Update - March 2021


Some progress and brainstorming since my last post! I'm stoked with how its working for me... I feel its helping me be a safer jumper, because I'm much more aware of my speed while flying. I've sent a couple of these to Europe for people to play with this summer and see what they think.

Various hardware and software updates...
- Faster more capable processor (ARM Cortex M0)
- Ublox GNSS receiver with integrated antenna (SAM-M8Q)
- Flat brighter LED
- Improved LED assembly can go inside or outside helmet
- Included a temperature component in the calculation for Indicated airspeed
- Logging of Indicated speed plus LED colour and some other things

I also put pressure sensors inside the suit (an excellent idea that a few other's have had) to see if I could derive airspeed from suit pressure. As you would expect there is a strong correlation, but the suit pressure fluctuates too much for a steady indication of airspeed.

I used two LPS22 pressure sensors, one in the legwing and one next to my body to measure the pressure difference, sampling at 10Hz. I tried using a 1 second rolling average to smooth the readings, but even that was fluctuating too much to be useful. So for now I've given up on that idea.

I'm now trying a version with 2 LED's, and also a different way of mounting on my helmet.




How fast am I flying?

Like everyone else I get my cues from visuals, noise, feeling and body position. But how good am I really at knowing my speed? The hero in me says I'm fine; reality says otherwise.

The more I jump, the less I seem to know, and last year I gave myself a scare by flying too slowly. In certain situations I'm quite good at convincing myself that I have enough speed... just to get over that ridgeline or pass the notch or whatever. A quick look at The List and you'll see I'm not the only one who does this. The problem is that even though we train this stuff, we are not infallible.

So after a lot of procrastinating followed by a lot of work, here is my version of an airspeed indicator for BASE.

OK so what is it?

Its not a new idea, but something I've thought about for a looooong time...

An LED mounted on my helmet that changes colour depending on my speed. Speed is measured by GPS, and normalised for sea level. Because I usually only jump in zero or light wind, the speed given is a rough approximation of indicated airspeed that could be got from a pitot tube.

The LED is angled towards my face and is bright enough to see in direct sunlight. I've mounted it just at the periphery of my vision.

The unit also has a removable SD card which holds a configuration file and stores log files.

So... there are only really three speeds I'm interested in:

RED: Too slow! Any speed below my best glide speed is too slow.
BLUE: Best glide speed (max L/D) give or take a few km/h.
GREEN: Fast. Flying faster than best glide and with more energy.

How it looks with my (very rough) polar curve...

Wingsuit polar

Speed thresholds / colours

Thresholds for each colour can be changed. The full range of colours are...

DIM GREEN - Standby mode when not moving (so not distracting at the exit)
FLASHING RED - Very low speed
RED - Low speed
BLUE - Best glide speed, give or take
GREEN - Faster than best glide
OFF - Very fast

However not all colours need be used, so for example I've got it set like this for myself...

Dim green (standby) = less than 80km/h
Flashing red = 80 to 130
Blue = 130 to 140
Green = more than 140

Maybe I'm stupid, but I've found having minimal colours is more useful as I can process the info more automatically while everything is going on. Flashing red means danger, that's simple enough for me.

The speed thresholds are set in a text file on the SD card, which can be easily edited.


Jumps are logged to the SD card automatically. 10 data points per second by default. Logging only starts once I'm moving and stops at the landing. What happens while I'm standing at the exit is that it continually buffers 2 seconds of data, so that once it detects I'm in freefall it can go back and write that buffer to the log. So the jump is logged completely from push to landing, and nothing more. The log file is compatible with the Flysight viewer.

True airspeed vs indicated airspeed.

True airspeed is your actual speed through the air. In zero wind this is the same as the total 3D speed given by a GPS.
Indicated airspeed is what you feel, and what is measured in aircraft with a pitot tube. At sea level (and standard air density) its the same as true airspeed, but at higher altitude its less. For example at 3000m if your indicated airspeed is 150km/h then true airspeed will be around 175.

We stall at a particular indicated airspeed regardless of altitude, so this is the measurement we want to know while flying. Speed Monkey takes GPS total speed, makes the assumption there is little or no wind, and using the altitude and a simple algorithm, converts it to indicated airspeed. The result is a little rough because atmospheric pressure and density are variable. More info here.

It wont work well if there's much wind, because GPS total speed will deviate from true airspeed.

The 3D speed of the GPS is logged as well as the indicated speed and the LED colour. So I can use the log file to fine tune the speed thresholds where I think my best glide is.

It looks like a Flysight

This is not a Flysight. Flysight is a genius tool for training flying performance. This is for safety, it helps me avoid making some stupid mistakes.

I don't use Flysight audio for BASE because I find it distracting. When everything is happening fast I prefer instant access to information using a visual indicator rather than audio.


Malfunctions. Could be bad to get an incorrect green light. Need to use some common sense.
Reliance. Could make me lazy about using my other senses for judging speed?
Distracting? Depends on where its mounted and familiarity. I only check it occasionally... when I'm transporting, entering a line, going for glide and such.
Skydiving. I don't think its much use, because usually there's some wind on a skydive.
SD card. Log is written to SD which has to be removed to copy to computer.


The photo above is my forth prototype (I'm now working on version 6 and 7). I'm stoked with how its turning out. Version one was the size of a small brick on my helmet. Version 6 is 5x5cm and 2cm thick. It needs more field testing, and by more people. To that end, there should be a number of them in Europe this summer.
Next step is to design a printed circuit board, depending on what interest there is.

ATSAMD21G18 ARM Cortex M0, 48MHz microcontroller
Ublox 8 GPS: SAM-M8Q 72 channel (GPS, GLONASS, Galileo constellations), 18Hz max update rate
180mAh rechargeable lipoly battery

Programmed in C++, to be open source on Github.


Michael Cooper set the bar with Flysight and opened my mind to what can be done with a GPS. Plus a continual source of guidance and encouragement.

Fran├žois Gouy gave me the idea years ago by hooking up LED's to a pitot tube (mounted on a wind vane on his helmet).

Where to from here

At this point its nothing more than an idea, a personal project, and a prototype. If there is strong interest I can work on getting a PCB done.

Anyway, I welcome any comments / suggestions / critisism, cheers.

Contact me... David Walden