Idea for an "econometer" for non OBD-II EFI...
 

HI guys,

Just came up with a strange idea this morning... you take a battery powered analog clock or watch, you set the hands at 12:00, take out the battery, hook it across an/the injector, note your odometer reading, take a trip on a test loop, note your mileage, and the actual elapsed time on the clock, which should have recorded actual injector "on-time". Using those figures and lb/hour figure for your injectors, you should be able to determine actual fuel economy for your test loop... if not actual fuel economy, something that's pretty close and that should be repeatable. i.e. if you get a 10% increase in these figures between tests with different devices or driving methods, then there should be a 10% increase, whether or not it was actually 32.2 to 35.42 mpg or 33.5 to 36.85mpg...

... only thing that's bugging me is whether it would actually work with a quartz controlled clock or not, I'm thinking it might with one that has a continuous sweep second hand, but not with a 1 second tick one, although I might be wrong, depends whether it counts pulses from the 32Khz oscillator or charges a small capacitor.

A quartz controlled watch wouldn't move a second until you had injected 32,000 times. You could wire a relay to connect or disconnect the crystal but I don't know how those watches or the crystals will behave under those conditions.

Ah yes, they divide it down to 1 second pulses...
https://electronics.howstuffworks.com...rtz-watch2.htm

That sucks... means you have to keep the watch powered but interrupt the crystal.... can do it with an AND gate, BUT makes it as easy to do for a digital watch... Think it would give a good enough resolution, when smallest likely injection event is about 100 microseconds (If it would go to 1% duty cycle at 6000 RPM ever) in the midrange it'll be timing events around 2000 microseconds 2ms, which will probably give accuracy to 1%ish

I'm not sure if you're talking about the measuring device or the injector, but I've read almost 45% duty cycle at 6000RPM@WOT.

Yes that's more realistic, I was wondering the minimum period it would have to measure, maybe see 1% if you chopped throttle.

It's not a percentage, but in terms of miliseconds and if it's below 1 milisecond, the car won't idle properly (since that's probably where you're going to get the lowest duty cycle).

Quantisation error, that's the word I'm looking for, if minimum resolution is 31.25 microseconds by interupting the 32Khz pulse at the crystal... which fortunately is usually pretty much the only component of a watch big enough to identify.