John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
I hear you loud and clear: MiiB. And I agree with you. I am impressed that you work with 'quality' loudspeaker design. It is an art, as well as a scientific discipline. Electronic quality is much more subtle, but it IS there, to find and use.
As far as the Edison cylinder is concerned: It MAY actually have some subjective quality that is not in today's recordings, for the most part. AND it might be because it is LESS PROCESSED than modern recordings.
About 40 years ago, I met Dr. Peter Craven and his associate, Michael Gerzon, and was invited to Peter's house to discuss and listen to hi fi reproduction. They had very simple electronics and a single Quad electrostatic speaker with a Shure cartridge that had interchangable stylii, for both 33 and 78 rpm records. They played 3 selections of the same classical performance: One was from the 1950's and sounded 'OK'. The next was a 78 and it sounded BETTER! And the third was a Dolbyised version of the same performance on 33.. It sounded the worst of all. So much for progress in that direction. Now, digital has been measuring better and sounding about as bad as before. Perhaps the right questions have been suppressed and 'specs' rule the engineer's world.
 
MIB,
We are in agreement that ceramic magnets don't belong in a high quality speaker , no argument on that fact. Alnico or Neo are much better materials for many reasons in a speaker. Now if you had significantly reduced the diameter of the motor assembly to reduce the reflective surface behind the cone I could go along with what you have done. that is not the case. The front plate is still there, not a very elegant solution.
 
Kind , there's is no front plate on out top of the line speakers, There we have a vertical magnet system, I believed you were referring to that. with neo, regardless if it's a ring or separate round elements you can reduce the top plate area quite a lot, but with conventional magnet structures you still need some area in order to create the right flux in the gap.

Steve- speakers are dynamically flawed by nature. take just the varying position of the VC in gap creates a varying inductance. Surrounds, spiders and box creates nonlinear springs, all contributes to dynamically flawed speakers.
 
Last edited:
MIIB,
I only took a quick look at your website so don't know all the things you have done, that would be a nice solution to remove a front plate but obviously we do need to create a gap for the voicecoil to operate in. As I said earlier your speakers may sound excellent, I can't make any comment on the sound. Just looking at things that I see that are more for marketing and appearance than pure science. There is way to much of that in speaker construction.
 
IIRC the 68040 FPU had 8 - 80 bit FP CPU registers. Apple used to have a demo where Mathematica could invert a nearly singular matrix on a MAC but not on Intel. This was circa 1989

Probably the other way around. The 8087 already had 64 & 80 bit floating point and the
68881 had 64. That was a long time before the IEEE standard.
OTOH that does not preclude rigging a demo. Just play with compiler options.

BTW I once worked for a company that made Unix computers based on 68K
and later on the Fairchild Clipper. Once we got complaints that the C compiler
would emit asm files that the assembler could not translate.
BIG HEAD SCRATCHING!
It turned out that the Motorola SysV C compiler still emitted floating point
instructions for the VAX. :) But Moto got it working, eventually.

Gerhard
(currently ripping Virgil Fox, G.Gould & Erik Sati from CD to USB stick for car use)
 
vacuphile,
I'll leave metal diaphragms to both compression drivers and dome tweeters with Be. As far as a metal cone for a dynamic driver I will pass unless that driver is only used in its pistonic range, above that range the high Q breakup modes of those cones are not something nice to deal with. I appreciate anyone working on cone materials, it is a very complex situation to deal with. Paper, or what we call paper has been a very useful material for most speakers and is still the norm. I don't use paper myself and understand the many issues that have to be addressed working in that area. Get the diaphragm material right and you are very far along to making a great sounding speaker. All the littlest details change the sound of a speaker, there are many factors to balance.
 
Just for reference. The thermal noise level of a microphone is around -135 dBm. 32 bits above that is + 60 dBm or 1,000 watts.

Now in acoustic power -8 dB is about the lowest level than can be heard. So 187 dB above that is just a bit under complete atmospheric modulation.

Can't wait till someone tries for 48 bit encoding. That upper limit would require destroying the solar system. :)

Wasn't there a band in the "Hitchhiker's Guide to the Galaxy"
that was best listened to from a high orbit around the planet?

B4 that there will be thermal runaway. When the microphone gets hot to the melting point,
thermal noise will increase even more. Ouch!
 
