Behavioral Modeling Question

Additional information... transfer function: Unity, but if you change direction Iin must change by Ihys before you resume the unity slope ;-)

Larkin's input solicited... I'm in need of a laugh... I'm tired :-) ...Jim Thompson

This moron killfiles me, and then solicits my help.

If I helped him, how would the idiot expect to see it? And why would this creep expect me to help him, when he has gone out of his way, repeatedly, to grossly insult my wife?

We're still waiting for his solution to the "trivial" mosfet biasing issue in the thread above.

What's wrong with this old fool?

And why does he, the self-described Master Designer, need help with a simple problem like this?

John

How do you define unity gain with hysteresis? The threshold must be relative and never absolute. So if there is a delta-I a little less than Ihys and the signal reverses slightly, then it has to start measuring hysteresis again from the new starting point? Does the output just hold its value in the meantime, and snap suddenly to a new value when Ihys is reached?

Me too, and I'm pretty sure most people are.

So, you've finally noticed that? I've been withholding some of the most obvious answers I have, because he's an asshole.

Apparently he gets off by crowing about killfiling people he disagrees with, which is pretty much everybody.

Cheers! Rich

Being a Master Designer must be a lonely thing.

John

Imagine two unity slopes, separated by Ihys, so hysteresis is local, direction change requires a delta Ihys before you resume slope.

(I'm trying to behavioral model a magnetic field sensor including the zeroing current effects.)

...Jim Thompson

Are you thinking of something brutal along the lines of

if (dI/dt >=0 ) Iout = Iin else Iout = Iin-Ihys ?

Or do you want something more realistic with smoother continuous transitions like tanh() hysteresis transitional behaviour?

Regards, Martin Brown

Like this...

Solved ;-) ...Jim Thompson

-- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

Yeah, it is (sigh).

Thanks, Rich

is

That solves the up slope case what about the down slope case? Just an addon on the way to random case.

?-)

Observe arrows ;-) It covers both slopes. I'm still mulling (and asking questions of the mfg), is there a time constant.... should it gradually return to mid-point if current is held at a constant value? ...Jim Thompson