Comparator with adjustable hysteresis

Cost no object? Less than a dime per unit? What?

It seems that such a dramatic change in frequency with load is both unnecessary and would lead to more switching losses at low loads -- but I'm not you, so maybe there's a sensible reason.

Parts of the NE555 essentially work like a comparator with a wide hysteresis with one switching point adjustable. Would your concept work with a comparator that has each switching point adjustable? If so, you may be able to do something clever with a pair of comparators and an RS flip-flop, like the innards of a 555.

There are some comparators with analog-programmable hysteresis, but a more flexible approach is to use two comparators driving an RS flipflop. You can set the upper and lower trip points anywhere.

See my patent, 3,638,041 Sample and Hold Trigger Circuit, PDF near bottom of my home page. Hysteresis value is proportional to a current source. ...Jim Thompson

Hmmmm- "Hall-Effect Sensor IC Bandwidth ? Today, the usable bandwi dth of most linear Hall ICs is ? 20 kHz. Signal voltage changes lit tle up to this frequency. However, noticeable phase shift becomes distinct at somewhat lower frequencies. Some variation is apparent amongst different ICs and vendors, but the rolloff is quite steep beyond ? 20 kHz. A lthough the cutoff frequency for the ?3 dB rolloff of all linear HE Ds is inconsistent, 20 kHz to 25 kHz is a valid approximation."

The classic approach to adding hysteresis is to divide the output with 2 Rs & feed that to +in. It's fairly obvious one can adjust the divider with 1 or more FETs. This is fairly trivial, you don't say what the issue is.

Thanks for all your posts!!!!

I used 2 comparators + 74hc74 + some current-mirrors an resistors

Best Regards Harald