Last edited:
Ok the cone material is not metal, in fact the aluminum is the soft part of the cone, the cones we use a ceramic (aluminum-Oxide) made in a layered plasma process, app half the material is Hard ceramic the rest is soft aluminum, This makes for s very stiff, but also very damped membrane (sound velocity propagation is very different in two materials).

The fundamental resonances are just a bump (in speaker terms) of 3 dB, a 4" membrane has a resonance at app 12.5KHz.
 
MIIB,
You are using a plasma process to create that material, that is beyond the norm with metal cones. Most use simple anodizing of the aluminum, which you would get by simply allowing the aluminum to oxidize anyway. So I assume you have a thicker layer of oxide material to your cones. I agree that two materials combined would reduce the high Q flexural modes in a cone as long as the propagation speeds are different enough. I have done a very similar development with composite materials, just not willing to give up the material properties or chemistry involved. Even the chemist doesn't know what I am using the material for or the final combination of chemistry.

Me and my private notebook are the only ones with that information.
 
Steve- speakers are dynamically flawed by nature. take just the varying position of the VC in gap creates a varying inductance. Surrounds, spiders and box creates nonlinear springs, all contributes to dynamically flawed speakers.

And those flaws can be greatly diminished when you're not designing for "Wife Acceptance Factor" and convenience (narrow vertical towers). And higher efficiency can go a long way as well, reducing thermal compression. Done right, I think speakers can be very dynamic.

se
 
Disabled Account
Joined 2012
------ Now, digital has been measuring better and sounding about as bad as before. Perhaps the right questions have been suppressed and 'specs' rule the engineer's world.

I agree with this. esp with ADC/DAC. By the numbers we are told everything is fine. But it isnt. And, I have pointed to where at least one spec is the problem and it doesnt go far enough in defining the real dynamic range in an app, IMO. The true dynamic range is much less and audible. There is probably enough 32/64 DSP and/or CPU power right now that with high speed converters, enough dynamic range can be had without audible consequences intruding as much as it does now. The technology is there today but, according to some tea leaf readers here, not the will to try it. No guts no glory.

Oh well, now onward and downward (bit-wise).

John, Where does the headphone play in all this.... with quality products in abundance, we can't leave them out for quality of sound..... they eliminate the pesky room interface issues and IMO can hear the analog and digital artifacts and lack of clarity below 0FS more easily than ever before.

THx-RNMarsh
 
Last edited:
Disabled Account
Joined 2004
Probably the other way around. The 8087 already had 64 & 80 bit floating point and the
68881 had 64. That was a long time before the IEEE standard.
OTOH that does not preclude rigging a demo. Just play with compiler options.

I checked my old Inside MAC it must have used the co-processor to do 80 bit FP. I used to program inline assembler on Codewarrior because FP variables declared register were ignored with all settings or optimizations. I didn't really care about the 80 bits per se I just wanted to do complex multiplies in place.
 
IIRC the 68040 FPU had 8 - 80 bit FP CPU registers. Apple used to have a demo where Mathematica could invert a nearly singular matrix on a MAC but not on Intel. This was circa 1989

The original IBM PC, when equipped with an 8087 FPU, also had 8 80-bit FP registers (effectively 81 bits unless denormal flag set) and could do extended precision fp math in hardware (sort of). Bit of a pain to program because tjose registers were a stack and you had to keep track of the, and results were always written to register 0.
 
Kind, app 1/3 of the thickens is Oxide, the process takes 60 hours, liquid High voltage and High Current. The process cost ten time the price of the cone. For even better performance we use an extra 10 um layer of Diamond.

Steve high efficiency and bigger size has other issues, noise is one of them, membrane break up and is an other, there's No free lunch. and those speakers are also dynamically flawed in just the same way as smaller speakers are.
 
Disabled Account
Joined 2004
The original IBM PC, when equipped with an 8087 FPU, also had 8 80-bit FP registers (effectively 81 bits unless denormal flag set) and could do extended precision fp math in hardware (sort of). Bit of a pain to program because tjose registers were a stack and you had to keep track of the, and results were always written to register 0.

I believe you guys, but I still distinctly remember the demo. Maybe as you say many folks didn't bother because of the difficulty using them. I certainly doesn't matter now. Remember the Motorola byte order issues?
 
Status
Not open for further replies.