Hi everyone, for my dissertation I'm looking at knock sensor signal processing. At uni we have a 1.1 rover k series engine on a dyno that I have
to use. The engine is running on the original rover ecu, but also has the provision to be wired into a megasquirt ecu. However the megasquirt ecu was
installed by another student many years ago and isn't currently working. I want to take knock signal data next week with the engine running like
normal, and knocking.
Is there a way to force the engine to knock on the standard ecu? The ignition trigger on the camshaft doesn't seem to be adjustable
My thoughts are to remove the distributor cap bolts and rotate the cap approx 10 degrees, giving me 5 degrees of ignition advance. Then possibly just
a big washer on the two distributor cap bolts that will hopefully catch onto the boss that the bolt used to go through, to keep the cap on. Will this
work even though the engine is controlling the ignition advance curve?
Yes K series sensor is on the flywheel. I used to run a k-series with an MS-2.
The easiest option would be to put a toothed wheel and sender on the front of the crank and then you can vary the timing (There are a couple of wheel
patterns for the K series, EU2 and EU3)
Just seen your pic it's an EU2.
Cheers
Ian
[Edited on 10/3/16 by mackei23b]
Rotating the distributer won't change the timing as this is controlled by the ECU, it will fire the coils in relation to the load, speed, crank position etc...
Rotating the dizzy cap won't alter the spark timing. The ECU does that and the cap/rotor just distributes it to the right plug; Rotating a
distributor used to work because the timing was controled by points or hall effect trigger, but not any more.
Don't know about adjusting the timing via the standard ECU. Suspect you'd have to get the megasqurt working.
Thanks guys, I was just thinking about it and realised that rotating the dizzy wouldn't work.
Looks like I'll be having to get the megasquirt working then. Since all my data will come from a fixed rpm there should be minimal tuning needed
to get it to run enough for my tests.
A little brake cleaner in the intake will induce knock they also knock like the devil on EeziStart.
The SAE will have loads of papers on that kinda thing, have you done a quick google scholar search? What do the guys calibrating knock sensors or
characterising knock with schlierin photography etc do?
I guess WOT / high load and some engine speed where cylinder filling / mixing is poor would get you most of the way there.
I wonder if bringing the engine as close to knock as you think you can with the dyno then streaming in a little volatile gas into the induction air
stream, or heating the inducted air would work - you want the effect to be easily modulated I suppose
Keep the ideas coming guys. Not sure what the uni will think about spraying easy start into the intake while its running. Can't really be in
the test cell while its running.
The project is to use signal processing techniques to accurately detect engine knock vs no knock. So I just need two sets of data
I had a customer car a few years ago with a knock sensor that had come unscrewed and was dangling by it's wire; The engine knocked like crazy under load. So... what happens if you just disconnect the existing knock sensor from the ECU?
There isn't a factory knock sensor on this engine, we've added one for data acquisition.
Read Sir Harry Ricardo classic work "The High Speed Internal Combustion Engine" it may be ancient but it contains all the pointers you need. You really need to think it out yourself.
Sounds like the easiest thing would be adding a factory pattern trigger wheel to the front of the crank with an adjustable mount for the factory crank sensor
I'm well aware on how to induce knock when given control of the engine. But was wondering if there could be a way to easily crank up the ignition timing using the rover engine to get my results.
Just reduce the fuel pressure to the injectors, make it run lean, put it under load, away you you go knock knock!.
Ian
Are you doing this at a College/University? Does that College/University have a Chemistry Dept.?
If so, get hold of some n-heptane or whatever you can find that's volatile, add that to regular gas.... I've experimented with fuel mixtures
in a younger life... it will knock!
I had a customer with an Astra that kept going into limp home , when we road tested it we got a fault code that meant the engine was knocking . When investigated I found the oil level was low , topping the oil level up cured the fault. A bit drastic but it could work in your case?
quote:
Originally posted by rusty nuts
I had a customer with an Astra that kept going into limp home , when we road tested it we got a fault code that meant the engine was knocking . When investigated I found the oil level was low , topping the oil level up cured the fault. A bit drastic but it could work in your case?
