Hub quiescent load

[DJO]

I've got a 3rd party USB C hub (pqi 6 port - of which 4 are useable) and the bits that work are fine but it's one hell of a power hog. With the Gemini shut (Stock Android) and the hub plugged in but nothing connected to it it slurps about 9% of the battery capacity per hour, with Ethernet connected and WiFi off it was about the same. The alleged power passthrough doesn't work either, still drains 9%/hr with the power connected.

So I was wondering if the Planet Computers hub has a lower or similar quiescent load and if power gets through to the Gemini?

Didn't use anything fancy to test, just looked at the battery level in Android so my 9% estimate is pretty rough but the load is still about 10x that without the hub plugged in.

[vader]

Silly question. Does it have a connection LED? 9% an hour seems like a lot. That is something like 400mA draw. Could be that the CPU can't sleep when the USB is plugged in. We used to measure about 290mA on a much simpler board when it was running (no peripherals)

[DJO]

Silly question. Does it have a connection LED? 9% an hour seems like a lot.

No, except on the RJ45 socket but with nothing connected they were not lit

 [ Invalid Attachment ]

[RJL]

So I was wondering if the Planet Computers hub has a lower or similar quiescent load and if power gets through to the Gemini?

The Planet Computers Hub is the about the same, uses about 8% per hour. Which I estimated similarly to you, plugging in the hub (no devices connected to the Hub), putting the device into Flight Mode and shutting the device, and then looking at battery consumption after an hour. Doesn't matter which Gemini USB port you plug the hub into, similar draw either way.

By comparison, putting the device into Flight Mode and shutting the device consumes ~0% per hour, so it is definitely the hub.

You cannot charge via the Plant Computers Hub either. But Planet Computers never claimed that you could, so I guess this is "as designed".

[DJO]

The Planet Computers Hub is the about the same, uses about 8% per hour...

You cannot charge via the Plant Computers Hub either. But Planet Computers never claimed that you could, so I guess this is "as designed".

Thanks, that's exactly what I wanted to know. No point in getting the PC one as it's no better than mine so I'll stick with that.

Does seem to be a bit if a design flaw if you cannot use a hub and get power into the Gemini at the same time.

[Daniel W]

So I was wondering if the Planet Computers hub has a lower or similar quiescent load and if power gets through to the Gemini?

Sorry for being late to the party, and you seem to have gotten your answer already. FWIW, when putting my USB-C power meter between the Gemini and the Planet Computers hub (with nothing connected to it), the power drawn fluctuates between 50 and 200mA, seemingly averaging at 120mA. I get the same numbers when the Gemini screen is off. While power plugged into the USB-C port on the hub doesn't pass through to the Gemini, it does power the hub itself.

I repeated the test by RJL, except I plugged power into my hub. After one hour in flight mode (screen off), the Gemini battery level had dropped from 80 to 78%. For a single hour, that's a bit much in flight mode, so it seems the Gemini itself uses a tad more power with the hub connected, even when the Gemini doesn't have to power the hub. Since the Gemini can't receive power from the hub, it would reasonably have to use its battery to power its own USB interface, so it makes sense.

Taken together, this would suggest the power drawn by your hub might be a bit less than those 9% per hour, as some of the power is used by the Gemini to (be ready to) talk to the hub. It should also suggest you are less likely to see any real difference by using a different hub.

[Daniel W]

And now I'm even replying to myself... Anyway, I took a few more measurements with my USB type C power meter. Plugged into the right, as seen from the keyboard, Gemini port, and observed for a bit longer, the hub itself seems to draw between 30 and 230mA, averaging around 100mA. The port provides around 5.10 Volt. Moving the hub to the left Gemini port, the power draw stays at around 100mA, but seemingly with less variation. That port seems to provide around 4.80V.  I then plugged the hub into to USB C port on my Windows laptop. Now the hub just wanted around 30-50mA, with occasional spikes, from a port providing around 4.90V. Interesting. I would guess this is because the power draw to some extent is affected by what the host tells the hub to do, and whatever Windows says, seems to cause a lower standby current draw. That might suggest software optimizations on the Gemini could be able to make the hub itself use a bit less power.

As I kind of guessed, the hub seems to work fine with the PC. While I haven't been able to test the Ethernet port, it shows up as a "Realtek USB GbE Family Controller" in the device manager, and Windows says it's working. As the USB ports likely shares a common controller, they can, per the USB 3 standard, share up to 900mA among the connected devices. Apparently insufficient to power two of my external hard drives, one hard drive and one card reader worked well. Plugging in external power did, as expected, not help, since this limitation is in the protocols rather than in the amount of power physically available. Still, I'm glad I can connect an external USB 3 hard drive to the USB C port of my PC, using the Gemini hub, since most USB C to A adapters are USB 2.0 only (a bit slow for 400GB backups) and USB type C to USB 3 micro type B cables seems as common as hens teeth, Oh, well, I'm veering OT, so I'd better shut up now ;-)