BPI-R3-mini with OpenWRT

SoC: MediaTek MT7986A 4x A53
Flash: 128MB SPI-NAND, 8GB eMMC
RAM: 2GB DDR4
Ethernet: 2x 2.5GbE (Airoha EN8811H)
WiFi: MediaTek MT7976C 2x2 2.4G + 3x3 5G
Interfaces:

  • M.2 Key-M: PCIe 2.0 x2 for NVMe SSD
  • M.2 Key-B: USB 3.0 with SIM slot
  • front USB 2.0 port
    LED: Power, Status, WLAN2G, WLAN5G, LTE, SSD
    Button: Reset, internal boot switch
    Fan: PWM-controlled 5V fan
    Power: 12V Type-C PD

MT7986A 4x A53 two native 2.5G HSGMII

https://openwrt.org/toh/sinovoip/bananapi_bpi_r3_mini

I see a picture, but not a single feature you're looking for.

If you're going to make such a claim, please explain why you feel this way.

Otherwise, this is likely to be flagged as spam.

Beyond that...

OpenWrt doesn't have any 'official' devices except for, loosely, the OpenWrt One device which is in development. This is something of a demonstration of what is possible, but by no means a device of 'universal' appeal as people have different needs based on their applications.

  1. Small size and easy to carry
  2. Higher performance

MT7986A 4x A53 two native 2.5G HSGMII


MT7981B only one native 2.5G HGMII

And what makes this better than a device with more Ethernet ports? Seems to me that this is the configuration that you would find best, but others may need different things.

But OpenWrt doesn't make hardware.

Well, feel free to spin up a product, sales channels, and support services.

Since when?

Yeah, and which companies are officially sanctioned by OpenWrt? Please enlighten us.

Gotcha, so you’re just talking out your backside

Feasible, maybe...

I'm guessing that you've never developed hardware. It's not just an on-paper design work and spinning up a manufacturing partner. For wireless to work properly, you need to have serious skills in RF design and antenna theory, plus test equipment and faraday cages that will run several hundred thousand US dollars. Oh, and you probably need the cooperation of the SoC vendor -- there's usually a need for design level assistance from their field applications engineers. In order to get their attention, you often have to be able to present a large enough volume commitment such that the time they invest in helping you realize your design will actually benefit them in the long run. There's also documentation that you may need but not have access to without a proper customer relationship agreement.

Sure, maybe you can get your manufacturing partner to do that work, but then there are certifications -- don't underestimate how time consuming and complex it is to get all the regulatory approvals in each region. It's also expensive. And if it fails the testing, you've got a new hardware spin and the corresponding recertification testing.

Now that you've got hardware designed and manufactured, you need sales channels. And not just sales -- you also need reverse logistics because people return things. And there is hardware support and warranties for when hardware fails in the field. Oh, and of course a commercial product almost always need some level of commitment to support. A volunteer support community may not be sufficient -- when people buy commercial products and use the firmware that is pre-loaded from the vendor, there is usually an expectation of dedicated staff support.

So, as I said... you're welcome to spin all that up if you'd like. But for the record...

This is absolutely, positively not true. Please do not make stuff up.

1 Like