Hi John (JC),
when talking to JN, you don't even know what you don't know. Outclassed is one way to put it.
We are lucky to have people of this caliber here to contribute. If you want to really improve your product, you ought to be listening very carefully to what John (JN) has to say. The same thing can be said for Scott or a few others. I guess there is simply no way to show you this. I know I`m an idiot compared to some of our members. It doesn't bother me, I`m here to learn and these folks are doing the real teaching.
What I don`t understand is how or why you take issue with this. It is so obvious, how can you not see it
-Chris
when talking to JN, you don't even know what you don't know. Outclassed is one way to put it.
We are lucky to have people of this caliber here to contribute. If you want to really improve your product, you ought to be listening very carefully to what John (JN) has to say. The same thing can be said for Scott or a few others. I guess there is simply no way to show you this. I know I`m an idiot compared to some of our members. It doesn't bother me, I`m here to learn and these folks are doing the real teaching.
What I don`t understand is how or why you take issue with this. It is so obvious, how can you not see it
-Chris
I've learned a lot form you both, JN and JC; A HOLE LOT! Thank you both. I hope this can get smoothed over and we don't loose you guys like we lost Brad Wood when he got pissed and left.
I be outclassed! OK, JN knows E/M better than me.
By the way, MarkW4, I have a Stanford Research SR-1, you might look it up to see what I have available in my lab, as well as a QuanTech 2173 for low noise measurement. Of course, I have everything else, more or less. That is why I have been able to reach the equivalent noise of a 10 ohm resistor for the last 50 years. How about you?
Could I have better test equipment? Richard Marsh does, and Ed Simon, etc, but I will let them measure the hard stuff, and I do just fine.
The problem with working for a giant lab, like Richard Marsh once did, or even a medium level level lab like AMPEX, is that you have to work on the stuff that THEY want you to work on, not what you want to work on, so audio becomes an outside hobby. It wasn't enough for me, I wanted it all!
By the way, MarkW4, I have a Stanford Research SR-1, you might look it up to see what I have available in my lab, as well as a QuanTech 2173 for low noise measurement. Of course, I have everything else, more or less. That is why I have been able to reach the equivalent noise of a 10 ohm resistor for the last 50 years. How about you?
Could I have better test equipment? Richard Marsh does, and Ed Simon, etc, but I will let them measure the hard stuff, and I do just fine.
The problem with working for a giant lab, like Richard Marsh once did, or even a medium level level lab like AMPEX, is that you have to work on the stuff that THEY want you to work on, not what you want to work on, so audio becomes an outside hobby. It wasn't enough for me, I wanted it all!
If you have a suitable current probe its pretty simple to measure the current from the amp and the current going into the voice coils. The difference is going into the network (unless we have a new math for electrons). it would be pretty simple with an amp with low source Z to make the same measurement with and without the network and see the amplitude and phase relationship of the current for a given voltage across the driver.
Having dealt with a similar issue (power factor correction) what happens when you make the source Z lower the distortion from a non-linear load moves from voltage to current but its still there.
In my experience parallel networks just suck power from the amp. There was a generation of the Watt-puppy that took that to an extreme with a 1 Ohm impedance in the upper midrange. I'm sure it was very sensitive to amps and cables. And a hard way to deal with an HP peak. Flattening the impedance variations will make a speaker less sensitive to amp/cable changes. I still don't see how it could remove nonlinearities from the motion of the voice coil (or the magnetic circuits).
JN's feedback idea is really interesting. The feedback coil should be a finer wire to keep the gap small in a production implementation (one more patent idea tossed to the side). Unfortunately above probably 500 Hz with even the best drivers the nonlinearities in the drive are only part of the problem. Cones have a mind of their own. The one virtue is that the mass and the viscosity of air tend to reduce any higher order motions in cones/domes etc.
FWIW my General Radio Omnical microphone calibrator uses current drive for its transducer. They use it so the output doesn't change as the driver warms up.
Sorry, if I missed some of your test equipment, John. I wasn't trying to dis you as much as I was wanted JN to see that he and you are in very different situations.
Regarding myself, I prefer to remain anonymous. So, I will leave it at that.
