To give a qualitative picture on how „dry friction“ affects movement of of the diaphragm I set up a simplified simu
This simu assumes that there is 100% rigidity from VC to the points of velocity independent friction (spider and surround) – which is kind a optimistic – and also shows static and dynamic friction being the same (for simplicity of simulation).
First a plot that shows the situation when roughly 10% of the given max force is eaten up by friction.
Starting out with a 2Vpp source signal – membrane movement looks like that:
lowering source signal to 1Vpp signal – membrane movement looks like that:
lowering source signal to 500mVpp signal – membrane movement looks like that:
lowering source signal to 250mVpp signal – membrane movement looks like that:
lowering source signal to 200mVpp signal – membrane movement looks like that:
We see that distortion due to velocity independent friction starts to stick out as signal falls.
Right to the point where weak signals are no longer able to move the diaphragm at all.
We also see that movement starts delayed as the force has to reach a certain level first before movement begins.
Also we see that movement stops before reaching peak points as there is no longer enough force to pull to sine top.
The DC shift seen here would level out in reality of course – and we would more clearly see the rest position to be offset...
Well this might be a „little bit“ exaggerated – as we hardly would see it happen in such pure form and at such level – but the basic principle clearly is shown IMO.
Michael
This simu assumes that there is 100% rigidity from VC to the points of velocity independent friction (spider and surround) – which is kind a optimistic – and also shows static and dynamic friction being the same (for simplicity of simulation).
First a plot that shows the situation when roughly 10% of the given max force is eaten up by friction.
Starting out with a 2Vpp source signal – membrane movement looks like that:
lowering source signal to 1Vpp signal – membrane movement looks like that:
lowering source signal to 500mVpp signal – membrane movement looks like that:
lowering source signal to 250mVpp signal – membrane movement looks like that:
lowering source signal to 200mVpp signal – membrane movement looks like that:
We see that distortion due to velocity independent friction starts to stick out as signal falls.
Right to the point where weak signals are no longer able to move the diaphragm at all.
We also see that movement starts delayed as the force has to reach a certain level first before movement begins.
Also we see that movement stops before reaching peak points as there is no longer enough force to pull to sine top.
The DC shift seen here would level out in reality of course – and we would more clearly see the rest position to be offset...
Well this might be a „little bit“ exaggerated – as we hardly would see it happen in such pure form and at such level – but the basic principle clearly is shown IMO.
Michael
Last edited:
Quite in contrary to above – friction that depends on velocity does not show such ill behaviour.
An example for „velocity dependent friction“ could be the air that is pushed through the VC gap by the cup.
Both types of friction contribute to Qms, but have way different impact sonically.
Michael
An example for „velocity dependent friction“ could be the air that is pushed through the VC gap by the cup.
Both types of friction contribute to Qms, but have way different impact sonically.
Michael
Ah the evil zero crossing again!
We don't like that! It's why some of us charge couple our capacitors a la jbl, something other people using CD 2-ways may wish to consider given that they're mostly already using high Qms drivers and Dr. Geddes recommends low crossover distortion amplifiers
We don't like that! It's why some of us charge couple our capacitors a la jbl, something other people using CD 2-ways may wish to consider given that they're mostly already using high Qms drivers and Dr. Geddes recommends low crossover distortion amplifiers
Ah the evil zero crossing again!
We don't like that!
....and Dr. Geddes recommends low crossover distortion amplifiers
he might have been shooting the wrong suspects.. but thats an old story ...
Michael
Last edited:
If you just want a judgement which is better, I really can't tell with the graphs flipping like that. I'm used to laying them side by side and studying them that way. The only thing I can tell from the way they are is that they will sound different.
I'm starting to doubt that coulomb friction is significant. If it were THD measurements would surely show it.
If this kind of friction is measureable, smaller test signals will reveal it more significantly, thus the THD measurements would show higher levels of distortion with smaller test signals. Wouldn't it?
That was my point. I have looked at THD at low levels and didn't see anything that would indicate this type of "notch" distortion.
I'm certainly the first to agree that this kind of distortion isn't any significant for the vast majority.
Otherwise it simply would not have been hidden for centuries.
