Re: Re: Re: Re: Nice ideas, but...
Konnichiwa,
As it so happens, yes. I used in the end electrect mike capsulas. Not particulary linear either, but in the mid 80's about the best I could get hold of and AT LEAST there was no hysteresis in the system.
Well, no points for guessing then.
Sayonara
Konnichiwa,
Pjotr said:
As it so happens, yes. I used in the end electrect mike capsulas. Not particulary linear either, but in the mid 80's about the best I could get hold of and AT LEAST there was no hysteresis in the system.
Pjotr said:
Well, no points for guessing then.
Sayonara
Re: Re: Re: Re: Re: Nice ideas, but...
They work fine used as an acceleration sensor but are quite noisy. Used as a sound pressure sensor you are asking for problems indeed. Piezo’s are much less noisy and can have much more extended response at the low frequency end. Hysteresis is not that bad as you tend to put up here.
Regarding your grumpy comments on it you must have end up in frustration then.
Cheers 😉
Kuei Yang Wang said:
They work fine used as an acceleration sensor but are quite noisy. Used as a sound pressure sensor you are asking for problems indeed. Piezo’s are much less noisy and can have much more extended response at the low frequency end. Hysteresis is not that bad as you tend to put up here.
Regarding your grumpy comments on it you must have end up in frustration then.
Cheers 😉
Re: Re: Re: Re: Re: Re: Nice ideas, but...
Konnichiwa,
I tried the Philips approach first and was not impressed. Good Mikes are very quiet, can handle more than high enough SPL's and usually are pretty linear.
Not at all, but as drivers kept going bang when things got lou I ended up building a fairly complex protection system in, derived from PA "Processor" applications to protect the Drivers, including a slow thermal limiter and a variable high pass to limit driver excursion.
In the end my dual 8" MFB active system still sounded not really better than any of my others speakers at the time (all medium / high sensitivity systems out of Pro Audio / Studio applications) when driven byt Studio grade Valve Amp's, so I conluded it was in real terms a waste of time, if an excellent engineering challenge and interesting to solve.
That said, I have ALWAYS preferred to implement and design as accurate as possible open loop systems over looped systems. I have since also noted that having too low distortion at low frequencies is a liability, sonically speaking, not an asset, but that as they say is another story.
Sayonara
Konnichiwa,
Pjotr said:
I tried the Philips approach first and was not impressed. Good Mikes are very quiet, can handle more than high enough SPL's and usually are pretty linear.
Pjotr said:
Not at all, but as drivers kept going bang when things got lou I ended up building a fairly complex protection system in, derived from PA "Processor" applications to protect the Drivers, including a slow thermal limiter and a variable high pass to limit driver excursion.
In the end my dual 8" MFB active system still sounded not really better than any of my others speakers at the time (all medium / high sensitivity systems out of Pro Audio / Studio applications) when driven byt Studio grade Valve Amp's, so I conluded it was in real terms a waste of time, if an excellent engineering challenge and interesting to solve.
That said, I have ALWAYS preferred to implement and design as accurate as possible open loop systems over looped systems. I have since also noted that having too low distortion at low frequencies is a liability, sonically speaking, not an asset, but that as they say is another story.
Sayonara
I had very good luck with the cheap Panasonic electrets
when I used them on the Shadow acoustic tube. I simply
put them in the center coming out of the dust cap (the magnet
had a hole for a brass tube) and applied about 6 dB worth of
feedback to keep the pressure near zero on the speaker cone.
More feedback than that gets problematic, so when you're
using this technique for motional feedback, it's wise to use
feedback sparingly.
when I used them on the Shadow acoustic tube. I simply
put them in the center coming out of the dust cap (the magnet
had a hole for a brass tube) and applied about 6 dB worth of
feedback to keep the pressure near zero on the speaker cone.
More feedback than that gets problematic, so when you're
using this technique for motional feedback, it's wise to use
feedback sparingly.
Assuming that one were to build a state-of-the-art driver from scratch to be powered by an amplifier with a relatively high output impedance, would it be possible to achieve critical mechanical damping (i.e. Qms = 0.5) through driver alone? If it were possible, what other factors would there be that would make such an endeavour pointless?
Is sensitivity also calculated by the same formula, then? i.e. (constant * Fs^3 * Vas)/Qes?
Is sensitivity also calculated by the same formula, then? i.e. (constant * Fs^3 * Vas)/Qes?
Sure, but were I designing such a wideband driver for such an amplifier (and, in fact, I have been 😉 ) I would not limit myself to a single damping value, as that would limit the possible alignments. I would prefer the driver be adaptable to most or all of the alignments we've discussed in this thread, including various measures of flow-resistance external to the driver itself.
Cost be darned, I would start with an underhung field-coil motor with full faraday treatment, a variable eddy-current brake, and perhaps magnetic suspension. Now both damping and Cms would be variable with field current. In conjuction with an adjustable Zout amp, you could tailor your response shape quite a bit.
