That is an interesting expression, 'dimensionless numbers' and yet they have real world ratios attached to them. In other words, they are useful.
But DF is a pointless ratio, in fact there is no useful ratio at all. Sorry. But of course, that I understand can be difficult to unlearn. You see, I used to think the same, only when I understood better the equations of Thiele and Small did I realise that there was no value in DF numbers. But DF96 did say something that is correct, it is indeed a reference to the output impedance of the amplifier, but only indirectly as you now have to calculate back and need two numbers to express what a single number already does perfectly.
You say it "seems clear that it represents the ability to deliver adequate drive to the load under conditions with demanding reactive loads." Again, this is a false impression. I have designed and built a transconductance amplifier with an output Z of 270 Ohm and it has no problem delivering 40 Watt into 8 Ohm load. The output Z is not the same as resistance, so correct me if that is not what you are thinking.
Let's face it, DF is the idea that the amplifier can add damping. Since the reference point is zero Ohm and if the amplifier's impedance it is a positive number, then the amplifier can only worsen damping. So indeed higher output impedance makes damping worse. Again, being fed the line that amplifiers can add damping, this is difficult to unlearn. Motional feedback can yield a negative impedance. That would be the only way to increase damping. Like Philips did decades ago.
BTW, what has been posted here is 100% EE compatible. If anybody reading this thinks otherwise, then proof, please.
Not worthy of a response, blatantly untrue statements! And defamatory too!
So you put a say 20R resistor across the loudspeaker terminals and despite the amplifier having decent damping/output impedance and no current limiting operation the sound changes......why ?.
With a 1kHz sine, I have observed change of the amount and proportion of harmonics, mainly H2 and H3.
You lower the output level? As I said playing with amplifier output impedance/cable impedance/speaker impedance is old stuff. Remove the fantastical and inaccurate claims and I would not bother saying anything. Remember if the shoe fits, i.e. typically the same folks that randomly cobble things together and "just listen" are equally guilty of not having the intellectual curiosity about what lies at the bottom of their results.
You do realize DF is frequency dependent within the audio band. A 20R resistor is not.
jn
Which fidelity information on which component were you looking for?
Can you cite an example of actual sound reaching your ears that misled you?
Of course it's strange to you when someone asks about listening test (WRT damping factor) when you were trying to spread FUD once again.
Of course seems that way to you. After all, you are in audio business and your objective here is to help your business.
It would help if you stopped mixing up plots of the modulus of the impedance and complex impedance. I'll ask again, do you not know how to express the circuits you present in terms of complex algebra or do you think the audience is incapable of understanding things at this level?
Scott, he already said he cannot use and does not understand maths of complex numbers. So the discussion is useless.
It's not the only way (if you mean sensor-based motional feedback). You can simply synthesize negative output resistance all the way down to -Re of the driver and call it a day. At that point the driver is fully velocity-controlled by itself (VC being actuator and sensor). Of course not really useful in real life because of unstable Re vs. temperature and nonlinearity of the actuator/sensor.
I'm trying to understand this and make sense of it, especially the part I put in bold.
Can anyone help? Am I trying to read too much into a few seemingly contradictory sentences?
You lower the driver resonance Z peaks. Those peaks, especially at the low frequencies cause the most back EMF. Limited Z of reson peaks limits/lowers the back emf.
Thx=RNMarsh
OK guys, now everybody is going to be confused about damping factor. Both sides are making the argument even worse.
Now, what is damping factor? It is simply the effective output impedance of the amplifier at some nominal frequency, let's say 1KHz, divided into 8ohms. That is it, so far as I know. Amps back in the tube days had plenty of problems keeping a high damping factor. It usually varied from 1(in rare cases) to 30( triodes), and would vary significantly over frequency. Why? Because the limited negative feedback possible (about 20dB) was not enough to give greater feedback to lower the effective output impedance further, AND on each side of the spectrum (below 100Hz and above 10KHz) the amount of feedback available was even less for practical reasons.
Solid state since the late 60's changed the situation. Direct coupling that was allowed in transistor amp design, allowed much more negative feedback to be applied, and it created an even greater damping factor that extended to much lower frequencies. However, at higher frequencies, the feedback rolloff was at least as bad as with tubes, and the effective damping factor at 10KHz or above was not very low. This is because high feedback amps must start rolling off their gain, internally (dominant pole), very early to not go unstable at ultrasonic frequencies. It is the same gain structure as a typical op amp, just with much more current and voltage out.
