DIY Fuel injection thread.

[Roman]

Hmmm okay thanks... Yes it's happening just around the rpm region where VVTI first turns on, so it could be a rattly pulley. I've tried adjusting the angles but I dont think I've tried shutting it off completely.

 

Unfortunately I cant listen directly with my knock sensor anymore as it's wired into the ECU and there's no audio output from it.

 

But I can definitely hear knock (or something knocking!)

[Roman]

Hmmm yep this could be it. 

 

 

12 degrees is the position when there's no advance applied. 

Then its bumping a few degrees ahead, just as rpm and throttle goes up that slight amount.

[Roman]

Went for a drive

Previously my VVTI map looked like this:

 

 

So I tried setting it all to zero between 2000 and 4000rpm except for full load area.... Seemed to  sound the same once car had warmed up.

So I went the other way, and set it all to 25 degrees including the low down areas where I'd always had zero.

It looks as though this may have sorted it.

Maybe the time that it clatters, is when its just coming off its base position, or is just above it.

So set to stay well above it and so far so good. Will need to go for a longer drive to confirm though.
 

Maybe this type of oil is contributing, recently started using a Penrite 5w30 when I've previously been using magnatec 5w30.

I've noticed that oil pressure goes a fair bit lower at idle now, possibly running hotter oil temps or something.



Sheesh I think I owe you a few beers by now Downtrail! And a few others here too. Thanks for the point in the right direction 

[Spencer]

This thread is great.

 

That Atlantic ECU looks like a late 80's 4age stock ECU with the software opened up, only two injector drivers! 

[fuel]

who needs sequential injection anyway?

[downtrail]

 

Went for a drive

Previously my VVTI map looked like this:

 

 

So I tried setting it all to zero between 2000 and 4000rpm except for full load area.... Seemed to  sound the same once car had warmed up.

So I went the other way, and set it all to 25 degrees including the low down areas where I'd always had zero.

It looks as though this may have sorted it.

Maybe the time that it clatters, is when its just coming off its base position, or is just above it.

So set to stay well above it and so far so good. Will need to go for a longer drive to confirm though.

 

Maybe this type of oil is contributing, recently started using a Penrite 5w30 when I've previously been using magnatec 5w30.

I've noticed that oil pressure goes a fair bit lower at idle now, possibly running hotter oil temps or something.

Sheesh I think I owe you a few beers by now Downtrail! And a few others here too. Thanks for the point in the right direction 

 

 

 

You will probably get rid of it by interpolating the table a bit so its not so sharp when it comes on. say at 40kpa map 2500 rpm have your 0 then at 2750 - 6.5 deg etc do this the whole way around and see if it smoothens things up a bit.

[Roman]

Phil is now one of those blokes,

Who's quite sick of my mitsi jokes.

He saw no potential,

in fully sequential,

As fuel isnt why his car smokes.

[Roman]

You will probably get rid of it by interpolating the table a bit so its not so sharp when it comes on. say at 40kpa map 2500 rpm have your 0 then at 2750 - 6.5 deg etc do this the whole way around and see if it smoothens things up a bit.

 

"Zero" degrees on that map is actually 12 degrees.

 

Because I'm using the PID for VVTI control from an Altezza that has a different offset on the cam angle sensor I think (12 degrees different I'd imagine...)

 

So it's maybe not as sharper transition as it would seem. However yes this is a good suggestion, thanks. 

 

Thinking about it, I've had a rattle at idle for frigging ages, and I've always thought it relates to valve clearances even though I've checked several times and they're all within spec. So I might have a good listen and see if it sounds like its coming from the pulley. If so I'll just replace it.

[bigfoot]

Ok. Next computer related question:what is a good driver to download to run a usb-serial adaptor?the one that came with my cable doesn't want to work

[h4nd]

chipset?

[Ned]

It needs to be for that cable, can't just download one willy nilly

[Roman]

download drivers willy nilly until one of them works

[downtrail]

"Zero" degrees on that map is actually 12 degrees.

 

Because I'm using the PID for VVTI control from an Altezza that has a different offset on the cam angle sensor I think (12 degrees different I'd imagine...)

