User:CounterPillow/PineTab-Next

From PINE64
Jump to navigation Jump to search
Note: This is not about an official PINE64 product, whether in development or not. It's merely the proposal by one community user.

In the April Q&A, it was mentioned that the PineTab was taking a back seat and the question was as to whether the SoC should be updated if it were to be reintroduced at a higher price point. With the A64's age, I am in favour of this. My proposal for an upgraded PineTab would pretty much be a completely new PineTab.

Potential Use Cases

Besides general tablet use, a Linux tablet could be useful for more than just what general consumer Android tablets do!

  • Stock keeping with barcode scanning in warehouses
  • Control interface for industrial applications
  • Data logging for scientific field work
  • Home automation user interface
  • After-market car infotainment system
  • Potential business and government use for a hardened fully auditable tablet system
    • very lucrative contracts that ensure long term financial viability

Suggested Specs

SoC

RK3568 (probably the plastic packaged RK3568B2 due to supply issues)

Why? Unlike the RK3566, it has PCIe 3.0 x2, which means an M.2 added into the tablet would run at a better speed. Furthermore, it supports USB 3.0 OTG instead of just USB 2.0 OTG. There's also CANbus, for industrial/automotive applications. Software support is not significantly different from RK3566.

Why not RK3399S? Probably uses less power (measure this to verify), has a newer GPU. 4x Cortex-A55 at 2.0 GHz should be "ok enough" for light web browsing.

RAM

4 GB of RAM should be enough and keep price "low enough". Definitely needs to be 4 GB if higher screen res is picked.

Storage

  • 64/128 GB eMMC, removable
  • SD card reader
  • Storage expandability with, ideally, M.2 2280. Shorter M.2 22xx would be an unfortunate compromise.

Screen

The same screen as is being used in the current PineTab. 1280x800 is likely a good enough resolution at that screen size, GPU will probably be happier at that res than at, say, 1920x1200.

Firmware

Tow-Boot pre-flashed on SPI in factory. Updates through fwupd.

Connectivity

  • USB-C with 3.0 speeds and OTG
    • 3.0 speed important! Allows things like docks with Gigabit Ethernet etc.
    • if not too expensive, DP alt-mode on that connector too.
    • if DP alt-mode too expensive, consider HDMI output for feature parity.
    • perhaps the UART moved to USB-C for parity with PineNote, otherwise, use the 3.5mm jack.
  • 3.5mm headphone jack with mic
    • make sure the audio output is less noisy than on the Quartz64
  • Wi-Fi, Bluetooth 5.0
  • PinePhone style pin connector on back for alternate back covers
    • I2C/SPI
    • CAN maybe for industrial uses
    • second USB 3.0 port's lines (including USB 2.0), for e.g. an SDR back cover


Camera

Keep the current setup, if possible:

  • Main Camera: Single 5MP, 1/4″, LED Flash
  • Selfie Camera: Single 2MP, f/2.8, 1/5″

Needs software work in mainline to actually work, since rkcif is not upstreamed.

Why no upgrade? Not sure if that's within the price point or within the RK3568's capabilities, but the cameras will be good enough for e.g. scanning bar codes.

Battery

Original 6000 mAh battery should probably be fine? Make it removable again!

Misc

Probably use RK817 or similar for PMU and audio codec.

Price

Up the price to $199 instead of $99, it's a significant upgrade. $219 would probably be fine too.

Accessories

Optional accessories could include:

  • "rugged" rubber case for drop- and dust-proofing the tablet for industrial/field work uses
  • USB-C dock with
    • Gigabit Ethernet
    • DP or HDMI video output
    • USB ports for keyboard/mouse
  • alternate back covers
    • Industrial GPIO back cover using an I2C/SPI to GPIO bridge and passing through the CAN
    • transparent plastic back cover for late 90s/early 2000s electronics look
      • nothing to hide!
    • SDR back cover that connects to the USB 3 lines on the connector

Challenges

There are a few challenges with this design.

  • M.2 2280 might be too large to comfortably fit into the device.
  • Camera interface (CSI) needs significant software work in mainline.
  • Specs may be too ambitious for price point.
  • We're still waiting for RK356x firmware sources.
  • MCU for suspend firmware is essentially still a complete unknown, not even the TRM really goes into it.
  • RK356x hardware bugs with regards to the GIC, shouldn't be an issue in this form factor though
  • Back cover connector's high speed signals will probably be hard to electrically engineer, but not impossible

Advantages

  • Upgrade justifies higher price point, which in turn can be used to pay for the increased price of some of the components (e.g. screen.)
  • RK3568 is a long term support chip, Rockchip will keep making it for the foreseeable future.
    • Later in the new tablet's life cycle, it may still be useful for industrial/home automation purposes, as opposed to general web browsing.
  • USB-C at 3.0 speeds is a significant upgrade for convergence purposes.
  • Mainline support for RK356x is improving steadily.
    • GPU works, significantly better than A64 and RK3328 GPU
    • Video output merged likely soon
    • PCIe 3.0 being worked on
    • hardware H.264, MPEG-2 and VP8 decoding works and will likely be merged soon
      • HEVC being worked on
  • Better memory controller on RK3568 compared to A64 means much improved real world performance