Greg G

And today's concept: lockout. Slow cruise, car goes into a high idle with idle valve at 60% duty...

Greg G

Finally got some decent progress done on the idle today!

These are the idle settings that seemed to work. No more wild oscillations, just a big dip when AC goes on. It will catch the droop 99% of the time! PID theory is a bitch...and it's weird this PID, it likes more I as opposed to P and D. Does that make sense?

Any suggested improvements are very welcome!


Tables: please give me suggestions!
Breakpoints are now easily correlated. Can I steal another column from the high rpms to give to the idle area?

I have a volcano in the center of the VE table from VE analyze live...is my spark map too aggressive??? I kinda blended the stock map loaded on with the more aggressive maps on the forum

Are my AFR targets OK?


Graphs:


A few things strike me:
1. battery voltage dips during AC... is this a bad battery/alternator?
2. weird timing dip during the voltage droop...due to battery? Goes lower than commanded by the spark table.
3. small surge on AC switching off, due to a timing bump. Was thinking of flattening the 1600 rpm column, but will that kill engine response?

MSQ and MSL attached. Comments would be most appreciated!!! Pleeease help me with the spark map, before I screw something up

JasonC SBB

Momentary voltage dip during a/c kick-in is normal. That's a large electrical load suddenly placed on the system.

You say it catches the dip 99% of the time - it may be worth logging a couple of on/off cycles and stopping the log once it catches a dip, then examining what the duty cycle is doing differently when it does dip.

18* of advance at idle, while it saves gas, reduces the effect of "more advance when it dips" to actually catch a dip.

JasonC SBB

MS guys:

What is the reason/advatnage of having a different set of MAP and RPM breakpoints for fuel, spark, and AFR targets?

Braineack

iTrader: (62)

cause you can? Help you get resolution in areas you need it. For the most part I like to keep them all standardized, but for my A/C on spark map, I like a bit different rpm and load points around my idle cells, things like that.


His log looks to be idling along 14°, but his map clearly doesnt go below 18°, so we need to figure that out, which will help the dip. It drops down as the rpms decrease, and agian it should be rising to around 20* according to his spark map.

Remember, the CPU can only react to an event, as soon as the a/c is triggered the load will increase, idle will want to drop and MS will have to react to it. The Idle Duty does a prety good job reacting, but it's that 4° drop in timing, the increase in load and the richened AFRs that makes the idle speed drop.


Does a more advanced spark map really help MPGs at idle? I idle at 12° @ 850RPM and 18° with the A/C on.

Greg G

When I had the idle at at flat 14, the dips were about a 100 rpm deeper, bringing me close to a stall. Is it too simplistic to raise it up as I have? That lower left area is flat, but I did put a slope between AC idle (40 kpa/750 rpm) and no load idle (29 kpa/1000 rpm).

I did turn on the idle advance, though I really can't tell if and when it's working. Can anyone check the settings I have? It's supposed to advance the timing further once rpm drops below 800 and load is greater than 32. Timing is raised a flat 15 degrees- I don't yet kow how to tune that curve....

Greg G

Well, the current idle is the most stable so far. Same settings as shown above.



I just don't understand why there's a momentary dip in spark advance on AC activation, from 17 to 10 then 14 then back to 17. The table is a flat 18 at that area...ideas?

Also, PID is at 5/30/10. Looking at the PWM idle duty, a bit of oscillation. Suggestions? More D? Less P? P is already 5!

mxnige

Been reading this with interest. I've got this all to come with my turbo'd 96 euro 1.6.
DIYPNP arrived today.

Looks like you're making good progress though. Keep at it.

Greg G

Good luck! It's a big learning curve, even if you think you did your homework ahead of time! Nothing like actual hands on experience to cement the knowledge in your brain

Is your DIYPNP built, or still for assembly? Don't forget to add the sequential injection kit.

Make a build thread when you start! It helps keep you pointed in the right direction. Especially when you feel like this --> or -->

Greg G

I found the reason for the weird spark droop on initial AC load! User error

Turns out the idle advance timing curve is in absolute values! I thouht I was adding 15 degrees to the spark map, when in fact I was commanding a lower timing on AC load. So essentially, I was making it droop more

I thought it was like a timing retard table...thanks to Braineack for schooling me on it...

Greg: "my idle sux"--> <--Braineack: "WTF are you doin to my box!?"

mxnige

I've still got to assemble it and yeah, i've got the sequential mod kit.

It's not so much the building that worries me, i've got experience working with pcb's. It's the troubleshooting when things don't work and learning how the software works that seems daunting.

I'll make a build thread. Probably start this weekend.

I don't have AC so thats one less thing to worry about, but good to see you figured yours out.

Greg G

Ahahaha funny you should mention that...

I got to the shop today and saw the car sparkling clean! Mechanic says he washed it. Aaaand now the AC compressor is always engaged! Not cycling at all! AC switch off--still on! WHAAAT THE FFUUUUUUUUUUUUU To top it off, it's a holiday tomorrow

The AC fan and blower still obey the switch. I guess that rules out everything up to the AC relay.

The good part is I got the opportunity to dial in the PID with the AC on

EDIT: found the issue. The wiring to control the AC fan relay was miswired to the AC relay. So instead of turning thte fan on when hot, it turned the compressor on! Duuuuuuuuh

Greg G

OK, I identified this. In neutral, quick throttle blips are OK. A little dip but liveable. If I hold the revs up, look at the idle valve. It closes and doesnt come back, so when I let off, I get a big droop and the engine dies.



I know it's some setting, can anyone point me to it?

EDIT: It must be this right? Close delay settings at the bottom? What are the right settings? Never close?

Greg G

Behold I have created idle! LOL

It's not too bad. There is an initial dip, pretty deep--to 500 for a bit, then slowly goes to 950-1000. I catch it with idle advance, then the idle valve comes in.



1. Should I use more spark advance to catch the dip? I'm just using 2 degrees advance, do you guys use more?
2. Can I increase the ramping up of the idle valve?

Running around, the car runs pretty well. But I noticed these "injector hiccups" in the log. What are they?

Greg G

Looking at my logs, a light bulb went off in my head! I decided that instead of using the idle advance table to react to dips, I would try to keep using the idle advance table all the time when AC is on.

The results...promising! Better control of oscillations. Will work on it some more, but it looks like I'm making progress

Braineack

iTrader: (62)

a light bulb? isnt that what I told you to do?

Greg G

Yup, but now I understand it

Greg G

Been letting these concepts percolate in my head. So here's what I've got as of today:

I lowered the spark advance in the idle areas to 15, and adjusted the idle advance settings in such a way that as long as the AC was on, it would stay with the idle advance curve, which i set to 22.





These are the results:
Initial AC dip is about 400-450 RPM, but it never stalls out now. spark advance catches it, and idle valve (slowly) ramps up. Am trying to figure out how to make it ramp up faster. PID gets into a mild oscillation-- (PID: 4/30/10) Still need to fine tune closed loop idle, was working on making the spark/ve tables solid.



I did notice that when I burn settings onto the box, the idle valve goes to cranking value immediately...seen in the left side of the graph. I wish there was a way to command that instant response to an AC trigger, or if not to AC on, to conditions like TPS<2, Map >x, RPM<y, maybe MAPdot/RPMdot. Something that gives an immediate response to a rapid RPM drop below your set idling value...



msq and msl attached.

Thoughts comments suggestions are very welcome

Greg G

When I unplug my idle valve connector, I expect idle to drop--but it goes up!

Left side--unplugged; right side--plugged


What's up with that?

Greg G

Valve mode setting is normal (0%=off)