I do need to defend JC on the issue of resources. If you really understand what you are doing the tools can be limited and still allow for major progress. Harrison did not build his clocks using the best references available, just a really good understanding of the problem and a way to address it. The tools keep you honest. You need to know what to do with them.
However high energy physics is not something you can do on a shoestring.
However high energy physics is not something you can do on a shoestring.
One nice improvement to have would be a high quality line level preamp that improves digital sources. Now that you have a reference system that do sound good, perhaps you can design something using regular off the shelf parts your overseas contractors can build. And maybe you can share some bits for us members doing diy. 🙂
And a side note... We use I for steady state and i for time varying current. Same for V and v. So why do we use R and Z?
Secondly anyone who has played a bit realized that LR&C do have the companions of lr&c. But the difference is that they are often varying in value from other than the v or i. The extreme example would be in designing a proximity fuze.
Taking this a bit further I x R = V and i x Z = v. As Z needs not be constant then i would not be either. But that is not distortion of i.
Secondly anyone who has played a bit realized that LR&C do have the companions of lr&c. But the difference is that they are often varying in value from other than the v or i. The extreme example would be in designing a proximity fuze.
Taking this a bit further I x R = V and i x Z = v. As Z needs not be constant then i would not be either. But that is not distortion of i.
Last edited:
Joe, it seems you just forget that speakers are not pure electric devices, they are mechanical devices with inertia and various resonances.
You DO need to dump them as much as possible. So, you need very low Z output amplifiers. A lot of things have been tried, dumping those resonances with ferrofluid, as an example, or with double coils or accelerometers in servo systems. But all the attempts to drive a speaker in current give poor results or the benefit too low in regard to complexity in my second example.
Speakers are what they are, and you cannot change their behaviors as long as you continue to use them as they are.
That the reason why, on my side, I just help the amplifier giving to him a flat impedance load, and the other benefit is that simplifies the passive filters design. Too, it helps to minimize the effects on the response curve of any resistance (cables) in serial. (Not a good thing for magical cables sellers ;-)
Jneutron, clever, as usual, I read-you silently with a great interest.
It's difficult, but we have to try. The very first chapter of the module I tought at the University of Geneva is mostly about lateral thinking. Sometimes they get it, mostly they don't. But at least I tried, and maybe in time it'll become something useful.
I admit having problems, too, to decipher what you actually want to say, but with regard to above snippet it seems you wanted to say what I've been saying as well, several times on this forums:
For any given driver most likely neither pure voltage drive nor pure current drive is optimal. Rather there is a drive impedance "profile" vs. frequency that optimizes the bevavior (this is application specific and as well a matter of preferences, of course). High-Z typically is good for the region above resonance, while at resonance we usually want to have some electrical damping, I personally strive for close to aperiodic for open-baffle or closed box. This optimizes recovery from errors (overexcursion etc). Any distortion signal (motor distortion) and external exitation ("knuckle testing the cone", SPL from adjacent drivers) will undergo the damping set up by the source impedance therefore we will want to avoid high-Q's because otherwise errors signals would cause ringing in their settling. It doesn't matter to which SPL response the whole thing is EQ'd.
It is very important to remember a basic fact about any speaker driver:
When the measured terminal voltage (or current, for that matter) is EQ'd precisly to a given target the SPL response will always remain the same, not matter what the source impedance was. That's obvious one might say (we didn't change the driver itself) but we really got to be aware of what that actually means.
[sidenote]And as for driving a passive XO with current drive, after establishing a flat, ohmic impedance this is true as well, same terminal voltage or current (XO input) gives identical SPL output curve.[/sidenote]
The difference, and the optimization we can actually arrive at, though, is in the "fine print": the general distortion characteristic in its various aspects and the details of any overload/error recovery are dependant on drive impedance characteristic. There are drivers where this dependency is so small that we just take our voltage amps and we're set, yet other drivers might be improved significantly.
So we might just shape an amp's output impedance characteristic to whatever found to be ideal and beneficial for that driver, then EQ the whole SPL transfer function to get a reasonable curve not to much different in character from a standard voltage curve. For example we would at least need to compensate the peak at resonance when we choose to run the driver underdamped around resonance, Q>=3 or so.... so setting up the active XO and overall EQ doesn't have to deal with the bump, makes life easier. Not a must, but nice to have.
Last edited:
Some programs are better than others, for sure.
