Ever since I started flying FPV drones back in 2018 the one thing that bothered me, was how our FPV goggles were the only thing that the FPV community treated as a black box. We buy a pair of them, we do our thing and that’s it! Maybe a DVR mod or power switch mod here and there and not much more!

In 2021 during the lockdown I didn’t have anything better to do, so I started investigating what makes these goggles tick and how could I design a pair. After looking at many teardowns of both FPV and VR headsets, I found that the entire headset design revolves around the displays. For FPV headsets, there exist two categories:

• slim goggles containing two microdisplays and pancake optics. While these are the most lightweight option, they have a high cost, with the bulk of it being the microdisplays. A well designed headset of this category is the Orqa v1/v2 by Orqa FPV.

• box goggles. Usually they come with a single 4-5 inch display and a Fresnel lens. They are big and cumbersome, but they are also cheap and can accommodate pilots with bad eyesight by having the space for glasses. A well designed headset of this category, is the eachine/skyzone cobra goggles.

Actually I kinda lied, some headsets belong in neither of those categories. Headsets like the DJI v1/v2 goggles, Topsky prime 1s/2s, SJ RG01 contain two small displays and simpler fresnel lenses. They combine a relatively smaller frame with the big FOV that box goggles offer.

• The DJI goggles are mostly well received. They use two 2.5’ displays, and they crop them in order to simulate a 1’ display, in order to bring the FOV to a manageable level. Their main disadvantage is that they work well only with the first gen DJI FPV system. They lack HDMI inputs and the analog input they have adds lots of latency.

• The same cannot be said about the topsky and SJ headsets. They both lack the same thing, high resolution displays, and this is the reason why they both didn’t succeed commercially. Also, their quality was quite bad since both designs tried to be as cheap as possible. Despite being bad products, I find their core design idea very interesting and I believe that if those two companies tried to make a more premium product, they would have succeeded.

What if I could combine these 3 headsets into one?

What if I could make a headset with high resolution cropped displays like the DJI v1, while having the form factor of the Topsky prime, AND being open source for anyone to tinker with?

The research

After some digging in Aliexpress I found some cheap VR display kits. These contain 2 displays, a generic driver board with an HDMI input, and a specialized board to bridge the displays to the driver board. These kits are not what I want. My requirements for the electronics are:

• analog video input
• hdmi video input
• the ability to scale down the display, to simulate a smaller one
• OSD and menu support

After more searching, no kits came close to fulfilling any of those requirements. This means that I will have to design my own video processing system and display driver.

The displays

As I mentioned previously, the main component of any FPV or VR headset are the displays, so I will first need to select a panel and then design the rest of the headset around them.

A major limiting factor when it comes to choosing suitable displays is my current living status as a uni student and my geographical location. I do not live in China, nor do I have the appropriate reputation inside the VR industry to communicate directly with display manufacturers or sellers. The situation becomes even worse when many of the manufacturers purposely exclude critical data from their datasheets(display initialization code).

After a LOT of web scouring for datasheets and initialization code, I stumbled across Project Northstar, an open source AR project that uses the BOE VS035ZSM 3.5’ panel, with a resolution of 1440x1600. While it is kinda big for my needs, it is easily accessible on Aliexpress and it is cheap. For now I can create a proof of concept and then jump to a better display if needed.

One thing that most of the VR panels that I came across have is their communication interface. It is called MIPI(Mobile Industry Processor Interface) DSI(Display Serial Interface), it is proprietary and requires an annual fee to view it’s specification. A quick google search can solve this issue. This common interface makes it also easy form me to switch to a different panel on the future.

The video processing

I haven’t been able to find an off the shelf ASIC that is readily available to me and can fulfill my requirements, so a FPGA will probably be needed to scale the video and add a OSD menu. As for the video inputs(analog and HDMI), off the shelf ICs will be used. A possible candidate are the ICE40 and ECP5 series FPGAs by Lattice. They are reasonably priced, and they are well supported by the open source FPGA toolchains.

The drivers

MIPI DSI is not a very well known protocol, thus few FPGAs support it, and those that do and offer adequate speeds come with a high cost. My solution for this, is the ssd2828 ASIC that converts parallel RGB video to MIPI DSI. This ASIC is readily available, and there is lots of info about it on the web.

Conclusion

This is quite a monumental project and will take some time to complete. The next post of this series will be about the MIPI protocol and making the display function.