"1 uSec is not human capability, but an arbitrary factor of ten better setpoint I wish to work with..my point being, your setup, I suspect, is incapable of resolution at that level.."
Of course it is. The limits are timing jitter (order of ns), amplitude jitter (order of -85dB) and nonlinear distortion (order of -100dB) - the respective importance of each of those depending on what exactly one is trying to measure, how, at what frequency. Absolutely no idea what the requirements of MRI superconducting wotsits would be as I've never worked in that area of electronics, but this is audio we're talking about - few tens of V/us, not much over 1A/us (actually with real music considerably less than 1A/us).
We are speaking apples and oranges...I am quite sure you are correct in your system's capabilities...I am not arguing that.
What I am saying is that you are not measuring the voltage correctly because of the high slew rate currents..ahhh,, guess I'll hafta explain...
Take a 250 watt 8 ohm non inductive resistor..as a load.
the physical inductance minimum this load can present: using terman's equation, about .15 uH. And, this is with zero actual resistor inductance.
Now, your 1 amp/uSec slew rate...the reactive voltage across the load: .15 volts at peak slew. The actual voltage across the load is the amp voltage minus this reactive component. you can't measure it...because the loop you form in connecting to this load, picks up the magnetic field that the loop is generating. The best you can do with this load, is to measure the voltage plus the error component. If you could run the test lead through the middle of the resistor, you will break that loop intercept, and read the EXACT load voltage. Note that this error becomes more and more significant as the impedance of the load goes down...this is because the resistive voltage becomes smaller with resistive drop, while the error component remains the same, and is based only on geometry.
You have exactly two choices when measuring slew rates at these impedances...use a Danfysik style current measuring toroid, which is in itself bandwidth limited, or go with a coaxial resistor where the voltage pickup is through the geometric and magnetic center of the resistor..the latter is exactly what I am making..however, I have experience in both, so can advise on the limitations of each.
Quote:
Originally Posted by RB
No, I can't post pictures of my setup because I haven't got a digital camera.
I can, and will do so here if you request.
Quote:
But of course you can just attach a pair of leads at this sort of frequency. It's not hard to prove - just change the length or termination method or something and watch the measurements stay constant within measurement limits
Actually, I have done so, from 4 Giga amps per second into 1 ohm, 6000 amp half sine into 250 uOhms (here I had to remove the b dot term down to less than 5 millivolts, as I was measuring the temperature of a 3 inch diameter diode during the current pulse), and even at 22 Khz into a .1 ohm resistive load..
What I am explaining to you is exactly what I found in errors, and the methods and designs I had to do in order to eliminate them. Experience in real setups, I am not speaking academically.
Gotta go for now, a tour to guide..
Cheers, John
Note:followup, Jan 14:
You said:
Quote:
"but this is audio we're talking about - few tens of V/us, not much over 1A/us"
Hmm...a few tens of volts per uSEc? Is that a correct number? If so, then, into 8 ohms, 20 volts/uSec is, 2.5 A/uSec...5 amps/uSec into 4 ohms..let's re-think the numbers...
A power load with .15Uh inductance....
.15 uH load inductance will be 375 millivolts error into 8 ohms, 750 millivolts into 4 ohms.
And, look at a zip cable...200nH per foot, 10 feet, 2 uH...at 5 v/uSec, 10 volts? TEN? You're looking at a 90 degree lagging cosine with 10 volt amplitude...one that you can't readily measure, due to e/m field errors, and a sine measurement will say "zero distortion"..and, the speaker needs the current as that is what forces the voice coil...not the voltage.
Look at my double coax, 10 nH per foot, 100 nH ten feet...into 4 ohms, it will react with half a volt..meaning the current will slew at the speaker more like the amp intended..
Hence my use of the term "sloppy" earlier on..it is a rather unkind sounding word, and really is meant to express how present test methods need to be upgraded. One cannot address current slew rate capability if one does not know how to measure it properly.
"Hmm...a few tens of volts per uSEc? Is that a correct number? If so, then, into 8 ohms, 20 volts/uSec is, 2.5 A/uSec...5 amps/uSec into 4 ohms..let's re-think the numbers...
A power load with .15Uh inductance....
.15 uH load inductance will be 375 millivolts error into 8 ohms, 750 millivolts into 4 ohms."
Yup, can't argue with that. But it's still only phase shift and it only follows the usual equations for LCR. Actually, I'm sorry, I was a bit cack-handed with my figures for real world slew rate. Full amplitude 20kHz sine wave from a 200W amplifier is just over 7V/us, or about 2A/us into 4 ohm. That's 4V error in your cable example, against peak amplitude of 56V, about 1/14, which sounds right to me - -3dB point will be about 300kHz. This is a problem why? It's still zero nonlinear distortion and we're just back to arguing the audibility of a few degrees of phase shift at 20kHz. You can measure that phase shift easily enough with or without any knowledge at all of current slew rate.
Yup, can't argue with that. But it's still only phase shift and it only follows the usual equations for LCR. Actually, I'm sorry, I was a bit cack-handed with my figures for real world slew rate. Full amplitude 20kHz sine wave from a 200W amplifier is just over 7V/us, or about 2A/us into 4 ohm. That's 4V error in your cable example, against peak amplitude of 56V, about 1/14, which sounds right to me - -3dB point will be about 300kHz. This is a problem why? It's still zero nonlinear distortion and we're just back to arguing the audibility of a few degrees of phase shift at 20kHz. You can measure that phase shift easily enough with or without any knowledge at all of current slew rate.
