Ground Balance Theory

[clancy]

Aziz,




G(t) = a*(t+p)^b where p is preamp+coil group delay (like “phase delay”), cannot be correct for general PI Tx.




Imagine -20V applied Tx coil for just 50microseconds, followed by back-emf period, i.e. short magnetic ramp with sudden switch off. Obviously shape of VRM decay must be function of this short magnetic ramp length. But logically same must be true for all PI systems with magnetic ramps, even long period ramps, VRM signal must be function of ramp length. And logically, long length ramps must have different VRM signal shape to short ones (not just size difference, your “a” value). And also then logically VRM decay signal for PI systems with current ramp droop from the Tx coil TC must be function of ramp length and Tx TC droop.




For PI of Eric, he have Tx coil TC<< length of his -ve low voltage Tx period. This give long period of near constant magnetic field with sudden switch off, and for his Tx signal, agreed, G(t) = a*(t+p)^b approxmate true for VRM decay signal, but G(t) = a*(t+p)^b must be wrong when Tx coil TC is not << length of -ve low voltage Tx period.


Clancy.

[Aziz]

Aziz,

G(t) = a*(t+p)^b where p is preamp+coil group delay (like “phase delay”), cannot be correct for general PI Tx.

Imagine -20V applied Tx coil for just 50microseconds, followed by back-emf period, i.e. short magnetic ramp with sudden switch off. Obviously shape of VRM decay must be function of this short magnetic ramp length. But logically same must be true for all PI systems with magnetic ramps, even long period ramps, VRM signal must be function of ramp length. And logically, long length ramps must have different VRM signal shape to short ones (not just size difference, your “a” value). And also then logically VRM decay signal for PI systems with current ramp droop from the Tx coil TC must be function of ramp length and Tx TC droop.


For PI of Eric, he have Tx coil TC<< length of his -ve low voltage Tx period. This give long period of near constant magnetic field with sudden switch off, and for his Tx signal, agreed, G(t) = a*(t+p)^b approxmate true for VRM decay signal, but G(t) = a*(t+p)^b must be wrong when Tx coil TC is not << length of -ve low voltage Tx period.


Clancy.

Hi Clancy,

G(t) = a*(t+p)^b does even take group delay of the amplifier into account. Inherently. But that isn't an issue.
Substitution of p:
p = g+q, where g=group delay

If we minimize the group delay g -> 0, the remaining parameter q which I'm referring to as p in my equation is the interesting part of the equation.

BTW, the formula doesn't say anything about the history of a pulse. Parameter a is dependent on many factors of course. But that's not the point if you look at the VRM response only.

Aziz

[clancy]

Thank you for reply Aziz.

“p = g+q, where g=group delay
If we minimize the group delay g -> 0, the remaining parameter q which I'm referring to as p in my equation is the interesting part of the equation.”


But if g = accurate group delay ((maybe about 1 microsecond for fast preamp and coil for very short back Tx emf?)), p must = 0 or be very small.
Because if p not small then when t=0 get finite G(t). Theory say this is wrong; it say G(t) must be like 1/t when t near 0 (for accurate g). I know electronics not allow to measure near t=0 of course, but formula must have good theory.
If p<0, when t near zero, get problem of course.


“BTW, the formula doesn't say anything about the history of a pulse.”


Useful science formula should predict from all variable Tx parameters.


I glad you have VRM data for good PI current ramp (when Tx coil TC >> length of -ve low voltage Tx period, small droop). I not know this data. Please post link or data. Thank you.


Clancy.

[Aziz]

Hi all,

I have got more measurement data to test VRM theory. If you want me doing this here, just upload the digitized response data here and we can see, what comes out. I will only publish the results, where the data has been provided. That should motivate you to upload the data here. That's a fair deal.

Cheers,
/\ziz,
the certified data processor (BTW, that's my real profession)

[Dave J.]

Trust nothing which cannot be reconciled with the Fish Equation.

Does it reconcile with the Fish Equation?

-Dave J.

[Aziz]

Trust nothing which cannot be reconciled with the Fish Equation.

Does it reconcile with the Fish Equation?

-Dave J.

How about your contribution Dave? You might have excellent, accurate & expensive instruments and you could deliver very interesting measurement data.
I can look at it and try to find your "Fish Equation" there.
Deal?

/\ziz

[Dave J.]

