I’m building a simple DIY HUD for my 2004 Toyota Corolla and would love feedback on my approach.
Goals:
Display speed on the windshield
Future additions: cabin temperature + fuel stats
keep cost low
Why not OBD2 or GPS?
The car only has OBD1 (no speed/revs).
GPS seems unreliable (lag, bad connectivity, higher cost).
Current Plan:
Tap into the Vehicle Speed Sensor (VSS) signal that runs from the transmission to the speedometer.
Use a solderless T-connector so I don’t cut wires.
The VSS outputs a 12V pulse. I’m scaling it down using a voltage divider (220Ω + 330Ω resistors), which should reduce 12V → ~4.8V for the Arduino Nano.
Display the data on a small OLED screen, then project it onto the windshield using a small reflective film (haven’t found the right kind yet—recommendations welcome).
Questions:
Is the resistor divider safe/reliable, or should I be using a transistor/optocoupler for the VSS signal?
Any recommendations for cheap but effective reflective film for HUDs (ideally from Amazon AU)?
Any other pitfalls I should watch for with this setup?
Here’s the OLED I was planning to use: Amazon Link.
Here is the schematic from Tinker cad Tinkercad VSS Reader (only program school offered for schematics that I could simulate the Arduino on. I am open to free alternatives.) if it helps.
It’s recommended. Count on electrical noise from the 12 volt car system.
35 years ago the low voltage side of the ignition coil was tapped for engine speed using an opto coupler. It worked fault free for 10 years. All based on a Z80, 64 bit aritmetics. For speed the gear in use was detected and tuned to give correct readings.
Thanks for the advice.
I’ve come from more of a coding background so I’m not overly familiar with circuits, if you could provide an example or something similar that would be much appreciated! And by noise could you be more specific as in >12v or just small random pulses not connected to the actual speed that need to be filtered.
Sorry, no such available. Look at datasheets for opto’s and application notes how to use them. An oscilloscope applied to the signal You want to use would be the best. A common noise is voltage spikes in the pulses. Random, short pulses, not likely.
Prepare for doing some tests and evaluate that. Designing circuits for guessed noise is a waste of time and efforts.
Yes, please do. The “common opinion” here is that cars of today are terrible noise emitters. I doubt it. I worked on a 1985 model car having no digital electronics, only an electronic ignition module that died after 100 000 kilometers when the car sat at the shop for service…