Sponsored

Performance HUD

AustinSupra

Member
First Name
Bryan
Joined
Oct 16, 2025
Threads
1
Messages
7
Reaction score
7
Location
Austin texas
Car(s)
2026 Supra
Thread title: Factory Supra HUD tachometer project — working V1, progress and next milestones

Hey everyone—wanted to share a project I’ve been working on: adding a live tachometer and shift lights to the Supra’s factory windshield HUD.

My car is a 2026 GR Supra 3.0 Premium, six-speed manual, running NBT Evo NBTevo_T24081I. The idea is an S2000-inspired RPM display that uses the existing HUD hardware and makes room for factory navigation and radio functions.

We now have a working V1 installed on my car.

What’s working

  • Live engine RPM from the factory engine-speed feed.
  • Large RPM digits, a 24-segment tachometer bar, and five orange/red shift lights.
  • Native operation on the head unit—no connected Mac or laptop needed during normal operation.
  • Automatic startup, with a settling period and checks that navigation and radio are idle before drawing.
  • Removal of the custom graphics when factory navigation or the radio selector needs the display, followed by automatic return once those functions release it.
The current layout fits within a 150 × 36-pixel bitmap, so getting readable digits, smooth updates, and correct colors into that space has been a big part of the work.

What we’ve actually verified

The current November 18 baseline survived an iDrive restart, and I completed a drive with no visible tachometer loss. The post-drive checks matched the installed files and configuration, recorded six clean handoffs to factory functions, and showed no new permanent application fault records.

That doesn’t mean the app ran as one uninterrupted process: it deliberately stops and resumes drawing around factory activity. The goal is for that behavior to feel seamless while preserving factory priority.

Earlier live testing also helped resolve unwanted symbols, flashing, color mapping, and graphics cleanup.

The difficult part

Drawing a tachometer was only part of the challenge. Making it coexist with the car’s own display behavior has required much more work.

This uses a shared factory guidance path, so it isn’t a completely independent HUD layer. We’ve had unsuccessful startup-integration attempts and head-unit restarts during development, followed by factory recovery. Those results shaped the current startup, cleanup, fault logging, and rollback approach.

I’ve been using ChatGPT Work to assist with research, code, test analysis, and documentation, alongside actual vehicle testing. Passing software tests is useful, but I’m keeping that separate from what has really been demonstrated in the car.

Next milestones

  1. Repeated startup testing: overnight sleep/wake and repeated cold starts on this exact final build.
  2. Longer endurance testing: more runtime and more navigation/radio transitions.
  3. Broader factory-function checks: warning and assistance-display interactions that haven’t yet been covered.
  4. Better documentation: clearer setup, compatibility checks, troubleshooting, and restoration instructions.
  5. Future features: explore configurable display settings and additional data such as boost, after the V1 baseline has more validation. Those features are not currently implemented.
Can other owners install it?

The source and compiled V1 components are on GitHub, but this is not a universal, plug-and-play installer.

So far, the verified installation is one car on one firmware version. The public download still requires a verified vehicle drawing profile, the target car’s own startup configurations, and compatibility checks. Another owner’s factory configuration should never be copied blindly.

There is a documented rollback path, and the original factory executables were not replaced. Firmware flashing or downgrading is not being proposed as a way to make other cars compatible.

The shift-light thresholds are display settings, not validated engine limits. Extended reliability and every possible factory warning interaction remain open testing items.

Project links

The image in the README is a software-rendered preview, not a windshield photograph.

I’ll use this thread to share progress, actual HUD photos/video, problems we encounter, and milestone results. If you have experience with NBT Evo, QNX, or the factory HUD’s navigation/radio behavior, I’d be interested in comparing notes.
Sponsored

 

nanaisu

Well-Known Member
First Name
Ronnie
Joined
Jul 2, 2024
Threads
2
Messages
342
Reaction score
512
Location
United States
Website
bananaisu.com
Car(s)
2022 Supra 3.0 Premium
I'll say that I'm a bit hesitant to use anything that leverages hooks just due to stability and what you mentioned above "...followed by factory recovery", unclear if you mean esys/ista level recovery or the "head unit self healing"
 
OP
OP

AustinSupra

Member
First Name
Bryan
Joined
Oct 16, 2025
Threads
1
Messages
7
Reaction score
7
Location
Austin texas
Car(s)
2026 Supra
That’s a fair concern, and I should have been clearer about “factory recovery.”

During earlier development, two attempts to temporarily replace the running controller caused head-unit restarts. Afterward, we verified that the original startup configuration was unchanged, the original controller was running without the experimental hook, and navigation, radio, iDrive, and the HUD were operating normally. That wasn’t an E-Sys/ISTA reflash or recovery procedure. “Self-healing” would also overstate what we established—we checked the resulting state and confirmed normal operation.

Those failures led us to abandon that controller-replacement approach. The current version uses a different startup approach, but it still uses hooks, so your underlying stability concern absolutely remains.

The current build has passed a restart and a drive with clean factory handoffs. Repeated overnight cycles, extended endurance, and all warning/assistance interactions are still unverified. I’m comfortable continuing development on my own car, but I wouldn’t present that evidence as enough to recommend installation to someone prioritizing factory-level reliability.
 

