VEMS : revs rpm dropping off when engine loaded

[A80Avant]

may i say that reidar has a nice explanation there BUT i have to say this to people who think that previous firmwares have had problems with idle. I dont believe it.

i am still on 1.1.81 and my idle is responding to changes nicely. I use much smaller values on the PID (something like 20/12/2) and much larger on integral limits and that allows the PID to make big changes if needed but slower!!!!

If you use large PID values, then the ECU is trying to correct a problem making large steps. However, if your integrals are tiny (such as reidars) then the amount of intervention from the PID you allow is very small, so its like shutting it off!. If your idle is always close to the desired/specified value of say 900rpm then all settings (ignition timing, IAC control, etc) do not have to do a lot of work to get it right.

In my opinion the PID is there to prevent over and under shoots, and if the rest of the settings are set properly, then it will react nicely to problems like high load and such.

Quattrofever: PID values will have NO effect whatsoever on idle unless you blib the throttle while channging them in order to see what happens to the idle as it tries to set back down to your desired value. PID is trying to control idle when it is away from it. So the easiest way to test your PID is while you make a change in the values, you blib the throttle slightly all the time and allow it to rest to the idle value and see what happens. Does it overshoot? does it undershoot? does it die? etc. Ideally you do that with all electrics on, find a PID setting that works there and then turn all off and see how much different that is (ie how higher the idle sets with all electrical equip off.)

In my opinion, as Raydar writes, the beginning of the procedure is right, BUT once i put in the ignition based timing settings, i DO NOT recalibrate the ref. curve. I leave it as it is, because that means that if i do that, then i dont know how the ignitiion based timing will eventually react to changes, and what the final outcome will be.

If i fix the curve and i have a value of say 35 to keep an idle of 900rpm @ 75C of coolant temp, and 15deg of advance, and once i turn the ign. based timing ON, then the idle drops or changes from that position, then there is something else wrong. Either some of the ign.timing settings are fishy, or the curve is not correct, because if my desired idle is reached by the curve alone, then ign.timing SHOULD NOT affect the idle even when turned on. Ign. timing and PID are there to control offsets of idle NOT the idle itself. That is done by the ref curve of the ISV.

Quattro, i think the best way to help you is to try and meet up at some point so i can see in person what the idle is doing.

I have to say, i use EGO on, my VE table is set to be almost 100% (ie zero) on idle for correcting to desired lamda, and even though i have a slight problem with my throttle valve being 70mm and allowing some air to pass through,therefore giving me higher idle than what i could get, i now have a specified idle speed of 950rpm, ign.timing of motronic value(my motronic had about 8-10deg advance on warm idle) and the engine feels like stock. No overshoots except one position on warm up at about 45-50C which for other reasons it tends to drop a bit as i go to neutral. But that corrects itself at 55C . it must be related with my warmup enrichments.

[ranz]

Now try to rev the engine and close the trottle to see if it catches it. The pid`s D, is 150 and i see that on most cars if the D is to low, the car got stalling issuses.

I had the same issue with previous FW 1.1.85, but now with 1.1.88 it seems that it also works with low D (previous FW also had boost PID issues), I played with it yesterday a little and finished with D only 9 and it catches idle wery good.

[Reidar]

Let me write you a quick guide The files posted in the vems wiki page is all from me, but not all are good in the new firmware releases.
You are using vemstune i guess?

First off, disable all corrections like the PID controller and spark control. Set you spark table to a fixed advance like 15-18 deg btdc, in the idle area..
Set the refrence point in: "idle control general" to you perfered IDLE speed at warm engine 900rpm or so. This is the reference the Idle ignition advance and PID is working with. so it`s critical you set the perfered idle speed before tuning the idle.

Tune your VE table to a steady value, pending lambda in this aerea will effekt the idle quality.

Now, set your IAC refDC to match your perfered idle speed (900rpm). this is usually a number around 40 on the audis. I have never used the flyback diode on the IAC`s so i set the frequenzy to 25hz, still i geat a dead-steady idle at most cars.

At this point you should expect a pending idle, probably plus minus 50rpm. And the controller might have problems catching the rpm when dropping from higher revs.
Now you only got a fixed opening at your IAC, so the valve itself wont try to adjust acording to the perfered rpm.

Now activate the Ignition advanced control. here are some sugested settings:
Max advance: 8deg
Max retard: 8deg
Rate of change: 10deg
treshold for activating iac: 3 deg

Now you need to keep an eye at you actual Ignition. let`s say your idle advance is 15 deg, and it`s now showing 12deg at the gauge, you need to adjust your refDC position to a lower value, or the other way around. If it`s retarding ignition, your iac value is to high, and if it`s giving more advance, your postition is to low. you should expect it pending plus minus 1deg from your ign table. then all is fine.

Now! activate the PID controller. This will now regulate the IAC so i can be adjuster acording to you conditions. PID tuning is quite hard to get correct, and if it`s done wrong, it will for sure make your IDLE worse.

I`ll give you some starting values for the IAC PID settings:
P: 60
I: 3
D: 150
Decrease limit: 5
Increase limit: 5
deadband: 30rpm

Now try to rev the engine and close the trottle to see if it catches it. The pid`s D, is 150 and i see that on most cars if the D is to low, the car got stalling issuses. Allso check your fuelcut settings, you want the fuel to resume atleast 100-200rpm before the perfered idle speed.