I would specify a cone perhaps of mixed-mode function where the smaller circumferences would use the shear-wave model (ala Jordan), with a star edge to break up reflections and a low-impedance juncture to more lossy doped paper outer regions, which would steadily roll on resistive wave termination. I'd pay close attention to waterfall plots, and tweak until I got it right. I would probably also specify a star-edge steadily thickening felt overlay on the outer cone edges to damp any hint of surround modes or transition diffraction. I believe it would also be possible to apply the felt to the cone in such a way as to control the treble pattern, to make it throw something like a 40x90deg. pattern, for example. There's many more possibilities for nifty suspension and damping...
And I've got a couple other tricks in mind to improve the onset response, raise the frequency of the first standing wave, and even to provide an inherent adjustability for baffle step, but I'll save those for now.
Cost be darned, I would start with an underhung field-coil motor with full faraday treatment, a variable eddy-current brake, and perhaps magnetic suspension. Now both damping and Cms would be variable with field current. In conjuction with an adjustable Zout amp, you could tailor your response shape quite a bit.
I would specify a cone perhaps of mixed-mode function where the smaller circumferences would use the shear-wave model (ala Jordan), with a star edge to break up reflections and a low-impedance juncture to more lossy doped paper outer regions, which would steadily roll on resistive wave termination. I'd pay close attention to waterfall plots, and tweak until I got it right. I would probably also specify a star-edge steadily thickening felt overlay on the outer cone edges to damp any hint of surround modes or transition diffraction. I believe it would also be possible to apply the felt to the cone in such a way as to control the treble pattern, to make it throw something like a 40x90deg. pattern, for example. There's many more possibilities for nifty suspension and damping...
And I've got a couple other tricks in mind to improve the onset response, raise the frequency of the first standing wave, and even to provide an inherent adjustability for baffle step, but I'll save those for now.
Let's take the example of a bass(ish) driver, i.e. something around the diameter of 12" or so. We'll be operating well above any breakup modes the diaphragm may have, so basically any material that's rigid enough will do. I think, for lower frequencies, the more rigid the material, the more pistonic the behaviour, so we could go for a woven carbon fiber/epoxy pre-preg cloth (radially pleated around the coil former, if we wanted, for extra stiffness in the axial dimension), or perhaps a diamond-coated (either diamond-like carbon, or crystalline diamond, if we wanted to go for the extra cost, both to increase mechanical properties slightly, but more to greatly increase thermal conductivity from the coil former, which should also be thermally conductive, since most materials that conduct electricity also conduct heat well) beryllium diaphragm - sorry for the long sentence.
Now, a Q of 0.5, as you well know, means a return from peak to nothing as quickly as possible without any oscillation, which is why I was asking about a Qms of 0.5, since with high output impedance Qes typically has no role in damping the behaviour of the diaphragm. Since this driver is no-holds-barred, why not use it in a dipole situation, where the in-room bass response benefits greatest from dipolar radiation? On an OB, the Qtc (Qtd) is about the same as Qts of the driver, so...
A magnetic suspension is feasible, but as you said yourself, a suitable restoring force might only be possible with a very long magnetic ring, which is heavy, unless a lighter magnetic alloy was used (an Al/Ni alloy as suggested by KYW). But then damping and Cms would not be independently variable... if they are, I am obviously not seeing the whole picture.
Re: your comments on the shear-wave model, do you have any links with information about it? How would you also merge the "low-impedance juncture" to a lossy doped paper outer region? Could you clarify "resistive wave termination" - or is that just a fancy term for a lossy surround? 😉
Now, a Q of 0.5, as you well know, means a return from peak to nothing as quickly as possible without any oscillation, which is why I was asking about a Qms of 0.5, since with high output impedance Qes typically has no role in damping the behaviour of the diaphragm. Since this driver is no-holds-barred, why not use it in a dipole situation, where the in-room bass response benefits greatest from dipolar radiation? On an OB, the Qtc (Qtd) is about the same as Qts of the driver, so...
A magnetic suspension is feasible, but as you said yourself, a suitable restoring force might only be possible with a very long magnetic ring, which is heavy, unless a lighter magnetic alloy was used (an Al/Ni alloy as suggested by KYW). But then damping and Cms would not be independently variable... if they are, I am obviously not seeing the whole picture.
Re: your comments on the shear-wave model, do you have any links with information about it? How would you also merge the "low-impedance juncture" to a lossy doped paper outer region? Could you clarify "resistive wave termination" - or is that just a fancy term for a lossy surround? 😉
Bass driver is fine, I just keep coming back to widebands because I'm attracted to the possibilites they present with current drive. 🙂 And OB is IMO a great alignment, I just don't think it's necessarily all things to all people. I'm leaning more and more toward trying for a cardioid-esqe pattern in my next buildup, so I might want a slightly higher Qm to bring back down with stuffing. ... Just where I'm comin' from. 
