Audi 90 with AAN/GTX3071 - Tuning Time

[Marty]

This last weekend I tried many things to troubleshoot my weak cylinder(s), but I've come to a conclusion that maybe the individual power feature is not very useful at idle, and 0 through low load accelerations.
(I never went for a drive because I have some body work in progress, thus all of my testing was done under the above conditions.)
What I observe when I start the car is that IP0,1,2, and 3 show various mostly positive values during cranking and afterstart. Once afterstart is over, IP 3 goes to -512, and IP4 goes to +508 and stays steady at idle.
- I'm confused about why this 'switch' would occur in general, it just just seems strange to me. 

Now, I believe I understand the Individual Power concept pretty clearly and I have an observation from my tests that I don't understand: when one IP is weak, both the adjacent IP's go to +508.
(For example in my case: IP3 shows -512, then IP2 and IP4 both go to +508).
    In this case I can understand why IP4 would go up, but why IP2?

Anyway in my testing, I did all of the following:
- traded out each ignition coil with a new one with no positive result
- attempted with 2 new ignition coils just in case there were 2 dead, or one of my replacements was also dead
- verified every single wire for both my injectors and ignition coils. I looked for proper wiring, shorts to ground and shorts to any other lead (at the EC36 connector)
- I swapped positions of my fuel injectors to see if it moved the position of the dead Individual Power output (test for bad injector)
- I replaced my spark plugs
- I replaced my spark plug wires

Much to my surprise and befuddlement, Not one of these things affected the Individual power output in any way... leading me to my conclusion about load above. The only other thing I can come up with is that I have a longer duration cam (from an Audi 7A) installed, and I know this can lead to lopey idle speeds (and thus strange power output values).
Once I get my body work done I'll do some more research with load on the engine. Unfortunately I'm still unable to determine which cylinder it is that's having problems. I read on the forums that there is a direct mapping (shown below), but I was not confident about this being the hard rule with such 'configurable' software. I was hoping someone could assure me(?).
For the record, I tried pulling injector connectors and observing as well as pulling plug wires and observing but the IP graphs and pillars were so 'all over the place' that I could not make heads or tails of it.


    Ign_out0 ->individual power2
    Ign_out1 ->individual power3
    Ign_out2 ->individual power4
    Ign_out3 ->individual power0
    Ign_out4 ->individual power1

[Marty]

Not to take away from the post above (still interested in responses), but how does the individual power measurement account for flat spots in the trigger wheel?

[Marty]

http://www.vems.hu/files/Marty/v3.3_u004574-2013.06.18-19.07.18.vemslog

I've been driving my project car around for a few days now and it's pretty fast  . Having a good time getting to know it, but now some more tweaks are in order.
I still have a few questions that I could use help with :
1) Is this the rule written in stone for individual power?:
    Ign_out0 ->individual power2
    Ign_out1 ->individual power3
    Ign_out2 ->individual power4
    Ign_out3 ->individual power0
    Ign_out4 ->individual power1

I need to know this so I can figure out which cylinder is problematic.

2) I have a GReddy Profec B boost control valve, but it doesn't seem to be working properly. Seems I am operating on spring pressure alone. Can I use VEMS to control this valve successfully? All the wiki discussion seems to be research related and not entirely conclusive. Can anyone help me set this up (or understand how)?

3) last question on individual power: for some reason 1 cylinder seems to do the exact opposite of all the others (ex. When 4 cylinders spike to the top, 1 cylinder drops to the minimum); its like it mirrors all the others. What could this mean?

[Marty]

Perhaps I can make one of my questions more explicit
ALWAYS TRUE   <or>    SOMETIMES TRUE    <or>    FALSE?
    Ign_out0 ->individual power2
    Ign_out1 ->individual power3
    Ign_out2 ->individual power4
    Ign_out3 ->individual power0
    Ign_out4 ->individual power1

[gunni]

As I don´t know the ins and outs of the individual power stuff I can´t really comment,

But what tests have you done with  your profec valve so far ?

The vems can run it assuming it´s a 2pin solenoid.