[Sorry, you Frenchmen, Aziz is torturing me beyond my ability to resist, I'm revealing the secret equation now........]

Fer gawdsakes, Aziz, with all your fancy math you don't recognize the simplest case of the Poisson Distribution?

--Dave J.

[Aziz]

[Sorry, you Frenchmen, Aziz is torturing me beyond my ability to resist, I'm revealing the secret equation now........]

Fer gawdsakes, Aziz, with all your fancy math you don't recognize the simplest case of the Poisson Distribution?

--Dave J.

*LOL*

Have a look Dave,

I have found your "Fish Equation" finally. Yeah!!!!



Cheers,
Aziz

[Aziz]

Busted 1/t VRM Law.

Hi all,

the infamous 1/t VRM law has been busted several times now.
Cheers,
Aziz

[Tepco]

1\t fallen on most samples I measured too. I questioned here, few times, Eric's measurement method, probably responsible for perfect 1\t result, but never mind. What I mentioned, collecting many, say 32 or 64 samples, is probably “ultimate” GB method without any theory, and yes, will work on Mars too, with any mineral composition etc. With 2 or 3 samples, detector receive relatively small amount of information, so knowledge of soil behavior is important for optimal utilization of this limited amount of data. With enough samples detector can “bypass” entire GB design process, just recording soil response in some calibration sequence and subtract it , GB without any theory at all. Only, someone else will do this, I mentioned my attitude toward code writing, last time it was 4-5 years ago, I prefer soldiering iron, not keyboard...
Actually I tried something similar, proof of concept, but not for GB purposes time ago, with multiple samples. With very limited processing power, 8bit PIC on development board, external ADC but less capable than today’s on-chip solutions, it was capable to, for example, accept 1E coin, and completely reject another same diameter coin made from different alloy, just on “record and reject this” basis using just 8 samples. Now it is about time to made this on a single chip, future is unfortunately digital.

[Aziz]

Hi Tepco,

the 1/t law might be valid for TX pulses long enough to reach the flat top TX coil current condition (until almost all magnetic domains has been aligned to the external magnetic field). I think Eric is using the flat top coil current by simply limitting the coil current.

Anyway, "we" are proving some VRM theories by ourselves.

-------------
Now to the forum trolls:
Did you scratch your head long enough about the trivial formula G(t) = a*(t+p)^b ?


Aziz

[Tepco]

My longest pule was around 50-60uS, far from flat-top condition, shortest was less than 500nS. Setup built with log-amp, preamp, band filtered, just above noise floor, and fed to old DSO for averaging. Eric used variable pulse width and integration time for different delays. This with “flat top” is probably correct. What I get is, ferrite is always 1\t, but samples,even ordinary brick have a “tail”.

[Aziz]

My longest pule was around 50-60uS, far from flat-top condition, shortest was less than 500nS. Setup built with log-amp, preamp, band filtered, just above noise floor, and fed to old DSO for averaging. Eric used variable pulse width and integration time for different delays. This with “flat top” is probably correct. What I get is, ferrite is always 1\t, but samples,even ordinary brick have a “tail”.

(emphasis mine)

I don't think so. 1/t law will get busted on "ferrites" too. There are so many ferrites out there.
Care to provide some data peer checking me? I can use advanced data mining tools. And I work accurate and precise.
Aziz

[Tepco]

What I used for “ferrite” reference is very large E100 core ordinary soft ferrite made by Kaschke.de, material type K2004, leftover from some commercial stuff. There are zillions of ferrite types out there, high or low permeability, noncrystalline, alloys like metglass (ok. this is not ferrite) etc, some of them can produce weird behavior. Main issue with this is, as always, completely improvised test setup, inadequate equipment and hastily done measurements typical for DIY. At first opportunity, when I have some time (winter) I will retest all this using more decent equipment and publish measured data. This old crappy DSO for example don't have any PC interface and can store only 3 traces, so all “data” is camera picture taken from CRT. More detailed testing of behavior with different, including very short pulses is interesting.

[green]

I ran some decay curves on some clay from the yard awhile back. With a amplifier gain of 160 the slope was about -1.4. With a gain of 1600 the slope was less than -1. Should gain change the slope? I'm not getting much signal from the clay. A strait line for about about 1 decade. Need to improve my setup. I'll post some plots when I get back home in a couple weeks.