[rusEfi] 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Your chance to introduce yourself and your vehicle
ssmith
Posts: 92
Joined: Sun Oct 17, 2021 10:21 pm

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by ssmith »

https://datazap.me/u/scottsmith/2022-01-09-215459-v2?log=0&data=0-1-24-25-28-29

Seems steady at high rpm, choppy at low load. I'm not sure why 9.5s-12.5s has such high swings while 15.6s-20.7s has small swings. It's like it was caught in some sort of ringing/oscillation feedback and had to be pulled out of it to settle back.

If you look at the transitions to high RPM (time=21s and 28s), the high fuel pressure is steady, but low for a while. But at low RPM, the fuel pressure bounces all over the place. I think these two things together mean the peak angle is off. At high RPM the P term doesn't seem to have enough of an impact, but at low RPM the P term is having too much impact. If you look at the cam profile, you'll see that a small percentage delta at low load has a big impact in the # of degrees the pump is activated, but has less of an impact at higher loads. If we shift the peak position, then we'll be riding on a different part of the curve.

I don't want to keep fiddling with the P and I terms because I think the two load points are suggesting different changes. Really these should be set after getting the right peak angle.

Under Controllers/HPFP Cam Configuration, try a test with Angle of first lobe = 10 and = 20.
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Starting is sporadic. Not sure if that's about sometimes not building pressure or else but it either starts or it does not :)

Biggest complaint: we have sporadic FATAL only on cranking. With 12 cranking attempts I got two of those FATAL. Those are _probably_ HardFault or something because TS console connection just disappears, with rusEFI firmware fatal we just get nice red indicator and connection is alive.

Note cute pressure rise on deceleration.

angle=10 log attached
angle=20 log did not save yet another TS glitch can I please make it not today.

IAT temperature does not seem very realistic.
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
ssmith
Posts: 92
Joined: Sun Oct 17, 2021 10:21 pm

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by ssmith »

AndreyB wrote:
Mon Jan 10, 2022 5:10 am
can I please make it not today.
No rush. I didn't see any log file attached, btw.

I did have an idea for a future improvement. It's along the lines of a feature the DTA S100 ECU has. It can measure injector deadtime by changing the number of injections per cycle from 1 to 2 and halving the amount of time. In theory if deadtime is set correctly this should result in no lambda change.

The same idea can be applied to the HPFP cam profile - X% every lobe should be the same as 2*X % every other lobe. If that isn't working, then either the peak lobe position or the cam profile (or both) aren't correct. I wonder if a self tuning mechanism could be made to take advantage of this. There are other factors that affect pump efficiency but it'd be a good ballpark. I bet the ECU could run a set of experiments on an idling engine and come up with a rough cam profile/peak angle. That'd be really helpful for amateurs since it's not like many people want to remove the HPFP and measure the lobe profile or TDC.
ssmith
Posts: 92
Joined: Sun Oct 17, 2021 10:21 pm

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by ssmith »

AndreyB wrote:
Mon Jan 10, 2022 5:10 am
Biggest complaint: we have sporadic FATAL only on cranking. With 12 cranking attempts I got two of those FATAL. Those are _probably_ HardFault or something because TS console connection just disappears, with rusEFI firmware fatal we just get nice red indicator and connection is alive.
Can you run it under GDB / openocd? The HPFP code might be tickling some race condition with the way it's doing scheduling?

And don't overburden your starter! Don't want to burn out another one. Maybe wait a minute between start attempts, and don't leave it cranking too long if it isn't starting on it's own.
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Ops
Attachments
2022-01-09_23.58.41 at 10.zip
(6.38 MiB) Downloaded 489 times
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

ssmith wrote:
Mon Jan 10, 2022 5:17 am
I did have an idea for a future improvement. It's along the lines of a feature the DTA S100 ECU has. It can measure injector deadtime by changing the number of injections per cycle from 1 to 2 and halving the amount of time. In theory if deadtime is set correctly this should result in no lambda change.
hello https://github.com/rusefi/rusefi/issues/3274
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

ssmith wrote:
Mon Jan 10, 2022 5:21 am
And don't overburden your starter! Don't want to burn out another one. Maybe wait a minute between start attempts, and don't leave it cranking too long if it isn't starting on it's own.
Totally planning to be a bit more careful!

