Well, that might not work very well on OB speakers😛
The sound of the cone being tapped contains also the resonance of the cabinet, probably by a very large degree.
So I'd suggest the damping should be considered as a whole system of amp-speaker-room combo. Among them, I think the biggest problem is in speaker itself, not amp.
In a single ended amplifier where the negative output are ground you power supply are the other halve of the bridge.
Due to this the power supply will directly have effect on the damping factor at lower frequencies where the capacitors in the power supply increase in impedance.
Feedback will reduce the measured effect of this.
I thought we were talking about typical SS amps - very few single-ended! I guess I could argue that an SE amp is "strange", and that the PSU there is part of the output topology. One can always find rare exceptions to almost any statement.
A single ended amplifier in this respect are not an SE valve amp but a standard solid state transistor amplifier...
Currently I am very happy with a digital amp and a wide-band midrange driver with high second order harmonic distortion and a large series resistor that reduces the damping factor. A SE simulator, if you like.
Last edited:
As I said, "a standard solid state transistor amplifier" is almost invariably push-pull so the power supply has much less effect on damping factor than other factors. I didn't mention valve SE. Transistors can be SE, but rarely are.
But are usually very good. And given the number of passdiy amps members of the forum have. much more common here than in the general population.
SE amps -- any amps for that matter -- cannot be waved aside in this discussion.
dave
Everything I could find with Google was someone "estimating" 15-20 for the SA-XR55. How can I calculate the effective damping factor together with the series resistor?
Huh. As a speaker engineer, I found this and the rest of the comment quite interesting. There IS a big tendency to overlook back EMF...
I've mentioned this fact before but I think it bears repeating. For a speaker with the typical LC crossover driven by a lowish output impedance amp, there is ZERO cone damping at the crossover frequency and precious little some way each side of it. This is because the back EMF basically has nowhere to go as it sees a parallel LC high Q resonance circuit, not a series LP filter as the woofer's forward signal does. Same with the tweeter and its HP filter.
The bad effects of this are easily heard once you know what to listen for. A few lossy resistors in the crossover help somewhat (or active amps of course).
The bad effects of this are easily heard once you know what to listen for. A few lossy resistors in the crossover help somewhat (or active amps of course).
'Back EMF' and forward current see exactly the same Thevenin output impedance, unless the amplifier incorporates a directional coupler in its output stage. A directional coupler would need to sense both voltage and current; I have never seen one in an amp and can think of no reason why anyone would wish to put one there.
There would not be much need for cone damping at the crossover frequency, only around and below the bass resonance. Elsewhere the air and enclosure provides damping. It is true that the crossover may present a parallel tuned circuit to the drive unit at the crossover frequency but it will be very low Q as it is damped by the driver's own resistance. Unlikely to have a 'bad effect'. Fiddling with a crossover will of course be audible.
There would not be much need for cone damping at the crossover frequency, only around and below the bass resonance. Elsewhere the air and enclosure provides damping. It is true that the crossover may present a parallel tuned circuit to the drive unit at the crossover frequency but it will be very low Q as it is damped by the driver's own resistance. Unlikely to have a 'bad effect'. Fiddling with a crossover will of course be audible.
I don't know what's typical, but the ICEpower modules I've used have a very low output impedance. About 20mΩ for the smaller amps and under 5mΩ for the large ones. That's measured by feeding 1A into the outputs.
They do seem to have a very strong, tight bass.
I think you miss the point. Agreed the overall electrical Q of the LCR network will be very low (around Q=0.8 for values used in my 3 way's LP section), but as far as the speaker coil is concerned (which is where back emf comes from) it 'sees' an open circuit at the crossover. In other words, an 8r0 voice coil in parallel with a perfect LC becomes an effective 8r0 circuit overall at the crossover - that's what you're saying same as me. However this also means that the voice coil looks back into an open circuit as 8r0 in parallel with infinity ohms is 8r0. So, no damping because it's just the same as if nothing at all is connected to the speaker voice coil (at the crossover).
A higher amplifier output impedance actually helps reduce this effect as the parallel LC becomes much more lossy due to the series effective resistance and the speaker coil looks back into a non open circuit.
OK, but as I said no extra damping is needed at the crossover frequency. Electrical damping only works at frequencies where there is tight coupling between the mechanical resonance and the voice coil. This means the bass resonance only, as at higher frequencies the cone flexibility prevents this.
I'm sure you can hear something, but what? Adding resistors into the crossover will change the amplitude and phase in the crossover region, where phase is particularly important due to having two drive units active.
Last time I checked a transistor was a directional coupler of sorts so output impedance is not necessarily (as in never) the same as the input impedance at the output terminals. Further, the VC moving in the gap always produces an EMF and power which should, ideally, be absorbed by the amplifier. The fact the cone flexes occurs at all frequencies and does not change the fact a VC is moving in a gap making EMF. I have quite effectively demonstrated many times an amplifier can somewhat tame a wild driver at any frequency. Speaker resonances cause peaks in the VC motion which generate peaks in the EMF which can be damped.
