Estimate minimum current draw for a device

Question

I'm trying to sum up the minimum current draw for all devices fed by a buck converter in order to determine the necessary inductance value. The max current is typically easy to find on datasheets, but I'm having trouble getting values for minimum current. For instance, the datasheet for the LTC2292 ADC specifies a max current of 95mA and a typical current draw of 78mA, but leaves the min column blank. I could play this safe and take of min current of 0A, but that heavily constrains the inductance value I can use and seems very unrealistic. Is it safe to take the typical value as the minimum current? Or, is there some other way to estimate this?

EDIT:

I'm seeing multiple answers to the question of how to choose an inductor for a buck converter use a 10% rule, where the current in question seems to be max current, although is not always specified. For instance, this answer explicitly states 10% of max current and this other answer also uses 10%, but does not explicitly state whether it is in reference to max current or typical current. Is it safe to use 10% of either max or typical current for choosing the inductor, rather than worrying about the actual minimum current draw?

Answer 1

The minimum will highly depend on the ic and what it is doing. Some microcontrollers can go into a deep sleep state and draw microamps, while their idle but not sleep state could be milliamps.

You could use the datasheet where quiescent current is listed (Iq), or if you have a specific board in mind, build one and measure the current draw while idle or asleep, using that as your baseline.

Your adc lists this as power dissipation. At full activity, 78 mA at 3V or 236mW. Then it shows shutdown and nap modes at 2 and 15 mW, essentially 0.65 mA and 5 mA. Leakage current may bring this up a bit, so 1 and 6 mA. But that's only if you use the sleep or shutdown modes.

You best case is taking advantage of sleep modes for all parts of your board, and measure to confirm. Obviously you want a quality ammeter.

The 10% is a rule of thumb. It won't beat actual measurements, but at these current draws 10% should be good enough. There are diminishing returns in lowering these tolerences.