Hi kgrlee,
thank you for the reply.
Here is my crude, experimental set-up, almost finished.
Kindest regards,
M
thank you for the reply.
Could you please explain how you use the Q-factors to arrive at your conclusion?
Thank you for the confirmation. Unlike you, I really like it. I think the female voices renditions sound amazing.
Here is my crude, experimental set-up, almost finished.
Kindest regards,
M
I didn't deduce this from the Q factors. I did it by looking at the Impedance.
If there is a single 'bass' resonance, (here about 500Hz) the height of the peak tells you the relative importance of Mechanical (or Aooustical) Damping and Electromagnetic Damping.
- If there is no peak (eg no magnet ), there is no Electromagnetic Damping.
- If the (single) peak is 2 x Rdc, Mechanical & Electromechanical Damping are equal.
- As the peak is much less than 2 x Rdc, the Mechanical Damping dominates on this AMT
This is how various programmes measure Thiele-Small parameters.
The 'peak' is quite narrow, so the Electrodynamic Damping is pretty small.
The Qs are another way of saying the same thing. Good MC units have low Qes and high Qms; ie high Electomagnetic and low Mechanical damping. This AMT is opposite. Hence the Qs confirm my conclusions
This is a HUGE subject and I could go on forever. eg I haven't even started on the double peak and other wriggles.
I was involved in developing the early theory of bass response in the 70s. We didn't use TS parameters cos both Thiele & Small were academics and the TS stuff isn't that useful for production & QC. Can't remember how MLSSA dreamt up TS stuff.
In the instruction manuals for MLLSA and other measuring devices/software, will be references to AES papers on measuring TS parameters. But I haven't dun dis hard sums 4 at list 4 decades and my single remaining brain cell hurts.
With my historian's hat on for a moment, Richard Small built on & extended the work of his colleague Neville Thiele, taking & extending mathematical simplifying assumptions based on electrical filter theory & applying them to drive units & particularly drive units in enclosures. Thiele in turn built on similar work done in industry by particularly James Novak at Jensen, with broader work also done by Leo Berenak and also Ted Jordan at Goodmans. They in turn were heavily extending & developing the theoretical foundations laid by Albert Thuras in the 1930s (also derived from contemporary electrical filter & mechanical engineering theory).
T/S parameters, taking it as a loose handle for fundamental, large & small signal values, were never intended for use as a commercial production QC method -they were / are largely about providing a simplified mathematical model of a driver's behavior at resonance & up to the mass-corner frequency under a given set of drive conditions, and allowing you to hit where possible a desired LF alignment in the smallest possible Vb.
T/S parameters, taking it as a loose handle for fundamental, large & small signal values, were never intended for use as a commercial production QC method -they were / are largely about providing a simplified mathematical model of a driver's behavior at resonance & up to the mass-corner frequency under a given set of drive conditions, and allowing you to hit where possible a desired LF alignment in the smallest possible Vb.
Last edited:
But if the membrane is very light, the air load would likely be a rare case of mechanical loading that is quite linear. Increase the 'cone' mass, and hey presto, the Q's all go up.
But there will be various modes and reflections just like with a soft cone. Unless some unusual radial mesh is used, the cloth would also be anisotropic from the grid alignment of the fibres.
Thin non-magnetic steel wire, like a guitar string (but not magnetic), could be tensioned to give the required spring constant, and it would have extremely low mass and energy storage. An alternative that comes to mind could be to use a laser cut stencil, but adjusting the tension might be a bit more complicated.
In any case, under a current drive scheme, reflections that return to the voice coil, whether from the cone or from the spider, would only produce voltage but no current.
OTOH, in spite of lower distortion under 'pure' current drive, I can imagine scenarios where a small amount of electrical damping could be useful to fine-tune the energy absorption of standing waves. An example could be a prominent throat resonance in a horn. Under voltage drive, the amplifier detects spurious voltages and 'stiffens' the cone, causing a series of comb resonances. Under current drive, the cone yields and the boundary is now at the back of the compression chamber. And somewhere in-between the 2 modes cancel.
Please note that the non-linear control techniques I mentioned are used to impart a linear transfer characteristic to the resulting amplifier. Strictly speaking, even linear systems behave non-linearly whenever the error is large.
No, I do not have any amplifier designs on the internet. What kind of amplifiers are you looking for ? Are you willing to try simulating those in SPICE?
Hi kgrlee,
thank you for your reply, even I have the (false) impression that I understand what you are trying to communicate.
Kindest regards,
M
thank you for your reply, even I have the (false) impression that I understand what you are trying to communicate.
Sorry to beat the half-dead horse, but my original question was whether you would consider the flat impedance to be due to the moving mass being dominated by the air load. It is my - perhaps incorrect understanding that the high mechanical damping may be not only due to the air load but also construction, e.g., shape and material of the membrane.
Kindest regards,
M
My work was based on another guy(s?) whose name I can't remember. Peter Fryer told me that Novak probably had the correct maths and I spent 2 days proving that his sums were the same as mine .. a Herculiean task as I neber wen 2 skul and kunt reed en rite
Beranek, Jordan & Thuras were red herrings. I wasted a LOT of time trying to make sense of Jordan as he seemed to be on the right track.
Too right ! Today, stuff like Klippel use supa dupa computing power to derive the important parameters from TS. Alas, accurate TS doesn't always translate to accurate important parameters. Much better to measure these directly.
Your question is very pertinent.
All I can be sure about is that Mechanical stuff (which includes Acoustic Stuff) dominates the Electrical stuff. It is probable the Acoustical stuff dominates the Mechanical stuff but to confirm this, you have to consider the factors you mention. It's been 40+ yrs since I've looked at a Heil.
