[Arm-netbook] RK3399

Pičugins Arsenijs crimier at yandex.ru
Fri Feb 9 16:55:54 GMT 2018


>  there are dual-cell solutions out there, google bunnie huang laptop
> power board.

I've found AXP259, will give it a shot - but will also look into Novena power circuit, thank you!

>>  Thankfully, we don't need to do as much as AXP209 does (and the computer card has an AXP209 in it anyway).
>
>  yes but it's placed into "5V DC input only mode" i.e. VBUS is shorted
> to DCIN (read the datassheet on this). cards must be treated as
> OTG-POWER-CAPABLE. this is IMPORTANT. you need to respect the fact
> that EOMA68's 5V power is DUAL DIRECTION JUST LIKE OTG POWER.

That is very important - thank you for noting this. It should be straightforward to solve, thankfully. I've read the AXP209 datasheet before, and this is pretty much what I expected from looking at the card.

>  so you cannot just shove 5V in to the EOMA68 card you ABSOLUTELY MUST
> have a one-way current-limiter at the ABSOLUTE MINIMUM (SY6280 does
> this job very well). this will terminate all and any possibility of
> doing OTG powering from incoming OTG Chargers but it is safer than
> fucking up and killing the card *and* the Housing *and* the battery
> *and* the DC mains power supply *and* the OTG Charger due to a current
> fight between conflicting incoming power supplies.

Makes sense. If my understanding is correct, this won't terminate the possibility of OTG charger powering the board. Say, you plug the card into the laptop (powered from battery), then plug a USB charger into the card. Given that the protection is, AFAIU, supposed to be right after the card's 5V pins on the PCMCIA socket, the end result would be: 1) protection IC being triggered 2) card being powered from the USB charger 3) all the remaining laptop internals being powered from the laptop battery. There's also one more question:

If a laptop is plugged in an AC source, plugging a USB charger will result into two power supply grounds being connected together - is there a possibility of anything destructive happening? Say, one charger is grounded and the other isn't, or they both are grounded, is there a possibility of current flowing?

>  bottom line: please do NOT design this circuit without public
> consultation and without my FINAL approval. remember, i am
> responsible for ensuring that the standard is safe for people to use.
> if you do not agree to this then you may NOT claim it is interoperable
> with EOMA68, you may not make ANY mention of EOMA68 anywhere at all.
> not even to say "it is quotes like quotes EOMA68...."

No worries! Will send you the schematic for a review - of course, to this mailing list, too (and I have other people to consult locally). You might want to make a set of requirements for EOMA68-compatible products that make them eligible for the "EOMA68" mention - I imagine that, if you don't set as much requirements as possible beforehand (especially when they're as important as this), it will be problematic for you to track down everybody who might create an EOMA68-capable housing and review their design practices. I also imagine that reviewing a schematic after a schematic will get tiring quickly.

Additionally, I suggest adding appropriate reverse current limiting to EOMA68 breakouts (available as Crowdsupply perks). If that's a feature that each and every breakout is very likely to need, you might as well make it easier for people to comply with the EOMA68 requirements (especially since all the people that ordered these breakouts are likely to do something out-of-line with them).

>  apologies but i really need to be strict about this as it is down to
> user safety. get it wrong and you could end up killing someone due to
> a lithium battery fire.

This is perfectly understandable - if my product ended up causing an injury for somebody through no fault of my own, it wouldn't be pretty.

Cheers!
Arsenijs



More information about the arm-netbook mailing list