Eddy current effects advice pls.

[Mosaic]

Hi all:

I was experimenting with the effects of a yellow 220µH toroid (shorted and unshorted) with a high current pulse 14AWG passing thru the centre toward the load.
I could use some help with the principle in play here.

Pls note the attached scope images. (constant avg current = 1.8A)
1: Toroid coil NOT shorted - note the blue trace current spike (40A/V) @ 224A peak, also the pulse width is 20µS


2)Same as 1 WITH the 220µH toroid shorted. Blue trace pk current is now 256A and the P.width is shortened to maintain the 1.8A avg.
The yellow voltage trace exhibits more (post pulse) kick back ringing harmonics and amplitude


3) Here is item 1 zoomed to observe the risetime (50nS), a bit of jitter.


4)Item 2 zoomed to observe the rise time, I observe a 'cleaner', less jittery signal.


Empircally, it would seem that the shorted toroid improves the pulse fidelity, current rise time and generates richer ringing harmonics. Also , noting the yellow V trace...the positive endpoint V of the capacitive pulse is the same in both 1 & 2 implying the charge moved is the same, but in less time! Wouldn't that imply that the 14AWG parasitic inductance is mitigated in some manner?

My question is what combination of electrical principles are at work here; Lenz's law seems applicable but how does it improve the behaviour of the pulse signal in the pulse conductor to give a 14% improved peak pulse and reduced jitter?

Edit: I took some more images of the pulse signal at the load (battery) which will include all cabling impedance losses. The first set of images were direct from the PCB output. The voltage scale is adjusted to compensate for the loss in amplitudes.

5) Same as 1 @ the load terminals:


6) Same as 2 @ the load terminals:


7) zoom of #5


8) Zoom of #6

 

[Mosaic]

Ok, I figured out the issue.

Shorting the coils on the toroid, (appears to have caused DC bias saturation effects from the induced currents) negated the increased inductive effects of the torroid ferrite on the pulse. Without the toroid at all....the pulse performance is the same as the 'shorted toroid'. Thus there is no 'benefit'.

It can be seen that the ferrite IS removing a fair amount of harmonics from the pulse ringing.

What I find interesting though is the fast rise time of 50nS is unchanged, but the slower rise time of the current pulse over the 16uS is increased with the ferrite core added.

In effect the ferrite is 'squaring up the pulse from a fast step rise (50 nS) with a slow (uSec speed)ramp to a fast rise with a flatter top. The cost is the peak of the slow ramp is lessened. This becomes more apparent as I add turns of the main pulse cable around the ferrite to increase it's effectiveness.

Is this the same effect as adding a bit of coax or stripline to a picosecond class avalanche pulse transitor? Elongation of the pulse w/o loss of the fast rise time?

The material of the core seems to be yellow/white (type 26) here (yellow on the circumference, white on the sides)
http://www.catzco.com/toroids.htm

Which has highest Q in the sub 1Mhz region. Looking at the graphs here:
http://www.digikey.com/Web Export/Supplier Content/TDK_445/PDF/TDK_InCompliance_Aug2010.pdf?redirected=1

Indicates that this type 26 'inductor' has lost its inductive impedance past 1Mhz and thus does NOT affect the 50nS rise time AT ALL. But it DOES affect the over 1 uS rise times significantly.

 

[dr pepper]

Pic 2 looks like a iron dustcore profile, and stripline on one side pretty much states it is.

Remenance could also be having an effect here, if the ferrite is 'square' this can be an issue.

Iron powder depending on the core is lossy at higher freqs, that might be causing the damping, that or the driving impedance is well matched.

Looks like this is a flyback, however depending on topology ringing can be worse at low load and better at high with push pull.

Are you like me using junk box ferrites perchance, that kinda thing its almost a sport for me.

 

[Mosaic]

It's these cores:https://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_387268_-1

 

[dr pepper]

Data for -26 ferrite isnt that common it seems, however looking at the al factor that stuff looks at first glance more for o/p filtering than for use as energy storage, might be ok for a gdt or something.
Shouldnt be hard to fudge a snubber to reduce the ringing if its a problem.
In the uk N87 or N97 is a good choice for power ferrite, but then us europeans have diffrent terminology.