Hope this helps
Reidar

[QuattroFever]

As Vas said, check the ISV. Also the butterfly position/opening of the TB can have influence on the idle RPMs. Have you tried to crack that open a bit?

it idles at about 1100 without any load on it then when you load it up it drops to about 850-950.#

if I drop the ISV min duty to 24 that seems to be the point at which the isv wont make any difference if I take it lower and it idles at about 800 loaded up.

[QuattroFever]

i've seen people do that but didnt see much point. Each point at idle changes the afr by .12-.15 or so. surely that wouldn't make as dramatic a difference as im seeing here?

spose its just a case of putting it all in a spreadsheet and doing the lot in one hit.

[Jamo]

It's not the injector latency, a lot of people use the 630's on the settings you have like Bram and Vas and
http://www.vems.hu/wiki/index.php?pa...sPage%2FReidar for example configs. I have some more stored I can put up but shouldn't need them those ones on there are good examples

I would start by turning off the injector angles until you've got it dialled in. Set all angles to the default 210 for all rev range and see what it's like.

You have the map filter disabled in ecu calibrations under the filter, set it too 1

Your VE table doesn't look right at all to me, values of 60 at idle is very small resolution. I would decrease the req_fuel and increase the ve values to get a better resolution

[prj]

This is exactly what happens if injector latency is set too low in Motronic with big injectors.
But I don't know about VEMS - might have other gremlins.

[QuattroFever]

@a80avant - started with no idle controls on and tried working through

@billman - stock engine with bits bolted on. stock block, stock head cams, gt3071 .63, wagner rs2 evo ex mani, IC and reverse IM. Siemens 630cc @ 3.8bar. 3" turbo back.

@ranz - Need to post a config but says invalid file, will try to post it as a txt when i get home. Maybe you can rename it back to .vemscfg once downloaded.

@prj - not sure on the latency settings, haven't walked that road yet, but was given a basemap from a car running the same injectors as me.

sorry, didnt write it in full. was using my phone earlier.

I tried tuning the PID in the order you described with no ign based idle first. then adding P....under electrical load etc. it seemd to make hardly any difference.

thats why I resorted to turning them all off. If the problem was absolutley no difference then something else must be up.

heres the config in .txt format, ignore the PID settings. I saved the config whilst messing with them.

latest config in txt.txt

I noticed earlier that the AFR goes lean when it bogs. when i blip the throttle to about 2200rpm its rich rich rich until about 1200 then lean = 18.91 till 1000 then rich = 12 then lean again at about 5-600 rpm when stumbling = 18.91 again.

ego correction was not activated till min 2000rpm and fuelcut was set at
cut = 4500
resume = 1600.

it looks like its being starved of fuel on the overrun? but by what if the fuelcut isn't active in that band.

[Bram]

As Vas said, check the ISV. Also the butterfly position/opening of the TB can have influence on the idle RPMs. Have you tried to crack that open a bit?

[A80Avant]

yes ok, you started like that,but once you fix the ref curve, you are supposed to fix the PID in order to avoid all these problems with the less hassle.

the positioning of the ISV valve is also crucial for vems.

If you make the ref curve for the idle as close to the req idle speed, then your PID values should be small to start with, and eventually should remain small, as there is no much need for a large correction.

[QuattroFever]

@a80avant - started with no idle controls on and tried working through

@billman - stock engine with bits bolted on. stock block, stock head cams, gt3071 .63, wagner rs2 evo ex mani, IC and reverse IM. Siemens 630cc @ 3.8bar. 3" turbo back.

@ranz - Need to post a config but says invalid file, will try to post it as a txt when i get home. Maybe you can rename it back to .vemscfg once downloaded.

@prj - not sure on the latency settings, haven't walked that road yet, but was given a basemap from a car running the same injectors as me.

[ranz]

can you please give me some specs of your car?

Original 3B with RS2 ex. manifold, dual plenum revers intake and bigger cooler. Stock injectors, but I'm using 4BAR fuel pressure.
On that conf the boost settings are not correct, I used previously 1.1.85 FW and it had some problems with boost PID settings, so they are not set correctly right now.

[Billman]

Post the config file also, it's a lot better to help you then.
Here is my cars conf right now: http://ranz.planet.ee/1.1.88.zip
I had to make also all kind of changes to get so far, but right now it seems to work good with any condition.

can you please give me some specs of your car?

[prj]

Is your injector latency configured correctly?
I have never tuned VEMS, but on Motronic and some other systems what you describe happens when injector latency values are off.

Does your AFR dip as well? Too rich/too lean? When your RPM's dip?

[A80Avant]

i have been to the stage you are describing and got over it.

have you done the procedure i have described as to how to tune the idle from scratch?

Main issue from what you say is that idle drops on those situations. 2 things come to my mind:

1) your accel enrichment might be too low or high!---> do a log and see momentaraly how much your enrichment is at the point where you press the pedal in (accel). If it is momentaraly getting too lean (or too rich) but usually too lean, then you are getting a dry spot and because of the extra load( all electrics on) you end up with a drop in revs.

2) Your PID setting needs a bit of a tweak while all electrics are operating. I found that on my engine setting, reducing the D to a minimum or 0 made things improve when the engine is under heavy electrical load. BUT that also meant that as soon as the load was off, the idle speed would be settling way above my 900-950 mark, somewhere in the 1050area. That's why the rest of the idle settings should be set properly to avoid that.

If you like, we could arrange to meet at somepoint so i can have a go on your settings live and see how things respond.