Gauging Interest - New Hardware Product - Fully Featured Control Panel

Hardware
1

Hey Folks,

Just trying to gauge some interest… I hope this is ok on here??

I’ve been engineering an IP based encoder (Ethernet) which is USB based using some new cost-effective RISC-V silicon.

The solution provides a plug-and-play full framerate highish-res - up-to 4x 1080p60 or 6x 1280x800p60 - virtual displays per USB device with each being outputted via IP/Ethernet.

It would support both Windows and Linux on relatively low-end hardware (e.g 5th Gen Intel onwards) with no discrete GPU requirements (iGPU is fine).

Optimised for ultra low latency as well as multi-touch input.

A POE + USB-C powered Ethernet and WiFi6 Dual Band thin-client with 10.1" multi-touch display has been proven too.
Smaller screen sizes are possible, for example, square 4" display with native resolution.

With a bit of app engineering it’s possible to run multiple Chromium instances on each of the virtual displays headlessly & reliably.

Goal is for it to be truly plug-and-play, simply install the app and driver, plug the USB device in and connect to network.

Connect client(s) to network and you have multiple fully featured Home Assistant dashboards (any WebUI) which are also able to view cameras in as many rooms as you like.

Cost wise - actively trying to engineer a solution to be highly competitive.

Would you be interested in something like this - no headaches, no complexity, always available, fully embedded bare-metal firmware which provides industrial grade levels of reliability?

There will be other features too such as local push alerts, RGB, Neural Voice and Voice Assistant.

Any and all feedback is welcome :slight_smile:

Thanks in advanced

2

So you need both a PC in a closet with the usb attached as well as the PoE thin client on the wall?

3

Thanks for your reply. I wasn’t clear, apologies!

USB Encoder requirements (any one of):

  • A relatively low spec energy efficient PC, a Micro PC (e.g Dell OptiPlex 3040 micro), running Linux or Windows. (360fps shared total per USB encoder / 6x dashboards)
  • RPI 5 - can be a Home Assistant Yellow/Green. (180fps shared per USB encoder / 6x dashboards)
  • Other SBCs in future (subject to USB/kernel validation)
  • Almost any virtual machine that supports USB passthrough - e.g HassOS VM

Setup across all will be as easy as possible. Prebuilt Linux images will be available for standalone instances. An Addon for HassOS.

Up-to 2 USB encoders can be connected providing 12x unique dashboards. Multiple thin-clients can share a dashboard if required.

Thin-client connectivity options:

  • POE powered wall-dock offers a discrete seamless wireless connection for dockable tablet - room specific range, for ethernet prewired households
  • Tablet can be connected to almost any WiFi directly where the USB Encoder is connected - standalone use
  • Desk mount dock for tablet will be available (power only)

Thin-clients power options:

  • Mains wired power (not all countries have a neutral connection in their light switches however - e.g UK)
  • POE
  • USB-C

Smaller sized thin-clients (e.g 4") would be in two variants - wireless only USB-C powered for desk use or mains powered wireless only with relay (e.g light switch install)

Understand that it’s not for everyone. It will provide a scalable energy efficient multi-room control solution using embedded hardware that offers high performance, secure, locally hosted and customisable.

Thin clients will have a capacitive biometric fingerprint reader which offers rapid user switching to their dashboard in any room.

Currently exploring a single virtual display encoder implementation in software for the RPI 4 & 5 / HassOS with a governor to ensure no impact HassOS (no USB encoder). A single dashboard with biometric user switching might be enough for some users…

4

Why can’t the thin client run a browser itself? It seems overly complicated to have the usb encoder as another point of failure?

5

I can appreciate why you think that but I disagree - it’s not overly complicated, it’s a plug and play hardware accelerator that will be very cost effective.

No headaches & no tinkering with Linux or apps on each device. Unbox, plug it in, run a single quick setup, and it just works with every client connecting automatically.

Re the single point of failure - the same can be said for almost anything - including let say a RPI running HassOS - where do you draw the line?

USB is reliable if the hardware and firmware design is to exacting standards.

This has been engineered from the ground up to be reliable and for this specific purpose and to do it well.

Soak and load testing has been performed with stability for over 30 days has been confirmed and will continue. Across several hardware platforms too.

The thin client wouldn’t really be a thin client if it ran a browser, as such.
Cost is a driver for the thin-client, creating a cost effective multi-room product without sacrificing on quality is the goal.

Counterintuitively, having a centralised solution can increase quality and reliability by providing a controlled environment that’s easy to manage and maintain compared with discrete solutions such as Android tablets.

A significant advantage which I think most might appreciate - there is zero data collection from any device, no Android/IOS and is extremely secure and private.

Features such as multi-room music streaming that do not require casting technologies (Sonos, Google etc) are relatively trivial to implement with minimal impact to the host PC/VM/SBC.

The cost savings on the clients permit other hardware features to be added whilst remaining competitive; such as wake-word & voice recognition, music streaming, audio & visual alerts as well as future expansion including sensors (e.g mmWave Radar Presence).

A fully independent secure wireless solution incorporated into the USB encoder is possible. Simply connect to almost any PC/SBC/VM and it will support up-to 6x 10" thin-clients and 8x 4" clients wirelessly (dual chip design) subject to WiFi range but must guarantee network level QoS something of which we have a longer term solution for.

Hope it helps