On the other hand we possibly should not classify „significant“ with a mic and a osci at hand – I guess we would be in trouble to debunk more subtle effects that way – you wouldn't do it to demonstrate dither for example, would you?
Also, as much as I'm interested in any flame or snake oil discussion – I don't think its the right time - and also - there will be endless time left for endless discussion about significance AFTER we possibly have traced velocity independent friction down – if thats possible at all.
As for measurement I found that a *relative* approach might work best to begin with.
Looking at the spectra of the trapezoid waveform basically created by VIF, we certainly have it easier to look at changes in the relation of 2nd and higher order distortion correlating with SPL.
Usually (IMO), higher order order distortion comes up with any speaker at low SPL – we should sort out these drivers that come up with higher order distortion at lowest SPL levels to find the most „dynamic“ ones.
As a simple relative measure (not necessarily the best I admit), I have checked at which SPL level higher order distortion overtakes 2nd order.
As this is a mere relative measure, I gained advantage by not doing this acoustically but by direct measurement of the current when feeding the speaker with constant voltage.
A loooot of *if* an *when* – I know...
Michael
Otherwise it simply would not have been hidden for centuries.
On the other hand we possibly should not classify „significant“ with a mic and a osci at hand – I guess we would be in trouble to debunk more subtle effects that way – you wouldn't do it to demonstrate dither for example, would you?
Also, as much as I'm interested in any flame or snake oil discussion – I don't think its the right time - and also - there will be endless time left for endless discussion about significance AFTER we possibly have traced velocity independent friction down – if thats possible at all.
As for measurement I found that a *relative* approach might work best to begin with.
Looking at the spectra of the trapezoid waveform basically created by VIF, we certainly have it easier to look at changes in the relation of 2nd and higher order distortion correlating with SPL.
Usually (IMO), higher order order distortion comes up with any speaker at low SPL – we should sort out these drivers that come up with higher order distortion at lowest SPL levels to find the most „dynamic“ ones.
As a simple relative measure (not necessarily the best I admit), I have checked at which SPL level higher order distortion overtakes 2nd order.
As this is a mere relative measure, I gained advantage by not doing this acoustically but by direct measurement of the current when feeding the speaker with constant voltage.
A loooot of *if* an *when* – I know...
Michael
I wouldn't be so quick to dismiss what a mic can pick up. Any mico detail that is in the recording is only there becase a mic picked it up in the first place. And certainly a mic is much, much more sensitive than the ear, regardless of how golden that ear might be.
I have not even tried to do it acoustically - possibly I should have.
My line of thinking was that with electrically measurement we get cleaner signal to analyse as all the noise around - which is at my place around 30-40dB - will be heavily dampened (a speaker will make for a rather bad mic).
Sure - in case of steady signals analysis like THD, we can increase time interval and hence statistically lower noise but IMO its always good to start out with as clean as possible a signal.
A second thought is, that with direct measurement we simply exclude one pretty unknown variable : the mic quality
#############
Regardless of measurement technique - do you get the rise in higher order distortion at falling SPL as well and what are your conclusions ?
#############
My crumbly measurements so far can be looked up here:
http://www.diyaudio.com/forums/mult...le-dipole-beyma-tpl-150-a-12.html#post2044838
and here:
http://www.diyaudio.com/forums/mult...le-dipole-beyma-tpl-150-a-12.html#post2045077
Michael
My line of thinking was that with electrically measurement we get cleaner signal to analyse as all the noise around - which is at my place around 30-40dB - will be heavily dampened (a speaker will make for a rather bad mic).
Sure - in case of steady signals analysis like THD, we can increase time interval and hence statistically lower noise but IMO its always good to start out with as clean as possible a signal.
A second thought is, that with direct measurement we simply exclude one pretty unknown variable : the mic quality
#############
Regardless of measurement technique - do you get the rise in higher order distortion at falling SPL as well and what are your conclusions ?
#############
My crumbly measurements so far can be looked up here:
http://www.diyaudio.com/forums/mult...le-dipole-beyma-tpl-150-a-12.html#post2044838
and here:
http://www.diyaudio.com/forums/mult...le-dipole-beyma-tpl-150-a-12.html#post2045077
Michael
Last edited:
Sorry to come back to this so late Michael. Please look again.