I think there is a large innate component, too. I've worked with some people that have a lot of real world experience, but it still hasn't made them useful.
This is the clip I used to illustrate the science vs intuition dilemma. The "intuition" guy is the real Pelé, the soccer legend.
I was thinking about this. Don't most people use square wire these days for voice coils? I think this technique is worth investigating, but first we have to find someone who can make these coils accurately enough without messing up the raw speaker performance. JN has the budget to hog everything from billet unobtanium which we don't.
It's got to be worth messing around with a bit?
Now onto the dumb question. Once we can more accurately measure force and position can we actually correct cone/dome problems or are we rapidly turd polishing? I was musing that, if we can at least make sure the voice coil is where we want it, then we can consider making cones stiffer as we don't lose as much with the added mass due to the corrections.
Let me deviate a bit from my 'normal' behavior (that is 'be silent and watch').
Yes there is (at least to my opinion)
Generated-power / Motor-power = 100% equals zero impedance.
Generated-power / Motor-power = 0% equals (extreme) high impedance.
It may only be a difference of terminology and/or formulation.
...[/QUOTE]
Is it really needed to discuses like this?
I'm not!
And how do you know that? Your are (I'm wondering) working for the NSA and …
An so you will go on, it really stats to look like trolling.
Frans.
Yes there is (at least to my opinion)
Generated-power / Motor-power = 100% equals zero impedance.
Generated-power / Motor-power = 0% equals (extreme) high impedance.
It may only be a difference of terminology and/or formulation.
...[/QUOTE]
Is it really needed to discuses like this?
I'm not!
And how do you know that? Your are (I'm wondering) working for the NSA and …
An so you will go on, it really stats to look like trolling.
Frans.
And you decide not to show this important bit of supporting measurement?
Yes Joe, why the rush? I understand the advantage of less cable impedance sensitivity that Tournesol mentioned, but you have not shown enough for me to infer sonic advantages to gain from your "improvement" compared to "conventional" design. Another way to put it Joe, do you have something you can show on the aspects of sonic presentation found to be improved by your method compared to "conventional" method? I respect your discretion and waiting for your team to reach an agreement is not a problem. However, posting more opinions and premises that are not in consensus/agreement to be right will generate more confusion and negative comments. 🙂
I see a crossover schematic, drivers but no enclosure description. Interested people can independently verify your findings in part. However you should not blame them if they independently generate measurements not in your favor.
Well, in that case you are most probably the most academically gifted individual I have ever spoken to, and I really mean that.
JN, please do not take what I say here as anything other than encouragement, because that is what we all need - encouragement.
As I have been trying my best to follow this thread over the past few weeks, there are a few things I would like to make clear about myself. Firstly, I am someone who has an innate disability with mathematics - much to the dismay of my late father who was a shop floor master machinist in the aerospace industry. So, at an early age, I was made painfully aware of my own limitations, and have subsequently made a lifelong commitment to overcoming them. I reckon I have come a long way. I have worked in medicine, trained to degree level at art school, become a trained and qualified teacher in adult education, become an experienced photographer with an encyclopaedic knowledge of the medium, self taught in computer graphics, and have a good understanding of materials, especially sculptural aspects of metal, wood and paper. And I love music and sound. What may not be apparent is my capacity for critical thinking. I am one of a rare breed who is simultaneously an artist and a scientist. This is what I do, I live and breath it. I don't have a laboratory to work in, but I do like to think that everything I touch and engage with my senses is what feeds into what I call the 'imaginarium' between my ears. It is my gift to the world, for which I am truly grateful.
One thing I am certainly not is a troll, or a so called 'creative' type. I have very good and precise understanding of language. Some of the insults I have had thrown at me on this thread are so shameful, that I have actually thought of dropping DIYaudio altogether. What has been worse is seeing highly intelligent people turning on each other like animals tearing lumps out of each other. This debased level of Internet disinhibition is not what the world needs right now. Do you all understand that there are thousands of people out there reading the discourse between you? That you are actors on a world stage? That words have meaning and therefore are of consequence?
While I am writing this, my wife is in surgery being operated upon, completely surrounded by an ocean of kindness. Such a contrast.
I would like all of you to stop squabbling and start being nice to each other.
ToS
- Status
- Not open for further replies.