I’m wondering how the narrow-band lambda signal is converted to a value. Maybe he’s lucky and his car’s ECU gives an analogue output, but afaik if you monitor the signal directly it’s essentially binary (too lean, too rich). There doesn’t seem to be much on it in the wiki.

Um MS in piggyback? that is a very stupid thing to do. Having TWO ECM’s trying to control injector pulses will cause big problems. Sounds like someone does not understand car fuel ECM control.

Megasquirt is a REPLACEMENT of your car’s ECM. it does not piggyback and run with it. I know I have been doing custom ECM’s on cars for a decade, I even have one running a motorcycle for ignition timing.

That said, most of what he is doing is really overkill. Some of that information is useless outside of tuning. Throttle position is worthless without base injector timing. I can get a ODB-II interface to http://www.tunerpro.net/ tunerpro RT for real tuning. you can install a MAP sensor to get more useful information for what this display is for coupled with the RPM’s and ttemp sensors.

It’s neat, but completely useless for anything but monitoring.

@nes:
http://www.autospeed.com/cms/A_110624/article.html
http://www.autospeed.com/cms/A_110625/article.html

0-1v for the narrow band range. If you’re really interested, I’d recommend getting a wideband with narrowband emulation.

Wait, why does the build include 8 thermocouples, boost control, but no wideband or logging?

@nes
“. . .afaik if you monitor the signal directly it’s essentially binary (too lean, too rich).”

http://en.wikipedia.org/wiki/Oxygen_sensor#Zirconia_sensor

While it is true that the sensor is inaccurate away from stoichiometric most ECU’s in closed loop measure how many times a minute the O2s passes the midpoint, necessary to keep the Catalytic converter lit off with fuel and oxygen alternately.

Since the car needs to run at 14:7 (stoichiometric) the only reason for a different type of sensor is acceleration enrichment tuning, which can be done seat of the pants, or trial and error. Obviously this system doesn’t offer that.

I too am curious as to why he didn’t purchase a kit from DIYautotune, they are only $140 http://www.diyautotune.com/catalog/megasquirti-programmable-efi-system-pcb22-unassembled-kit-p-28.html

You can also control the spark with MS, it supports many kinds of Coil packs, dizzys and COP systems

Seriously, *chill out*.

I don’t think “Steve” claimed his project would solve all the world’s problems, so why are you being so harsh?

No offense @fartface, but stop and think for a moment about what you actually said… that this monitoring system “… is useless for anything but monitoring.” Duh. That’s like saying your glasses are “useless” for anything but lookin’ through.

And you, @trey… I don’t disagree with your initial comments that reading OBD might be a cleaner approach to monitoring an engine, but does that really make this project “retarded?”

Since you can’t see the value in this project, let me explain it to you.

(1) “Steve” came up with an idea and turned it into real hardware. That alone should serve as inspiration to all those who post here to crap on other people’s projects, but never seem to get around to building anything themselves.

(2) I’ll bet money that he learned a lot about processors, coding, and handling mixed signals. Anyone who attempts to replicate his device will learn similar lessons.

(3) Somebody else may look at what Steve has done, and decide that his gizmo is perfect for their non-automotive data collection application. @trey, you can’t use OBD if the article you are monitoring doesn’t have OBD.

(4) Somebody may look at what “Steve” has done and decide that *parts* of his system might be useful for a completely unrelated project they are building. If borrowing bits of technology from one system to integrate it into another unrelated system is not hacking, I guess I don’t know what is.

Some people bitch about Hackaday not posting enough real hacks, and yet, frequently when something is posted, the author is lambasted for his “retarded” or “useless” project. What incentive does someone have to share their work when they can expect that kind of attitude?

Trey. There is good reason not to use OBD data when your car doesn’t have it.

“Um MS in piggyback? that is a very stupid thing to do. Having TWO ECM’s trying to control injector pulses will cause big problems. Sounds like someone does not understand car fuel ECM control.”

Umm someone has designed one from scratch and works with them on a daily basis, thanks ;) You can easily piggyback an MS, just (rather obviously I thought, but clearly not!) leave off the injector driver connections.

“MS: its not legal, as it does not give OBD Errors i would fail emissions and a at a police check i loose my license plate…”

Sure, but you don’t have to run it as the soul ECU, thats why I said “at a later date, if you want to”

My point was simply that if you wanted a cheap and easy solution to monitoring most of your normal engine parameters, and even control some or all areas (like boost, even if not fuelling) you can do so in conjunction with your stock ECU. I spent 2 years driving round with the MS monitoring my stock ECU to map out actuator trip points, actual AFRs, knock levels etc. Then I modified it slightly to take full control of fuel and spark and tweaked the firmware to work better on my application.

Sure it doesnt have 8 EGT inputs, I’ll give you that, but it does have 1 which is plenty in the majority of applications.

@capn; thanks! What a useful couple of articles. I also just read their papers on EGR valves. Interesting stuff.

I am considering replacing the carb on my ’72 vehicle with a closed-loop single point injection (just as many manufacturers did with their existing engines around the mid ’90s when catalytic converters were mandated in several territories). Always on the look-out for useful info.

I don’t want to run a MegaSquirt, mostly because the crucial bits of the design, such as the closed-loop fuelling are apparently proprietary. :-(

to all the haters out there:

my 71 opel gt DOES NOT HAVE an ecm. this solution allows me to digitally monitor and record my engine and car condition without any external modifications aside from a vss mounted in one of the tires.

This build will be going into my GT verbatim minus boost and thermocouple solutions as i really don’t need them… but i will be adding a air/fuel mix monitor.

I think its an incredible demonstration of arduino application in a real world scenario, and his problem solving of the mixed sensor array is inspiring.

How would I log all data, to show to the insurance company, when my Toyota runs away and totals a loaded school bus?

St.Jimmy, its an Arduino shield as i was to lazy to make it directly…

if i would do it again the atmega would be directly on the board.

im running it now for some weeks and have no problems with vibrations, all the ic´s sit in the nice sockets not the cheap ones and they work good.

for a real sales product some redesign would be good, but as the people who make will drive it know what they do its ok like this :)

BTW, what I found, is that specs cost abou5 $5k. Congress made it legal for everyone to own the rights to diagnostics… but, as I found, it was a pretty penny. I was actually after the ‘dead cylinder’ test. that was a long time ago. Perhaps the cost of the spec is much cheaper.

if you can read the throttle position and rpm, you should be able to figure out the fuel consumption from your carburator.

a fuel flow sensor would be even nicer, giving you the highest accuracy!

Well done steve, there are always people out there ready and wanting to knock your work, any monkey can type a message but you have to think for yourself to write/debug a program.