It's much easier with an Electrostatic cos the construction is simpler (??)
I want a simple general purpose amp that is stable and has Zo > 2k from 20 - 20k Hz. In a different thread, I've established it can probably be done with something combining Mills & Hawksford with Vanderkooy's simple amp. But I have to get off my beach bum *** to re-learn LTspice as I've lost a lot of my 21st (and some 20the) century stuff in HD crashes
Last edited:
Hi kgrlee,
I am rather interested in this; I had built a current drive before (based on feedback taken from a resistor in series with the driver), and would like to experiment with the Heil, especially now that you concluded that it is suitable for such.
Kindest regards,
M
How did you established the requirement for the Zo?
I am rather interested in this; I had built a current drive before (based on feedback taken from a resistor in series with the driver), and would like to experiment with the Heil, especially now that you concluded that it is suitable for such.
Kindest regards,
M
I don't know how you wanted to combine those amplifiers but in my opinion, the easiest way to increase output impedance is output current control. That is, the closed-loop output impedance becomes the open-loop output impedance times the loop-gain. However, in order to maintain the output impedance at 20kHz, the loop gain should be maintained similarly, which is possibly where the stability restrictions come into play. A non-linear control technique may not have the same restrictions for example, Bruno Putzeys' UcD amplifier that uses sliding-mode control.
For anyone interested in how the theory behind all this goes, there is the following:
https://eng.libretexts.org/Bookshelves/Electrical_Engineering/Electronics/Operational_Amplifiers:_Theory_and_Practice_(Roberge)/02:_Properties_and_Modeling_of_Feedback_Systems/2.05:_EFFECTS_OF_FEEDBACK_ON_INPUT_AND_OUTPUT_IMPEDANCE
Last edited:
OK >
Perhaps I should have written 'quite a bit', because all you need to do is read all the posts to see the misunderstanding.
There are even people that think simply adding series resistors to your amp's output (thereby creating a current limited voltage amp)
necessarily creates better sound,[>that is a Furphy.
RE. voice coil temperature > Heating a metal coil actually increases its resistance, thereby creating a degree of self-regulation.
The point I was making was that a driver that can't handle increase of voice coil temperature is not a good high power driver.
Other questions > Your own writing answers them.
Last edited by a moderator:
"Air Coupling" has effectively nothing to do with the flat impedance of drivers like AMT and Magnepan speakers,
it is the way they are designed/built that creates a purely resistive load.
My immediate requirement is for measurements. Zo = 2k would give 0.04dB error if the speaker impedance rose to 100R anywhere .. juu.uust .. accurate enough for TS parameters. If such a beast was simple to DIY, it would be further incentive to design drive units & speakers specifically for Current Drive.
Yes I know. Today with supa dupa computer power, you can just use a 2 port measurement like Clio & Klippel. But it's nice to just do a sweep without duin hard sums wij mek mi hed hurt
As you have chosen your speaker, your amp only needs to be stable and 'current drive' enough for your Heil. Many implementations including your suggested one will do fine.
Hi,
Have you experimented with this, do you have measurements? Electrical damping is due the driver damping itself over low impedance between the terminals, iow low impedance circuit, but only around the main resonance where current from backEMF is opposite phase opposing the movement. Higher up, the cone or dome would always "yield" as electrical damping doesn't happen because the current is perpendicular to movement and has no effect, or has magnifying effect if it's beyond 90deg due to Le. See https://www.edn.com/loudspeaker-operation-the-superiority-of-current-drive-over-voltage-drive/
Hi, not a Furphy, see Norton and Thevening equivalent and how to convert between the two. This is usually found in circuit analytis part 1.
While it is true a series resistor doesn't turn voltage amplifier to current amplifier, it turns the circuit equivalent which is all that matters, has same effect in acoustic domain. The amplifier doesn't matter (almost*) at all only the circuit impedance that is in series with driver matters, which means how much the driver can affect circuit current, iow acoustic output. You could have ideal voltage amplifier or ideal current amplifier, or anything between from real world, and acoustic measurements would show change in frequency response and distortion accordingly, which are both due to same reason: either the driver dominates circuit impedance, or there is some impedance in series affecting as well, perhaps dominating. This would be amplifier output impedance and any passive network there might be including cabling. Use circuit analysis to find it out, and manipulate for best acoustic performance in given situation.
*) while amplifier can have impact on sound they usually are orders of magnitude better behaving than drivers, just use good amplifier that is in good working order.
Last edited:
Perhaps, but the thing is that I understand that a constant current source can be constructed simply from a high voltage source and a high resistance. That is how the 'tail' CCS has traditionally been made since the early days of tube-based long tail pairs. In the case of an amplifier's output impedance, the main difference between 10 ohm, 100 ohm, or 1000 ohm is an arbitrary line in the sand where someone decides the effort is no longer worth the diminishing returns. So, if going from 0 to 10 ohms is already a furphy...
What self regulation? A current amplifier emulating 1kV behind 1k ohm, will supply a steady 1A whether the load is a cool 8 ohms or heated up to 16 ohms.
I think @abstract answered that above.
Yes, but not in the audio context. You may try adapting the UcD technique to control the output current (vs voltage).
http://amorgignitamorem.nl/Audio/HD702/UcD/ucd_aes.pdf
There's an old thread by @Eva adding a current limit to the UcD that some of the older members might recollect having read.
https://www.diyaudio.com/community/threads/self-oscillating-fun.103826/