t vago's MPGuino workspace thread, continued

[FlashStopFall]

Could someone point out where build instructions for an mpguino are posted? I'd love to start this project with a nano, but I can't seem to even find a pinout.

[t vago]

Originally Posted by FlashStopFall

Could someone point out where build instructions for an mpguino are posted? I'd love to start this project with a nano, but I can't seem to even find a pinout.

You can start here -> https://ecomodder.com/wiki/MPGuino

[FlashStopFall]

Originally Posted by t vago

You can start here ->

That page is outdated and poorly labeled. It links to your mpguino code, but not it’s own? (v0.86 was the last maintained I presume, and I was hoping to look at that code). It should really be rewritten. Anyway, I understand the schematic, I was just hoping someone had better instructions, or maybe a documented build thread with some newer hardware. And maybe some simpler code. I guess it’s not that big a deal though, I’m just going to try to build one with a raspi pico from scratch.

[Fastvan67]

FlashStopFall, please do share your progress.

[FlashStopFall]

So, I’ve made a decent amount of progress. I’m using a raspberry pi Pico running Micropython for my mpg gauge. I’ve got enough code written for just instant mpg to appear on the screen, and everything should work when I hook it up to my car this weekend. I’ll have to rewrite the frame buffer code for this screen library before I can display any more characters. Using PIO, I can get the frequency and duration of the injector pulse, and also the frequency of the VSS pulse. It’s working on my test bench, just waiting for diodes to arrive.

[FlashStopFall]

So I’m ready to start testing the MPiG thing I made. I hooked it up and got an injector open duration, but only about half the time, and only when idling, not when changing rpm’s. I bought a cheap oscilloscope and found out the waveform looks like this (image attached). I think that first; I have it programmed to read low to high which is backwards but more importantly; I think the voltage spike coming from the injector coil field collapsing is messing up my readings. My wiring is just like in the mpguino schematic, but with different resistors to get me down to around 3 volts. I’d slap a diode on there but I’ve read that can make the injector stay open too long. Do you guys have any ideas for either 1: hardware to get rid of the spike or 2: programming to read it better? Or would you mind explaining how the mpguino does it?

Upon further inspection, that signal is actually 14v, and isn’t being dropped down to 3v, yikes! So that’s problem #1, but I think my other two problems still exist.

[t vago]

the MPGuino code uses two interrupts to measure fuel injection pulses. The leading edge is measured in an input capture interrupt which is separate from the trailing edge. For late model Chrysler products, the leading edge should be a falling edge (InjTrig = 0) due to how the auto shutdown relay powers the injectors and ignition coils, and how the PCM completes ground on each injector to fire them.

Yah, 0.86 isn't all that capable compared to later versions. I'm in the process of completely re-writing my own fork. My code now includes a serial terminal interface which can easily be interfaced to bluetooth or USB, which of course is still in work.

[FlashStopFall]

Originally Posted by t vago

the MPGuino code uses two interrupts to measure fuel injection pulses. The leading edge is measured in an input capture interrupt which is separate from the trailing edge. For late model Chrysler products, the leading edge should be a falling edge (InjTrig = 0) due to how the auto shutdown relay powers the injectors and ignition coils, and how the PCM completes ground on each injector to fire them.

Yah, 0.86 isn't all that capable compared to later versions. I'm in the process of completely re-writing my own fork. My code now includes a serial terminal interface which can easily be interfaced to bluetooth or USB, which of course is still in work.

t_vago, that’s awesome. Thanks for the help too.

[FlashStopFall]

So I forgot to plug my hardware into the injector wire, I was just measuring the injector ground. This is how the signal actually looks, and it looks good, so I think my code is the problem.

[t vago]

Started adding USB support, researched a faster square root algorithm, and started merging the SWEET64 trace functionality into the serial debugging monitor terminal.

The first page of this thread contains the latest description of my MPGuino fork.