How does it support HDMI IN? Only over the flex cable or over the HDMI port?
Introduce ROCK 5B - ARM Desktop level SBC
Glad to hear that, then i misunderstood your post before, sorry about it
btw, I look through Taobao and see that Allnet China is selling a RMB $30 for $300 discount coupon (essentially same as the one here?), but it mentioned a Q3 delivery date; does it mean that Rock 5 is now further delayed? I ordered the coupon via the link Radxa Team posted in this thread though.
and just wonder: is it not possible to add a microSD slot? as it is much easier to replace microSD than eMMC, even though it is slower (SDR104 as I remember).
microSD card slot is already there. check the board picture, next to M.2 connector.
Thanks. I was thinking it were nanoSIM slot for LTE/5G modem
Not at that price.
With enough GPIO you can add a 3G/4G/5G module of your choice. perhaps even make it into an expansion board that people can buy.
You could try and look at the pin mux https://www.cnx-software.com/pdf/RockchipĀ RK3588Ā DatasheetĀ V0.1-20210727.pdf
I had a try but wow that is a complex one and non the wiser, some does seem to be muxed with the mipi whilst some seem to be dedicated pins.
Might be a question if mipi has been chosen over hdmi-in which would be a shame as there are pins for cec-tx & rx which would be cool for control.
I fear weāve still a misunderstanding.
For the RK3588 upstreaming It doesnāt matter whether Radxa ādonatesā free boards or cash to Armbian since this (upstreaming) rarely happens there.
Armbian is a build system and some sort of infrastructur. With Armbianās build system itās rather easy to collect and maintain a set of patches, test them for interoperability and package them so they can be deployed to userās machines.
Thatās the reason Armbian could in the past and still can often deploy functionality that takes months to years until landing in vanilla mainline kernel. But still the real work has to be done first before the results of this work can be collected and packagedā¦ and thatās writing/porting drivers ā essentially what the linux-rockchip community is doing and what can be observed here: https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git/
I mean the SIM slot, not the modem itself. Do you know that many LTE modems do not come with a SIM slot on the module? But that means he top m.2 slot should be A+E key instead of B key then, to be used for WiFi but not modem.
but the 4G modem modules I know of are all USB 2.0 based, so I donāt think using GPIO to drive modem is a good idea. 5G ones should be PCIe x1 or USB 3.0 AFAIK, but I have a 4G modem and a 4G plan only (installed in a USB enclosure that has SIM slot), so I canāt say for sure.
@enoch if you look at the pictures, you will see top m.2 slot is clearly E-key, for Wi-Fi or AI accelerator. Via affordable adapters, one can convert either m.2 slot to another key, including adapters with sim card, to use a modem in either slot - or just use USB adapter with SIM slot, like you already do, if modem supports USB.
damn this sbc looks really tempting
The GPU that is currently getting Mesa work well at least G78 which is Valhall.
I think it goes G710 is an updated G78 but still Valhall and the G610 is the one below aimed at sub premium, think it just has less shaders.
Then Quad A76 2.4Ghz + Quad A55 1.8Ghz and all the rest potentially this is a very juicy SBC and true Desktop level that isnāt just wishfull speaking.
If you want to support but donāt mind the long haul to get kernel & driver support then the $5 coupon for early adoption to get a discount is also tempting.
The GPU is a huge leap in performance over any SBC I know.
Do you have also an EU distributor to link?
Please, on release date, provide a case, plastic or metal will work.
a bit hard, as Iām getting older, my eyesights are not that good anymore
I tried an adapter beforeā¦ I have a MikroTik RB14e (I āfoundā it from reading Wiki) so I got a B-key to mini-PCIe adapter for my m.2 B-key LTE modem, but it never work; in turn I used an external chassis instead. I remember there were discussions on what āoptionsā the top m.2 would be used for, but perhaps Iād confused Rock 5 with the 3 or E23 / E25.
yep, though the chip itself is not āreadily for direct consumptionā; if you were given a RK3588 chip, you would create a Rock 5 yourself (of course it is a much more complex task)? maybe you can, but I definitely canāt (and wonāt have time to spend on this as a hobbyist).
When real work is done (with that I presume you mean drivers), it has to be ported to the mainline linux, which is not cheap nor simple. After its ported, someone has to maintain it - which is also real work. If none of that happens, things fall apart.
A lot of that dirty work falls into the hands of doers of Armbian community and regular maintainers, which receives close to no financial support for the time they waste doing that.
There are many examples, but I am sure you are familiar with the story about Allwinner video acceleration job, done by professionals. They did the job. They wrote drivers and they left it to the community since they have to eat and couldnāt stay more on the project.
How is the state of that āsupportā within mainline? Years after āreal workā ādriver developmentā was done? I know itās not what we have expected.
After real work, work done and wishes with force from someone else customers:
SPI boot support
someone wants display which he knows it works with legacy (so its a problem of software, armbian in this case)
(https://forum.armbian.com/topic/19127-rockpi-4b-and-raspberry-pi-7-inch-touch-display/
USB OTG
DP also need support to mainline
fan support for mainline
Cameras as well:
U-boot upgrades are always trouble or at least time waste:
Bluetooth fix
Downgraded DDR blob for RockPi 4 to fix 1GB model boot issue
Switch ROCK Pi 4 to all blobs u-boot for Radxaās SPI compatibility
ā¦ and on and on. But since only āreal workā counts, why do I waste time?
If free sample comes without contract of some sort, samples are not even unpacked. When you will understand that board support is for us expense?
All HW vendors sell āDebianā / āUbuntuā as a part of the package. So we can say they certainly want to look like a software developers. They sell āLinux HWā or at least association in some way.
Why would Armbian maintainers be stupid to invest once again (check about and multiply by 10-50 times for next hardware) their private time and money to support:
- end users which can only cover up to 0.5% of costs but demand top service
- vendor that doesnāt or is unable to cover costs that are made
- competition which only need to invest into sales
Try answer such questions ā¦