[Marty]

As far as boost control, I had to change to another injector channel because I'm pretty sure I smoked the mosfet on my first attempt...
Now reinstalled, and operating with proper flyback protection, but configuration is my problem.
Physical testing:
- I have verified the valve clicks fine using a 9v battery
- Disassembled the valve and cleared out any debris in the valve added some lubrication
- Verified installation (while closed the compressor housing feeds the waste-gate diaphragm, when open pressure vents to atmosphere)
VEMS Testing:
- I've tried generally messing around with values but I don't really know what I'm doing.
- I have the map set to allow 250 to 280kpa at all rpm ranges, but I never get more than 190kpa, so I assume the valve is working overtime
There seem to be a lot of options for controlling this and I don't understand what each means.
Ideally, I want to put boost values on the map with RPM and MAP as the vertices and I don't care about anything else. Obviously I would try to tune the PID to get rid of flutter and overboost, but I don't think I'm even near that point yet.
Does anything stand out on this config page?

http://www.vems.hu/files/Marty/BoostSettings.vemscfg

[Marty]

I'm wondering about something in my config; I still have a hard time grasping the Injector and Ign outputs but if someone can answer my question here, that will open the door to my understanding the rest.
- I have a sequential setup,
- my injector channels are:
    cyl[1] = inj[1]
    cyl[2] = inj[2]
    cyl[3] = inj[3]
    cyl[4] = inj[4]
    cyl[5] = inj[5]

- similarly, my ignition channels are:
    cyl[1] = ign[0]
    cyl[2] = ign[1]
    cyl[3] = ign[2]
    cyl[4] = ign[3]
    cyl[5] = ign[4]

When I look at this it seems wrong, but due to some things I think I know , I'm confused about what the correct configuration would look like.
Note: the car runs, I've been driving it, and I believe the ignition channels are correct (I was helped with these previously); so my question is 'how should the injectors be configured ?

[Marty]

I just edited the image I attached, with cylinder #'s on the ignition side. After looking at it again with those it looks correct to me.
I am still investigating why I have one cylinder that appears completely dead most of the time on the Individual power logs, but I am still highly suspicious that it is simply appears that way due to the flat spot on my trigger wheel. No one has yet been able to confirm or deny it.

[mattias]

In a coming VemsTune version this will all be a lot clearer.  The information displayed using the "real cylinder order" will be much more useful and will reflect which cylinder belongs to which individual power and knock gauge.  You could actually have used that to your advantage instead of putting text in the screenshot.

To make more sense of what you've got it's my opinion that you have set it up correctly. Now you can tweak the injector angle curve to find the idle and low rpm angle that gets you the richest lambda reading (until you correct = lower the VE table that is).

As far as your trigger, if the rpm reading is still jagged there is definately something bad going on. It should throw trigger errors if it misses any teeth, but it apparently does not? If it's not another "misaligned primary trigger sensor vs starter ring gear" I would look into using other values for the "divbyX" primary trigger setting, what are you using now?

[Marty]

I have a 60-2, not a flywheel trigger, and I have verified the alignment is good.
Sorry, I'm not familiar with VEMS jargon to know exactly what value divbyX is; is this what you're referring to?
 

[Marty]

I'm still banging away at this project, and it's running much better power and smoothness now.
I'm troubleshooting boost now and I'm hoping for some thoughts on this:
I'm using a GReddy boost solenoid, and it's plumbed as follows
Common (Com)            -> Turbo outlet pressure
Normally Open (NO)     -> bottom of the OEM audi Wastegate with 1.9bar spring
Normally Closed (NC)   -> Atmosphere

What's happening is that I'm hitting peak boost (235 kpa in this case), and it immediately tapers off. I see this behavior pretty much no matter what my PID settings and RefDC are.
To trouble shoot, in the linked log, I just set min and max DC to 70 for one pull, and then 80 for a second pull. There is a minor difference in the 2 pulls but regardless of that, what I see does not reflect what I expected.
http://www.vems.hu/files/Marty/v3.3_u004574-2013.11.05-14.23.49.vemslog
Is this a "haha, your problem is obvious" situation or am I missing something in the configurations or what could be going on here?
Note: I've tried gunni's write up settings, and adjusting those here and there. I've also tried various valve frequencies and so on. when max everything out I manag to hit 270 kpa before "boost taper" (as I'll refer to it), but nothing has gotten rid of it so far.

Could it be that it's exhaust pressure opening the gate?

