what leads to a lot of the claims about the superiority of active crossovers, many of which go by the wayside when the complete system is considered.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.