I had another short session with the MacBook. I think get the behaviour of "SSD 3.3V" transitions i.e on 2015 13-inch MBP + WD SN550 1TB.
When is "SSD 3.3V" = 0.00A. Should you believe it's truly close to 0A?
When the machine is connected to AC power, idle current is always around 0.26A (or 0.23A with NVMeFix). Only when disconnected from AC power, things start to change. Idle current remains 0.26A (or 0.23A...) but after idling between 2 to 8 minutes, it drops to 0.00A.
"SSD 3.3V" readings should come from Apple's SMC that measures current supplied to the SSD 3.3V power rail. So I tend to believe it's accurate account of the actual power usage (most of the time except when it's 0.00A). One counter example for the exception.
Disconnect the machine from AC power. Let it idle to 0.00A. If you re-connect AC power now, "SSD 3.3V" (as well as "PBus") will go back to some "sane" values immediately. But for this experiment. Let AC remain disconnected. Fire up BlackMagic and run it a few times, "SSD 3.3V" remains 0.00A through out the runs! That shows the reading definitely not correct when 0.00A is displayed. I could see SN550's temperature rising towards mid-40C. So it's is working hard.
Hopefully this is just a glitch in Apple's SMC after the SSD gets right out of lower power states. I don't know what's the exact behaviour of the original Apple SSD. I remember seeing 0.03A as lowest but can't recall if it's on AC power or not.
Actual power usage when "SSD 3.3V" reads 0.00A
Simply can't tell but we could make an educated guess. I believe APST is enabled by Apple for compatible 3rd-party NVMe SSDs. For SN550, take 0.26A as the "active idle" current. It yields about 858mW "active idle" power consumption. Very close to 964mW measured by
Tom's Hardware (sorry, can't find a better source). Assume the difference comes from better efficiency in Apple's NVMe driver or simply due to Tom's tests in a desktop environment.
A pristine battery for 2015 13-inch MBP is 74.9Wh. SN550 idle current thus induces about 1.15% battery drop per hour. An original battery should have aged quite a bit over the years, this percentage will appear even larger. So with "Standby" disabled (
sudo pmset -a standby 0
). Let the MacBook sleep a couple of hours. At the end, calculate the % battery drop per hour over the whole sleep period.
For SN550, if the battery loss is lower than 2% per hour, safe to bet "SSD 3.3V" = 0.00A may actually mean close to 0A consumption. If the battery loss is higher than 2%, SN550 idle current is probably not the culprit of your battery drain BUT we aren't sure "SSD 3.3V" = 0.00A actually means close to 0A idle current either.
However, we have a fix to solve the "drain" by triggering hibernation sooner. Apple's default is 3hr. During this time, if the machine drains 2% per hour, accumulated drop will be about 6% regardless how many more hours the machine hibernate after the first 3hr. If overnight "drain" is, say, 25%, it actually implies a likely 8.33% drop per hour.
This explains changing "standbydelaylow" and "standbydelayhigh" from 3hr to 25min helped achieving near "zero" drain in our overnight sleep.
Finally reached peace of mind. Cheers.