 

So it's maybe not as sharper transition as it would seem. However yes this is a good suggestion, thanks. 

 

Thinking about it, I've had a rattle at idle for frigging ages, and I've always thought it relates to valve clearances even though I've checked several times and they're all within spec. So I might have a good listen and see if it sounds like its coming from the pulley. If so I'll just replace it.

 

 

Ah i see theres a callibrate for the vvti sensor under triggers menu. perhaps callibrate it so it reads as it should. also will make less confusing when im looking at it like wtf its at 12 but at 0 hahaha. you could also try changing the frequency thats output. see if that helps it from jumping around so much in your data log. 

[Roman]

Oh god how have I managed to miss that cam angle offset haha.

 

Thanks I'll get that done.

[Celica RA45]

also dave have you reset the cam pulley back to zero for a re plot .on the motec we disconnect the power to the solenoids and rev it to 3000 so that we dont get cam creep and rezero them back to 0,then you can start moving them after that 

[Celica RA45]

and the pulleys get loud with thin oil try some 10 w -60 super syn from penrite 

[dwlee]

Roman, have you tried a custom PID setup? Looking at the default setting in Link's Altezza base map the filtering is set to the lowest level and derivative gain looks quite high (but hard to know without knowledge of how the algorithm was coded). The high derivative gain and a noisy position signal may cause oscillation in the control signal. Also, the large difference between the actual position and the target like at the start of the log might cause integrator windup, which can cause strange things to happen, so best to fix the offset error (hopefully link have a good anti-windup strategy build into their controller).    

[Roman]

Thanks Dwlee, no I havent considered a custom PID as I've considered it beyond the scope of my abilities (especially since theres a VVTI preset already there)

I think the high derivative gain is on account of that the pulley naturally tries to pull itself back to the most retarded position when the cambelt pulls on it?
 

However, the plot thickens!

 

Check this out:

 

 

When my RPM drops below the minimum threshold that VVTI operates at, the cam angle does drop to zero.

 

But when it comes to the VVTI operation rpm range, if I look at custom PID settings.
The minimum duty cycle for the oil solenoid is currently set to 20%. Which causes... 12 degrees advance.
 

Mind blown, have been driving the car for years like this and never bothered to check, just assumed 12 degrees was zero haha.

It likely means more power up top by retarding the cam that extra 12 degrees. Will be interesting to see if my fuel map changes by much.

 

Tomorrow I will reduce the minimum DC for the oil solenoid and see how it goes.

[dwlee]

I assume you want to be able to set the nominal zero position from which you advance and retard the cam and set that as your base position? The 12 degrees measurement maybe fully retarded and that's the reason they set the minimum duty cycle at 20%?

 

Some basics of PID control are that the input to the controller is the difference between desired value and the actual value, the error. The control signal is made up of the sum of three components that are related to the error:

• Proportional: proportional gain multiplied by the error. Think of it as the error is this much so we will apply that much control signal. Zero error equals no proportional component to the control signal
• Integral: the integral gain times the area under the error curve. Think of as the part that drives the control signal so that eventually actual error goes to zero.
• Differential: the derivative gain time the slope of the error curve. Think of it as the part that tries to resist the rate of error change.

The problem with PID controllers is that the measurement of the actual value is noisy and the real physical actuators have limits to how much signal they can transmit and the rate at which they can change their output. The noisy measurements will cause that noise to be amplified by the differential component of the controller. The physical limits of the actuator will cause problems such as; integrator windup (too higher integral gain) and or oscillation (too higher derivative gain). To higher proportional gain can cause actuator saturation.

 

In your case increasing the level of filtering will reduce the impact of measurement noise as will reducing the derivative gain. I would not recommend changing the integral gain and to a lesser extent the proportional gain. The best way to manually tune PID gains is to perform step change tests where you change the desired position from one value to another and look a how the actual position changes in the logs.

[Roman]

If you look to the right of the yellow vertical line on the time plot. When it applies 5% dc cam angle does drop to zero instead of 12 deg.