So in the case of your bass driver, at first blush you could just use a conductive former and skip all the other fluff that's been talked about. After all, if you're really going to stay pistonic, then the movement of the VC will be essentially identical to that of everything else, and you can do all your damping there.
However, even when a cone is completely pistonic and limited to a low band, if the motor is ever stressed, the cone will radiate a spread of harmonics no matter how the signal is filtered. Those harmonics could excite resonances far above the passband, which is when it suddenly becomes a good idea again to have additional damping elsewhere on the cone/surround. 🙂
Well, I was commenting on the necessary length of a magnetic ring in a XBL-style motor. In a more typical under or overhung motor, the ring would be about the same height as the VC, and you'd use some other mechanical means to catch a runaway VC.
Damping via an eddy-current brake and Cms via a magnetic ring both scale with B-field density.
Man, I wish there was a good source of shear-wave design info online, but I don't know of one. There was a great discussion of it on the fullrange forum a couple years ago, but I'm not sure where it went after the forum update.
By "low-impedance juncture," I just mean a tight bond between the dissimilar materials with a smooth gradual transition (provided by the starred edge). The idea being to provide an easy exit of high-speed waves off the inner part of the diaphragm so they don't reflect. Then they'd slow and be resistively terminated as they passed outward through gradually thickening lossy diaphragm material, so there would be very little energy left to reflect off boundaries of mechanical impedance mismatch, like the cone/surround transition.
So in the case of your bass driver, at first blush you could just use a conductive former and skip all the other fluff that's been talked about. After all, if you're really going to stay pistonic, then the movement of the VC will be essentially identical to that of everything else, and you can do all your damping there.
However, even when a cone is completely pistonic and limited to a low band, if the motor is ever stressed, the cone will radiate a spread of harmonics no matter how the signal is filtered. Those harmonics could excite resonances far above the passband, which is when it suddenly becomes a good idea again to have additional damping elsewhere on the cone/surround. 🙂
Well, I was commenting on the necessary length of a magnetic ring in a XBL-style motor. In a more typical under or overhung motor, the ring would be about the same height as the VC, and you'd use some other mechanical means to catch a runaway VC.
Damping via an eddy-current brake and Cms via a magnetic ring both scale with B-field density.
Man, I wish there was a good source of shear-wave design info online, but I don't know of one. There was a great discussion of it on the fullrange forum a couple years ago, but I'm not sure where it went after the forum update.
By "low-impedance juncture," I just mean a tight bond between the dissimilar materials with a smooth gradual transition (provided by the starred edge). The idea being to provide an easy exit of high-speed waves off the inner part of the diaphragm so they don't reflect. Then they'd slow and be resistively terminated as they passed outward through gradually thickening lossy diaphragm material, so there would be very little energy left to reflect off boundaries of mechanical impedance mismatch, like the cone/surround transition.
Bill F. said:Bass driver is fine, I just keep coming back to widebands because I'm attracted to the possibilites they present with current drive. 🙂 And OB is IMO a great alignment, I just don't think it's necessarily all things to all people. I'm leaning more and more toward trying for a cardioid-esqe pattern in my next buildup, so I might want a slightly higher Qm to bring back down with stuffing. ... Just where I'm comin' from.
So in the case of your bass driver, at first blush you could just use a conductive former and skip all the other fluff that's been talked about. After all, if you're really going to stay pistonic, then the movement of the VC will be essentially identical to that of everything else, and you can do all your damping there.
However, even when a cone is completely pistonic and limited to a low band, if the motor is ever stressed, the cone will radiate a spread of harmonics no matter how the signal is filtered. Those harmonics could excite resonances far above the passband, which is when it suddenly becomes a good idea again to have additional damping elsewhere on the cone/surround. 🙂
You mean, you want those things to reduce HD?
Well, I was commenting on the necessary length of a magnetic ring in a XBL-style motor. In a more typical under or overhung motor, the ring would be about the same height as the VC, and you'd use some other mechanical means to catch a runaway VC.
My mistake. For a bass application, I'm pretty sure either an XBL^2 or NRT topology would be ideal, or perhaps even a mix of both.
Damping via an eddy-current brake and Cms via a magnetic ring both scale with B-field density.
Yes, I'd realized that, but you wouldn't be able to change both separately (i.e. on the fly) which is what I meant.
Man, I wish there was a good source of shear-wave design info online, but I don't know of one. There was a great discussion of it on the fullrange forum a couple years ago, but I'm not sure where it went after the forum update.
I hate when information gets lost. Ask your grandparents about their stories, if you're lucky enough to still have them around! 🙂
By "low-impedance juncture," I just mean a... space-saving ... transition.
