I have a few questions about the ADCMP605 latch/hysteresis input.
Since it can be set using a resistor to ground, presumably hysteresis is controlled by current going *out* of the pin. The datasheet claims the hysteresis pin has 1.25 V bias and 70 kΩ of input resistance. But this is inconsistent with Fig 5, which clearly shows 7 kΩ (not 70) in the region below 1.25 V. 70 kΩ is also inconsistent with the other numbers, viz. that the minimum resistance is around ~50 kΩ. With 70 kΩ the minimum resistance (to ground) would be more like 500 kΩ.
1) So is the 70 kΩ figure simply a typo in the datasheet or is there a mistake in my reasoning?
2) There are frequent references to current sources to control hysteresis. Shouldn't it technically be current sinks?
3) Hysteresis control seems rather imprecise in general. Is it worth my while to accurately control the control current or can I just use a voltage + resistor? If the quantity that actually controls the hysteresis is the current, the latter is less accurate; the question is whether the loss in accuracy is significant compared to the existing imprecision.