Enable Tests
What I was pointing to was the surface development of the major resonance. If you look carefully you should be able to see the actual steps of resonant activity, as the EnABL'd driver losses energy. The basic shape to the FR is really quite close, with the changes you noted, but the real indication of micro dynamic abilities of the driver are in that stepped reveal of the resonant node, as the amplitude drops. At least in my opinion, this shows that drivers are really very competent in this realm, in every respect, except how they couple with the air.
Bud, not sure if I'm looking at the same places as you do ?
Here I marked the two areas where I see the major changes:
The marked area lower in frequency is considerably smoothed by enABL IMO, wheras the marked area at higher frequency comes slightly alive .
This I didn't get - can you give some more explanation?
Michael
Here I marked the two areas where I see the major changes:
The marked area lower in frequency is considerably smoothed by enABL IMO, wheras the marked area at higher frequency comes slightly alive .
This I didn't get - can you give some more explanation?
Michael
Ahh I guess I got it now - you are referring to the shape of the resonance that drops in steps originally and which - after enABL treatment - is a smooth rush !
You possibly know we had some similar effects been discussing with looped reflections in the horn honk and wavelet thread..
Well - now I get your point, and I for sure agree that micro details *must* get less masked when you effectively lower looped reflections - or whatever here exactly is involved - by enABL treatment
Michael
You possibly know we had some similar effects been discussing with looped reflections in the horn honk and wavelet thread..
Well - now I get your point, and I for sure agree that micro details *must* get less masked when you effectively lower looped reflections - or whatever here exactly is involved - by enABL treatment
Michael
You need to roll your cursor on and off of the plot to see the comparison. The red plot is the untreated cone, the blue plot the treated one.
You are looking in the correct places however. My point was that the difference displayed in the waterfall, between the red and blue plots, shows a marked difference in portrayed ringing of the resonance frequency. If you look closely, as you flip between the red and blue plots, you can see a very well defined set of steps, looking much like rice paddies down the side of a hill, in the blue plot. The same stepped portrayal of resonant ringing is not shown on the red plot, yet it is the same driver, test set up and tester involved.
What I am pointing to is that the driver can already perform to this micro level. To me this shows that it is already capable of very fine resolution in the motor and suspension systems and that it has correct damping in the surround. All that is missing from the red plot, untreated driver, is the expression of that micro signal control into the air. Once that coupling obstruction is overcome, the driver shows us what it was already capable of, but was unable to provide as a signal into the surrounding air.
I side with John here, in thinking that the major constraint to improvement in driver micro detail performance is not to be found in the mechanical frictions, though improving them will certainly provide the potential for less distortion to these tiny signals. Where John and I differ is how best to allow those tiny signals to be expressed.
In summation, I think that both the software and microphones are more than capable of displaying these issues. The only fault to be found may be in our lack of experience in recognizing these issues when they are displayed.
You are looking in the correct places however. My point was that the difference displayed in the waterfall, between the red and blue plots, shows a marked difference in portrayed ringing of the resonance frequency. If you look closely, as you flip between the red and blue plots, you can see a very well defined set of steps, looking much like rice paddies down the side of a hill, in the blue plot. The same stepped portrayal of resonant ringing is not shown on the red plot, yet it is the same driver, test set up and tester involved.
What I am pointing to is that the driver can already perform to this micro level. To me this shows that it is already capable of very fine resolution in the motor and suspension systems and that it has correct damping in the surround. All that is missing from the red plot, untreated driver, is the expression of that micro signal control into the air. Once that coupling obstruction is overcome, the driver shows us what it was already capable of, but was unable to provide as a signal into the surrounding air.
I side with John here, in thinking that the major constraint to improvement in driver micro detail performance is not to be found in the mechanical frictions, though improving them will certainly provide the potential for less distortion to these tiny signals. Where John and I differ is how best to allow those tiny signals to be expressed.
In summation, I think that both the software and microphones are more than capable of displaying these issues. The only fault to be found may be in our lack of experience in recognizing these issues when they are displayed.