Understood... But are you using the starred edge between the rigid/lossy sections of the diaphragm, or between the lossy part of the diaphragm and the suround?
Well, if your motor is generating harmonics, your cone is gonna radiate harmonics. But a little intrinsic damping at least keeps the cone from magnifying harmonic spectra.
I was referring to the transition on the cone. I've thought a bit about using a sawtooth cone/surround transition, but I'm not yet convinced it would behave as I want. But it might be worth experimenting.
A little shear wave info...
I just remembered that a friend of mine has a little info about shear-wave cone design on his website. Not much, but maybe it'll get ya started. 🙂
Basically, the upshot is that a cone operating on the shear-wave model transmits shear waves across its surface at a velocity determined only by its physical properties; not related to frequency--a weakness of common cones operating on the bending-wave principle. When the shear-wave velocity is known, a cone angle can be specified that allows the forward component of the shear wave to be right at the speed of sound, which allows high frequency acoustic waves to be launched coherently with good dispersion.
I just remembered that a friend of mine has a little info about shear-wave cone design on his website. Not much, but maybe it'll get ya started. 🙂
Basically, the upshot is that a cone operating on the shear-wave model transmits shear waves across its surface at a velocity determined only by its physical properties; not related to frequency--a weakness of common cones operating on the bending-wave principle. When the shear-wave velocity is known, a cone angle can be specified that allows the forward component of the shear wave to be right at the speed of sound, which allows high frequency acoustic waves to be launched coherently with good dispersion.
Regarding the voltage vs. current drive discussion earlier in this thread, I would like some comments on power drive (i.e. voltage controlled power source, with product of output voltage and current proportional to the input voltage). The only thing I found in my search was a mention in one of the tubecad journals:
http://www.tubecad.com/email_2001/e0829/e0829.pdf (page 4)
Can this mixed feedback be achieved without the extra active devices, or is there a better way?
http://www.tubecad.com/email_2001/e0829/e0829.pdf (page 4)
Can this mixed feedback be achieved without the extra active devices, or is there a better way?
I'm going to modify my question - surely it is possible to do so, but are there any significant drawbacks?
Off topic - did you know the Atma-Sphere MA-3 tube amp delivers 475W/ch and has an output impedance of about 0.7 ohms?
Prune,
I'm not a resident amp expert, and perhaps I misunderstand your question, but this ESP article discusses mixed feedback without any additional active devices.
454,
In the right alignment, I don't see any intrinsic downsides to a transducer built for current drive with a Qms of 0.5 ...if that was your question.
I'm not a resident amp expert, and perhaps I misunderstand your question, but this ESP article discusses mixed feedback without any additional active devices.
454,
In the right alignment, I don't see any intrinsic downsides to a transducer built for current drive with a Qms of 0.5 ...if that was your question.
In the context of this discussion, I just got wondering about OTLs. IIRC, OTL designers usually resort to using negative feedback to bring Zout down to what they consider acceptable levels.
If we're no longer concerned with amp damping, does that open up new possibilities for building less compromised OTLs? For example: less/no feedback, fewer output devices, higher plate-impedance tubes with other desirable characteristics?
Comments from the amp experts?
If we're no longer concerned with amp damping, does that open up new possibilities for building less compromised OTLs? For example: less/no feedback, fewer output devices, higher plate-impedance tubes with other desirable characteristics?
Comments from the amp experts?
My guess would be that feedback around an OTL would
bring it down to the 10-50 ohm source impedances that
I have found useful with the various high-efficiency full-range
drivers. The higher values for the sealed box cases, and lower
values for rear loaded cases. 😎
We are still concerned with amp damping, it's just that we
recognize (and have since the middle ages) that there are
often optimal finite values, and that infinity is not necessarily
the best damping figure.
bring it down to the 10-50 ohm source impedances that
I have found useful with the various high-efficiency full-range
drivers. The higher values for the sealed box cases, and lower
values for rear loaded cases. 😎
We are still concerned with amp damping, it's just that we
recognize (and have since the middle ages) that there are
often optimal finite values, and that infinity is not necessarily
the best damping figure.
Whether you put a driver in a pipe, a baffle, a box, or a horn, it's the total system Q that determines the low-end response, and most alignments will contribute to system Q.
So if you have a driver/amp system with a Q of O.5, you better put it in an alignment that doesn't raise the Q much, or you'll be underdamped. IOW, your limited to OB, U-baffle, a very large box, or some type of largely aperiodic loading.
That's why I like the idea of an adjustable-Qms driver--less limiting.
So if you have a driver/amp system with a Q of O.5, you better put it in an alignment that doesn't raise the Q much, or you'll be underdamped. IOW, your limited to OB, U-baffle, a very large box, or some type of largely aperiodic loading.
That's why I like the idea of an adjustable-Qms driver--less limiting.