Also a software question: When I look at the RefDC table in a running VEMSLog, only the top RPM row is used, for some reason it doesn't seem to take MAP into account. In other words the little round highlight that traces around the table always stays in the top row(?).

[GintsK]

For me seems you have two issues:
some mechanical. And configuration.

Last week I tuned RS2 replica with ~1.5bar spring@0%DC.  And hit 200KPa @20..25% DC. This is usual mark. Not 80%.
You have 80%DC and no boost. So I suggest to check your valve plumbing again. This boost shape is typical when plumbing is wrong.

 Then - is your valve really acts. Few weeks ago I had situation where two different solenoids by wiring mistake was connected in series. Electrically. And this gave very narrow valid control range somewhere at 50%DC. Otherwise solenoid was either open or closed.

Second one is settings.
RefDC table at your spring boost should be close to zero. And then sloppy go up.
Boost ctrl works from second row. Not top. Because your boost target is 260 all the time. Fill boost vs TPS table to solve that.

You can start with PID 120 70 70 with limit 100. And probably finish with it. Strategy itself is close to perfect and allow to keep target extremely well.


Beside boost problem you have bad triggering problems!

Gints

[Marty]

Thank you, I'll verify the plumbing, check the valve and make the changes to my configs.
I'm laughing, because the Boost vs RPM table item is so obvious, for some reason it never occurred to me. 

With your familiarity with the Audi 20vt motors, could you verify my plumbing detail plan:
Common (Com)            -> Turbo outlet pressure
Normally Open (NO)     -> bottom of the OEM audi Wastegate with 1.9bar spring
Normally Closed (NC)   -> Atmosphere

The main reason I ask is a side topic: I can't figure out from the plan above how Exhaust pressure to the wastegate is overcome. Or in other words when the BC solenoid duty cycle is high, the pressure in the wastegate should go to atmospheric. Now you have exactly spring pressure which opposes valve open due to exhaust pressure with 1.9bar...

 - I don't know how much exhaust pressure builds up, I imagine it depends on turbo and exhaust size. Perhaps its negligible, but I thought part of the system design was when all other systems fail, exhaust pressure is the final mechanical safety(?)

-I know I need to resolve my trigger error, but I don't spend much time at 6500 rpm yet, so it's not at the top of the list 

[GintsK]

With your familiarity with the Audi 20vt motors, could you verify my plumbing detail plan:
Common (Com)            -> Turbo outlet pressure
Normally Open (NO)     -> bottom of the OEM audi Wastegate with 1.9bar spring
Normally Closed (NC)   -> Atmosphere

Ouch! This is NOT right!  COM goes to WG bottom. NO to pressure. NC to atmosphere. Idea is to make possible to feed full pressure to WG and also fully vent it to atmosphere. Now your WG chamber is either closed or connected to pressure. So - you collect pressure and that's it - chamber stays pressurized and boost is low.

Gints

[Marty]

Well, I've spent a heck of a long time troubleshooting the cause of really odd individual power readings and it seems like I have been over, and improved every detail of my setup at least twice, all with NO affect whatsoever on the problem.
Last night I concluded it must be my main trigger configuration.
Previously I thought this couldn't possibly be wrong because someone trusted provided me the config and viola! it solved my former no start problem. (It got me over the hump of "first time startup on VEMS" which I was having trouble with).

Last night I took a deep analytical look at the config, and I'm surprised that the current configuration hasn't blown my engine to pieces (or more likely melted it haha)!
btw, no discredit to the person who helped me(!), I believe I took it for granted that I didn't need to tweak the base configuration

I would like to run this by folks, because I have a bit of confusion still.
Basic parameters are:
- 5 cyl
- VR Primary Trigger
- Crank pulley mounted 60-2
- 60-2 missing tooth location is centered with the TDC timing mark on the pulley
- VR Sensor is located ~17 teeth before TDC (~102 degrees)

Here is the config given which allowed the engine to start and run (This is hard to believe, for me!)


I did a timing test this morning with ignition locked at 0 and I'm showing about 30 degrees retard. I'm surprised it's even that close to correct given the config above(!?)
Here is my proposed configuration but it seems there may be some special Juju required to configuring for a 5 Cylinder (which I may not have!). -Any reason this shouldn't work?