Now, for the other side: How important is damping factor (DF), the number?
Mostly useless, except as an advertising point. For example, I recall that one competitor claimed a DF of 5000 back in the early 70's. Almost impossible and entirely impractical to actually use. But is a DF of 100, better than a DF of 10? Maybe. Yet it can be shown that the relatively high series resistance of the voice coils completely compromises any high damping factor to almost meaningless in a real situation, even if it might change the overall frequency response of the speaker just a little. I would presume that Small, and others were taking this into account.
Now is a modestly high DF always the best? No. I have experimented with a DF of 1, on a particular direct radiator acoustic suspension speaker and found it sounded best. It was unfortunate, because the added series resistor, about 3.3 ohms threw away significant power, but when 2 or more of us heard it, it was obvious that the series resistor should remain.
Also, there is NEGATIVE DAMPING FACTOR, that will partially remove the effect of the voice coil resistance. The problem with this is if you use too much, then then you will get oscillation, so it is tricky.
Now, maintaining a good damping factor over the entire audio frequency range is actually fairly difficult, because of reduced feedback at higher frequencies, and the inclusion of any series inductors on the output (almost everybody uses them) that increase in impedance with frequency, completely dominating the DF at some point.
Enough for now. There is more.
Now, what is damping factor? It is simply the effective output impedance of the amplifier at some nominal frequency, let's say 1KHz, divided into 8ohms. That is it, so far as I know. Amps back in the tube days had plenty of problems keeping a high damping factor. It usually varied from 1(in rare cases) to 30( triodes), and would vary significantly over frequency. Why? Because the limited negative feedback possible (about 20dB) was not enough to give greater feedback to lower the effective output impedance further, AND on each side of the spectrum (below 100Hz and above 10KHz) the amount of feedback available was even less for practical reasons.
Solid state since the late 60's changed the situation. Direct coupling that was allowed in transistor amp design, allowed much more negative feedback to be applied, and it created an even greater damping factor that extended to much lower frequencies. However, at higher frequencies, the feedback rolloff was at least as bad as with tubes, and the effective damping factor at 10KHz or above was not very low. This is because high feedback amps must start rolling off their gain, internally (dominant pole), very early to not go unstable at ultrasonic frequencies. It is the same gain structure as a typical op amp, just with much more current and voltage out.
Now, for the other side: How important is damping factor (DF), the number?
Mostly useless, except as an advertising point. For example, I recall that one competitor claimed a DF of 5000 back in the early 70's. Almost impossible and entirely impractical to actually use. But is a DF of 100, better than a DF of 10? Maybe. Yet it can be shown that the relatively high series resistance of the voice coils completely compromises any high damping factor to almost meaningless in a real situation, even if it might change the overall frequency response of the speaker just a little. I would presume that Small, and others were taking this into account.
Now is a modestly high DF always the best? No. I have experimented with a DF of 1, on a particular direct radiator acoustic suspension speaker and found it sounded best. It was unfortunate, because the added series resistor, about 3.3 ohms threw away significant power, but when 2 or more of us heard it, it was obvious that the series resistor should remain.
Also, there is NEGATIVE DAMPING FACTOR, that will partially remove the effect of the voice coil resistance. The problem with this is if you use too much, then then you will get oscillation, so it is tricky.
Now, maintaining a good damping factor over the entire audio frequency range is actually fairly difficult, because of reduced feedback at higher frequencies, and the inclusion of any series inductors on the output (almost everybody uses them) that increase in impedance with frequency, completely dominating the DF at some point.
Enough for now. There is more.
BUT it is a 'perfect' voltage amplifier with milliohms output impedance so essentially zero drop in level and undetectable according to accepted standards of level change subjective detection.
I have 75R resistors across amp and across speaker....I sat outside while I had a few glasses of red and cranked the system a week ago and cooked one resistor as it turns out.
A few days later I sat in the chair, played a few tunes and noticed the tonal balance and central imaging wasn't quite right and when I investigated I noticed the burned o/c resistor at one speaker. The difference was subtle but enough to draw my attention that things weren't quite like they should be. Scott try it for yourself, connect a resistor across one channel and take a close listen.
Last edited:
Thanks and yes I know.....BUT the perfect voltage amplifier should not be fazed whatsoever, it must be that my amplifier is not so perfect as the DF spec would lead me to believe.
Yes of course I do....so shunt 20R supplies damping that the amplifier cannot, ie the perfect amplifier does not exist.
- Status
- Not open for further replies.