Denton

Well-Known Member
Joined
Jun 11, 2019
Threads
6
Messages
387
Reaction score
507
Location
BC, Canada
Car(s)
23 MT
During earlier development, two attempts to temporarily replace the running controller caused head-unit restarts.
Only 2? ;) I'm up to crash 535,231 & have straight out cooked the built in nav app(s) several times leading to some sweaty palm copying of my backup over the top.

HU forgot my name now too :)
 

nanaisu

Well-Known Member
First Name
Ronnie
Joined
Jul 2, 2024
Threads
2
Messages
342
Reaction score
512
Location
United States
Website
bananaisu.com
Car(s)
2022 Supra 3.0 Premium
Only 2? ;) I'm up to crash 535,231 & have straight out cooked the built in nav app(s) several times leading to some sweaty palm copying of my backup over the top.

HU forgot my name now too :)
if the head unit doesn't forget who you are, are you really doing it right? My copy/paste "unfuckshitup" block 🤣
Code:
mount -o remount rw /net/hu-omap/fs/sda0
on -f /net/hu-omap/ /bin/sh

rm -r /opt/aa/nbt-aad.old
rm -r /opt/aa/NbtAad.old

mv /opt/aa/nbt-aad /opt/aa/nbt-aad.old
mv /opt/hmi/ID5/data/ro/Plugins/NbtAad/libNbtAad.so /opt/hmi/ID5/data/ro/Plugins/NbtAad/libNbtAad.so.old
mv /opt/hmi/ID5/data/ro/Plugins/NbtAad /opt/aa/NbtAad.old
rm -r /opt/aa/NbtAad.old
 
OP
OP

AustinSupra

Member
First Name
Bryan
Joined
Oct 16, 2025
Threads
1
Messages
7
Reaction score
7
Location
Austin texas
Car(s)
2026 Supra
LOL I'm still doing a lot of exploration and discovery but holy shit that's a lot of crashes! I've been trying to do a lot of documenting and discovery slowly to guard against crashes where I can. But admittedly it's been a lot of coding and then as I'm sure you know getting in the car, grabbing the crash code, etc. You MUST be building something sophisticated with that many crashes! In truth its a bit of a side project for me that I'm sharing with others along with some learnings.
 

Denton

Well-Known Member
Joined
Jun 11, 2019
Threads
6
Messages
387
Reaction score
507
Location
BC, Canada
Car(s)
23 MT
You MUST be building something sophisticated fucking around a lot with that many crashes!
Yep.

I moved to using arming files on USB now though, so if something does cause a crash, I just don't arm the same test/function again (I'm shimming binaries).

It's been a fun couple of dumb weeks with basically 0 to show for it. Oh and I don't own a battery charger, so that's been fun too.
 
OP
OP

AustinSupra

Member
First Name
Bryan
Joined
Oct 16, 2025
Threads
1
Messages
7
Reaction score
7
Location
Austin texas
Car(s)
2026 Supra
I aspire for it to forget my name. Right now I'm trying to broaden the usable space.
 

rwense

Well-Known Member
Silver Sponsor
First Name
Evan
Joined
Dec 10, 2021
Threads
47
Messages
806
Reaction score
1,314
Location
Pittsburgh, PA
Car(s)
21 Nitro Yellow A90 Supra, 10 Honda Fit Sport 5spd
Awesome work, im interested in seeing the development.

A couple questions (and I know this is early so this more so thinking out loud)
1. Im curious to see the refresh rate of the display. Personally, shift lights should be as fast as the tach (see my CAN-based shiftlights here: ) have you played with altering that rate or have you found the limit?

2. I know for some people, brightness is a bit of a concern, especially with polarized sunglasses, it would be interesting to see if you could add a flashing capability to draw attention, especially for those with sunglasses.
 
OP
OP

AustinSupra

Member
First Name
Bryan
Joined
Oct 16, 2025
Threads
1
Messages
7
Reaction score
7
Location
Austin texas
Car(s)
2026 Supra
Both good points, and thanks for linking your CAN-based setup.

  1. The current HUD app targets an update every 250 ms—about 4 updates per second. The RPM digits, tach bar, and shift lights update together. That means the lights keep pace with our tach graphic, but it doesn’t establish how closely either follows actual engine RPM.
I haven’t measured the full delay from an engine-speed change to the image appearing on the windshield yet. It looked smooth during testing, but smoothness and responsiveness are different things, especially for a shift cue. At 4 Hz, there can be nearly 250 ms between updates before accounting for other delays. Measuring that end to end and testing whether we can safely increase the update rate are next steps.

  1. The current shift lights are steady; flashing isn’t implemented yet, but an optional flashing cue is worth exploring. My preference would be a configurable threshold and flash timing, with the ability to turn it off. It would need testing to make sure it attracts attention without becoming distracting or interfering with factory information.
Flashing wouldn’t necessarily solve the polarized-sunglasses issue, though—it changes the timing of the cue, not the HUD’s brightness or polarization. We’d need to test visibility with different glasses and daylight conditions, and I haven’t established whether independent brightness control is available through the interface we’re using.

For now, I’d prioritize measuring latency and improving the update rate, then evaluate flashing as an optional feature once the underlying behavior has more validation. HUD does have a brightness setting that the app responds to as it controls overall HUD brightness
Sponsored

 
Last edited:
 








Top