Agreed, it follows LCR...never been a problem with that..
A 4 volt cosine error out of 56 gives about 2.5 uSec time shift. Nordmark results are flatline at 1.5 uSec to 12Khz, I don't have the time to re-do for 12Khz, or any lower freq now, gonna go home..I know I can't hear 20K..
I am not worried about phase shift per se...it's the timing errors in the 1.5 to 5 uSec regime that humans are sensitive to that I am speaking of.
Zwislocki et al has 30 uSec capable at 250 hz...(360 degrees/4000uSec)*30 uSec=2.7 degrees phase shift at threshold, while Nordmark has 10 uSec, which is .9 degrees phase shift as threshold. (I'll do up to 12 Khz numbers on Tuesday, I'm outta time...holiday weekend.
300Khz is entirely unimportant, I'm sure we agree..but the time constraints for image localization is certainly in the realm of those numbers.
Enjoy your weekend Robert...it's a pleasure speaking with you. Till we meet again..
Agreed, it follows LCR...never been a problem with that..
A 4 volt cosine error out of 56 gives about 2.5 uSec time shift. Nordmark results are flatline at 1.5 uSec to 12Khz, I don't have the time to re-do for 12Khz, or any lower freq now, gonna go home..I know I can't hear 20K..
I am not worried about phase shift per se...it's the timing errors in the 1.5 to 5 uSec regime that humans are sensitive to that I am speaking of.
Zwislocki et al has 30 uSec capable at 250 hz...(360 degrees/4000uSec)*30 uSec=2.7 degrees phase shift at threshold, while Nordmark has 10 uSec, which is .9 degrees phase shift as threshold. (I'll do up to 12 Khz numbers on Tuesday, I'm outta time...holiday weekend.
300Khz is entirely unimportant, I'm sure we agree..but the time constraints for image localization is certainly in the realm of those numbers.
Enjoy your weekend Robert...it's a pleasure speaking with you. Till we meet again..
Cheers, John
Any evidence yet that these timing errors are an audible artifact for cables?
Any evidence yet that these timing errors are an audible artifact for cables?
None that I am aware of, although I still try the published research. So far, most of what the "believers" explain is just, floob...and everything I read from the "science" end has audibility holes in it..
I have accepted the fact that I will have to do the measurements and correlations myself...the research is quite lacking, both in the human audibility aspect, as well as the electrical test one.
Just finished a big honeydo project, this one took two months..I now have a break in which I can start building stuff again. That window will close July 10, 2005, and won't open again until February 2006. My expectation is to have the electronics and loads built before that break, and start the audibility part May/June 2006.
In the meantime, maybe someone else will have something built to play with..
None that I am aware of, although I still try the published research. So far, most of what the "believers" explain is just, floob...and everything I read from the "science" end has audibility holes in it..
I have accepted the fact that I will have to do the measurements and correlations myself...the research is quite lacking, both in the human audibility aspect, as well as the electrical test one.
Just finished a big honeydo project, this one took two months..I now have a break in which I can start building stuff again. That window will close July 10, 2005, and won't open again until February 2006. My expectation is to have the electronics and loads built before that break, and start the audibility part May/June 2006.
In the meantime, maybe someone else will have something built to play with..
Cheers, John
Who are you involving in your audibility tests? Procedures worked out yet?
Who are you involving in your audibility tests? Procedures worked out yet?
Three people, me, myself, and I...unless you count all the multiple personalities, or the voices inside my head...
Procedures are pretty straightforward, stimulus is still an issue, choices are pure freq's, like a (sinx)/x , or a cowbell type, lots of fundamental, good transient, well damped to provide reasonably fast decay. Long decay is an issue when not in an anechoic environment..
Once actual data runs have been performed, with analysis of the data, I'll review the stuff with a high level researcher in this field..There's one about 30 miles away, nice guy, who I've conferred with..if deemed valid, then a real sample size thingy with controls could be done..
But, for my purposes, I'l have a decent set of numbers for at least one (arguably) human..and can set electrical limits about ten times what I can hear.
None that I am aware of, although I still try the published research. So far, most of what the "believers" explain is just, floob...and everything I read from the "science" end has audibility holes in it..
I have accepted the fact that I will have to do the measurements and correlations myself...the research is quite lacking, both in the human audibility aspect, as well as the electrical test one.
Try running your ideas and findings about cable time shift by the folks discussing cable at Dan Lavry's pro recording forum -- Lavry , Bob Katz, a few others, have some serious technical knowledge to share -- no floob allowed!
Once actual data runs have been performed, with analysis of the data, I'll review the stuff with a high level researcher in this field
Is this gonna happen in our lifetime?
John the parameter limitations of human hearing with respect to frequency response, phase, directionality, spatialization, etc has already been mapped, written about and published in many scientfic journals. A few sources have already been presented such as Everest, AES papers, Harman Research, NRC, etc. There really is no magic here.
Seems like alot of effort on your part to measure the hair on a gnats a**
__________________
Gene DellaSala
President, Audioholics Pursuing the truth in audio & video...
Try running your ideas and findings about cable time shift by the folks discussing cable at Dan Lavry's pro recording forum -- Lavry , Bob Katz, a few others, have some serious technical knowledge to share -- no floob allowed!
I discontinued association with that forum..I found that they were not at the level required to add to the issue. I had hoped otherwise, spent a good half year there..to no avail..
They do have extremely good knowledge, but it is only useful for incremental advances..or solving known issues..not for charting new waters..
Linear Mode