As for damping at the crossover, the crossover is just a series impedance such that in the case of a low pass the damping decreases with increasing frequency around the crossover reaching values of less than one somewhere above the crossover frequency. Of course this assumes the crossover itself does not have a resistor or network across the driver terminals which could damp the speaker somewhat at all frequencies.
As for damping at the crossover, the crossover is just a series impedance such that in the case of a low pass the damping decreases with increasing frequency around the crossover reaching values of less than one somewhere above the crossover frequency. Of course this assumes the crossover itself does not have a resistor or network across the driver terminals which could damp the speaker somewhat at all frequencies.
There is quite a bit of discussion about damping factor in my book "Designing Audio Power Amplifiers". The first thing to recognize is that damping factor is simply a way of describing the source impedance (output impedance) of the amplifier. The term damping factor unfortunately tends to constrain the discussion of damping factor to how well the amplifier "damps" the loudspeaker. This is only a small part of the story, as damping factor affects frequency response, often in subtle ways across the whole spectrum.
It is quite important to recognize as well that amplifier output impedance (hence damping factor) is a function of frequency. Often if a single number is quoted for DF, it is at low frequencies. DF is usually smaller at higher frequencies.
The best way to measure damping factor is to apply a current to the output of the amplifier and measure the resulting voltage. It is an OK approximation to deliver thus current through a resistor driven by another amplifier. In fact, if the driving resistor is 8 ohms, the DF is approximately equal to the driving test voltage divided by the voltage that results across the output terminals of the amplifier under test. This approximation is reasonable for DF significantly greater than 10. A more accurate result can be obtained by using an 80 ohm series resistor and scaling the result by a factor of 10.
This is a small-signal measurement of the damping factor. As long as the amplifier has enough output current capability, its large-signal DF will be the same as its measured small-signal DF, and the amplifier will not care about the presence of counter EMF or its resulting current.
Always bear in mind that the source impedance of the amplifier forms a frequency-dependent voltage divider with the load impedance presented by the loudspeaker.
In theory, most loudspeakers are designed to have a flat frequency response when driven with a perfect voltage source - in other words as if they are being driven by an amplifier with infinite damping factor. In practice, however, most loudspeakers are "voiced" when driven with a real amplifier with a finite damping factor. It is important to recognize this, because the finite source impedance of a real amplifier affects frequency response. As a side note, consider a speaker manufacturer who loves the sound of vacuum tube amplifiers, which normally have a low DF. If they voice their speakers with a VT amplifier and then the end user listens to them with an SS amplifier, the frequency response may be considerably different.
DF affects frequency response in large part because loudspeakers usually have a highly varying impedance as a function of frequency - not just at the low end where woofer resonances occur, but usually at the crossover frequencies as well. It is not unusual for an 8-ohm-rated loudspeaker to have its impedance go as low as 6 ohms (or lower) and as high as 20 ohms (or higher).
Many believe they can hear a frequency response coloration of just 0.1 dB. If you do the math, it becomes clear that DF must be quite high for that loudspeaker impedance variation to not cause more than 0.1 dB of frequency response change, even when DF itself is constant with frequency.
Cheers,
Bob
It is quite important to recognize as well that amplifier output impedance (hence damping factor) is a function of frequency. Often if a single number is quoted for DF, it is at low frequencies. DF is usually smaller at higher frequencies.
The best way to measure damping factor is to apply a current to the output of the amplifier and measure the resulting voltage. It is an OK approximation to deliver thus current through a resistor driven by another amplifier. In fact, if the driving resistor is 8 ohms, the DF is approximately equal to the driving test voltage divided by the voltage that results across the output terminals of the amplifier under test. This approximation is reasonable for DF significantly greater than 10. A more accurate result can be obtained by using an 80 ohm series resistor and scaling the result by a factor of 10.
This is a small-signal measurement of the damping factor. As long as the amplifier has enough output current capability, its large-signal DF will be the same as its measured small-signal DF, and the amplifier will not care about the presence of counter EMF or its resulting current.
Always bear in mind that the source impedance of the amplifier forms a frequency-dependent voltage divider with the load impedance presented by the loudspeaker.
In theory, most loudspeakers are designed to have a flat frequency response when driven with a perfect voltage source - in other words as if they are being driven by an amplifier with infinite damping factor. In practice, however, most loudspeakers are "voiced" when driven with a real amplifier with a finite damping factor. It is important to recognize this, because the finite source impedance of a real amplifier affects frequency response. As a side note, consider a speaker manufacturer who loves the sound of vacuum tube amplifiers, which normally have a low DF. If they voice their speakers with a VT amplifier and then the end user listens to them with an SS amplifier, the frequency response may be considerably different.
DF affects frequency response in large part because loudspeakers usually have a highly varying impedance as a function of frequency - not just at the low end where woofer resonances occur, but usually at the crossover frequencies as well. It is not unusual for an 8-ohm-rated loudspeaker to have its impedance go as low as 6 ohms (or lower) and as high as 20 ohms (or higher).
Many believe they can hear a frequency response coloration of just 0.1 dB. If you do the math, it becomes clear that DF must be quite high for that loudspeaker impedance variation to not cause more than 0.1 dB of frequency response change, even when DF itself is constant with frequency.
Cheers,
Bob
- Status
- Not open for further replies.