Hey Nico
a) this sounds unusual, there shouldn’t be noticeable power consumption differences between an 8gb and 16gb fins, pinging @hraftery for his thoughts.
1 Like
b) this currently is not the expected behaviour for RMA fins; there should not be anything running on it and I believe this to be accidental. In the future, this behaviour will be standard but we will present a way to access the devices QA/RMA test reports and the splash screen will point you towards a getting started guide.
1 Like
c) we are also aware of this issue 
it’s been discussed again this morning and the conversation with the Etcher Devs has been pinned to get the usbboot issue fixed. We’re planning to stop maintaining our own usbboot library and conform with Raspberry Pi for stability. I’ll jump back in here again to let you know when that fix is pushed and ask you to give that a test again for us! Thanks for investigating and providing more context here!
1 Like
On (a) I agree, power consumption differences should be negligible. That said, power supply requirements during mounting are stringent, so perhaps your old setup was borderline. At idle it only draws about 120mA, but during the mount process this regularly spikes to over 300mA and importantly, if the voltage droops during this process the mount process may be interrupted. Technically an unpowered hub is only required to supply 100mA, and the CM3 in rpiboot mode actually only requests 100mA, which doesn’t help matters. So successful mounting of a Fin requires a USB bus that by default puts out 500mA without significant voltage droop, which most do these days even though it’s technically more than required by the spec.
@hraftery
Yeah thats what I thought to. I bought an USB voltmeter now to be sure, but will take a while for it to arrive. I will report back on those readings as soon as it arrives 
@hraftery @bucknalla
Well, I got a small new toy and some data (yeah its no a calibrated unit and such, but it gives a rough idea).
“Known good Fin” with 8 GB (plugged into laptop)
normal current draw: 0.08 A
max current draw: 0.14 A
Replacement Fin with 16 GB (plugged into laptop)
normal current draw: 0.15 A
max current draw: 0.16 A
Replacement Fin with 16 GB (plugged into powered USB hub)
normal current draw: 0.12 A
max current draw: 0.20 A
Bottom line: directly plugged into the laptop, the 16 GB Fin does never successfully boot into MSD / “harddrive mode” and gets flagged as faulty usb device. With an external, powered hub it does work. The 16 GB Fin does draw more current, but is absolutely in spec with USB 2.0 - so bottom line, it looks like the USB port on my good old EliteBook 25x0p is dying 
Very interesting! Those 8GB numbers are strangely low, but could still be within tolerance. It’s still possible the 8GB/16GB thing is a red herring, but there’s definitely something strange going on. Anyway, I wouldn’t write off your EliteBook too soon - it may just be better adhering to the USB spec. Since the device only requests 100mA, it doesn’t technically need to supply any more than that. Much of the USB world only works because most modern hosts contravene the spec by supplying more than negotiated by the device. Fortunately, your powered hub is one of those hosts, and is happy to supply the necessary current.
1 Like
FWIW, to flesh out my numbers earlier in the thread, here’s some current consumption traces I recorded both while idling and while mounting when powered by USB PRG on 8GB Fin v1.1.1.
Current up top, voltage below, x axis is seconds. Yellow is average, red is high frequency noise.
- 1st graph is from power on and left to idle. The current consumption pattern repeats after that.
- 2nd graph is same period but with Etcher running. Initialising starts almost immediately and is 75% done by the end of the 30 seconds.
- 3rd graph is a little further along (look for similarities with graph 2 to figure out time delta). By the 24 second mark the Fin is mounted.
1 Like