Honda Insight Forum banner

Controlling the Instrument Cluster's Display

74958 Views 470 Replies 65 Participants Last post by  Mario


A little project I'm working on. :) (Sorry for the horrid picture!)

Here's the elevator pitch: Directly control the LCD panel with a microcontroller to display custom information on the instrument cluster. The microcontroller will also read the data that would normally be going to the LCD and can choose to display that info or custom info.

Maybe you want to display OBDII parameters where the MPG is. Maybe you want to change the charge, assist, and SOC gauges to accurately reflect amps in/out, real battery SOC, etc.

I've got a few other ideas as well. Once I'm finished with this project I'll open-source everything I've learned and perhaps produce some PCBs to sell. We'll see!

I have work and other projects as well, so it may be slow going at times. But I'll try to give updates fairly regularly.
I don't think this has ever been done before. I hope you guys are excited!
If anybody is or wants to work on something similar, I'll be glad to share what I know; just ask.
See less See more
  • Like
Reactions: 5
141 - 160 of 471 Posts
2
So I hooked up my oscilloscope to one of my injectors to see what I got.



I started the car and let it idle, and looked at the waveform to see how long the injector was actually open.

Here, the car was reporting ~0.76ms, but we can see the injector was actually on for ~3.02ms:



That comes out to the value the car reports being almost exactly four times less than the actual injector on time. This held true for a few other samples with slightly different on times that I gathered.

However, when I re-did the calculations, it's still not right. It's at least in the ballpark of what the FCD shows, but too low now. (I'm now assuming 140cc/min instead of 145).

Here's a few more data points (the ms values are 4x what the car reports and fuel flow is 140cc/min):

3000 rpm, 34mph, 2.8ms, FCD shows ~95
= 73mpg (77% of 95)

1700 rpm, 37 mph, 5.0ms, FCD ~100
= 78mpg (78% of 100)

1900 rpm, 34mph, 5.3ms, FCD ~70mpg
= 61mpg (87% of 70)

1900 rpm, 55mph, 6ms, FCD ~100 (original data)
= 87mpg (87% of 100)

Looks like 4x the reported value isn't quite the "fudge factor" we're looking for. Maybe I'll buy an inverter so I can power the scope in my car while I go for a drive to see if it always stays around 4x or if it changes. I'll also have PEGASUS output RPM, ms, MPH, and FCD MPG to my logic analyzer so I can capture the values to look at later (I'm currently looking at the speedo, tachometer, FIN value and FCD and remembering all the numbers until I can stop to write them down).
See less See more
The insight uses high impedance injectors (probably 12-16ohm) which are typically slow to open and close so the FCD might be estimated/fudged to assumes a constant 145cc/min flow rate.

If you want to be scientific (I'm thinking you may do :)

you can use known pulsewidths to fire the injectors and measure fuel flow into a measuring cylinder over say 1 minute.

So hold the injector open for a minute to measure its static flow at your chosen fuel pressure. then if you run a 50% duty PWM signal you SHOULD get halve the volume of fuel, but you probably wont and from this you can calculate the injectors open and close time. It will be very dependent on injector voltage, the ECU should know if the system voltage drops from 12.4v to 10.8v it needs to compensate with longer injector opening times.
3000 rpm, 34mph, 2.8ms, FCD shows ~95
= 73mpg (77% of 95)

1700 rpm, 37 mph, 5.0ms, FCD ~100
= 78mpg (78% of 100)

1900 rpm, 34mph, 5.3ms, FCD ~70mpg
= 61mpg (87% of 70)

1900 rpm, 55mph, 6ms, FCD ~100 (original data)
= 87mpg (87% of 100)
Good data!

So, this makes sense to me. The injectors don't fully open/close instantly, there has to be a ramp up/ramp down. So the longer the pulse the car is calling for, the closer a calculation assuming instant on/off should get to the true number.
Yeah, those calculations don't account for the open/close time of the injectors, which will likely have to be guesstimated as part of the fudge factor.

The injector voltage is a good point, Puggie. I don't have very many data points, but the four I have seem to fall into two categories - 77% of expected fuel economy and 87% of expected fuel economy. I wonder if those correlate to the DC-DC converter being on, and thus the 12V rail being higher or lower? I'll have to start checking the 12V voltage as well.
+1 to the idea of measuring actual fuel flow into a graduated volume. Based on the data I've seen in this thread, I recommend a 20% duty cycle at 30 Hz. That'll let you determine the true mechanical area under the curve (i.e. the volume of fuel released in a given time).
A distant future project of mine was going to require fuel use per unit time as a input variable .. If a good method for this gets ironed out sooner in Pegasus development .. all the better :D
Took a break from coding to work on a vectorized version of the instrument cluster's display. :) This is going to be used in the eventual Pegasus user manual for showing menus, what the screens look like, etc. without having to use photos.

This is all vector art, I can change the colors of the segments to "turn them on" or off and the entire thing can be scaled as large as you want without it getting pixellated. I'll be releasing the file for this eventually.

See less See more
Dang, you are putting in so much nice work.
Now you could replace the oem cluster lcd with a vga screen and still have the oem look and feel, but could display.switch too anything you want. That actually might be an interesting project in it's own right. :)
Now you could replace the oem cluster lcd with a vga screen and still have the oem look and feel, but could display.switch too anything you want. That actually might be an interesting project in it's own right. :)
Yeah, it's definitely possible! I'm not going to tackle that project, but somebody else would be able to with the information I'll release about how to read the display info coming from the cluster.
Then we could 'skin' it with a Jessica Alba theme...
Since your such a genius, it would be cool to have an app that reads and looks like the gauges and would allow customizability.
splash screen?

So even in its basic released state, would a splash screen, that wraps all the gauges up to max for a function check and return to a "pre-ignition key on state" be programmable?

Useless, but looks cool:D

Randall
How did you do that vector oem display graphic?
It's done in Adobe Illustrator, pretty much all by hand. I used my photos of the display and made them as straight as I could, then used that as a guide for making all the segments and whatnot. The bar graphs were made by putting one rectangle at each end, then using Illustrator's "blend" tool plus a circular path to fill in the rest. The seven-segment digits were traced by hand with the pen tool.
It's mostly just drawing different shapes that you can add together and subtract from each other to get the thing you want.



So even in its basic released state, would a splash screen, that wraps all the gauges up to max for a function check and return to a "pre-ignition key on state" be programmable?

Useless, but looks cool:D

Randall
I was thinking about making a custom startup screen, actually. :) I could do it at this point, but I'm focusing on more important stuff right now!
See less See more
Please carry on!!!

I was thinking about making a custom startup screen, actually. :) I could do it at this point, but I'm focusing on more important stuff right now!
Absolutely :D

Randall
It's done in Adobe Illustrator, pretty much all by hand....
Pretty impressive, a lot of work. I was thinking maybe there is an automated tool you used to convert raster to vector... But no, just shoe leather, as they say...
Pretty impressive, a lot of work. I was thinking maybe there is an automated tool you used to convert raster to vector... But no, just shoe leather, as they say...
Thank you! Yeah, those tools exist, and they work well for certain applications. But when you need clean, sharp lines, perfect circles, consistent spacing, etc. it's just not good enough. It's not the right tool for this job.
I've been using Illustrator for several years so I'm fairly decent at it by now. Total, I'd say that whole thing took me about 4-5 hours of work.
That's it? 4-5 hours? That's not bad at all. I thought that 'by hand' would have taken more like 40 hours...
141 - 160 of 471 Posts
Top