Nice used one was only $31 shipped (right) and on this specific vehicle those are easy to swap once you let 13mm and 18mm into your life. Old starter (left) had some rough life.
image.png
image.png (45.02 KiB) Viewed 44359 times
Attachments
20220108_154747.jpg
20220108_154747.jpg (231 KiB) Viewed 44359 times
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
ssmith
Posts: 92
Joined: Sun Oct 17, 2021 10:21 pm

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by ssmith »

Alright, peak position of 10 degrees looks much better. Lets work with that, no need to try 20 degrees.

So funny with high resolution logging to see the individual pumps and injections.

From here I still don't think we should change P/I yet. I think the next thing to do is pump compensation. Transients look like crap because the I term sits at different levels depending what RPM/load you're at. Ideally this shouldn't happen too much, but since we don't know the exact peak nor the exact injector flow rate, the algorithm can't compute things very precisely.

The Controllers/HPFP Pump Compensation table lets us paper over these errors. Basically we want to add the DI:m_I_percent term to the cells based on RPM and fuel mass injected. We have a unit mismatch problem, though; the "Fuel: running" parameter is in mg/cyl, while the compensation table is in cc/lobe. So take "Fuel Running" and multiply it by N_CYLINDERS/FUEL_DENSITY/N_LOBES/1000, or 4/0.72/3/1000=.00185 (I think, I'm just eyeballing this).

TunerStudio or MegaLogViewer can probably automatically build a table given a log file assuming you drive around enough datapoints, but we don't have that since the car isn't really drivable. Since we don't have different load points for a given RPM, we'll just assuming that it's all the same (for now) and build a simple table.

What we do have: https://datazap.me/u/scottsmith/2022-01-09-235841-angle10-part1?log=0&data=1-7-11
At idle, around 900 RPM, m_I_sum_percent is -12.4%
At 1450RPM, m_i_sum_percent is around -7.6%
At 2400RPM, m_i_sum_percent is around 2.5%

So very roughly, we could change the HPFP Pump Compensation table to have:
0 RPM: -5.0
900 RPM: -12.4
1800 RPM: -3.9
2650 RPM: 3.0
for anything higher, have it at 3.0
Repeat the values for all the different load settings (0.005 - 0.120)

With this change, the high pressure reading should be much more stable in transients.
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Problem: half of cranking attempts build HP and start, the other half of cranking attempts pressure is not building and we are not starting. For instance on the log snippet here first and fourth attempts we have pressure build-up, with start on fourth attempt only, second and third no pressure no running. Open question which was is the causation happening here.

My main focus tonight was on the fatal which I've kind of fixed for now so please forgive not great data. To my surprise car has started without pre-warming on OEM ECU with my remote starter button, i.e. without any throttle application! It would stall eventually but that's minor.

https://rusefi.com/online/view.php?msq=818

Battery not in great shape, I need to charge for more experiments tomorrow and I really need to attempt replacing it or just get a new one. For now I will keep it on the charger constantly.


image.png
image.png (68.18 KiB) Viewed 44274 times
image.png
image.png (18.19 KiB) Viewed 44274 times
Attachments
2022-01-12_00.54.45.zip
(2.42 MiB) Downloaded 438 times
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
ssmith
Posts: 92
Joined: Sun Oct 17, 2021 10:21 pm

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by ssmith »

AndreyB wrote:
Wed Jan 12, 2022 6:10 am
Problem: half of cranking attempts build HP and start, the other half of cranking attempts pressure is not building and we are not starting. For instance on the log snippet here first and fourth attempts we have pressure build-up, with start on fourth attempt only, second and third no pressure no running. Open question which was is the causation happening here.
You can try disconnecting the wiring to all the injectors (maybe just unplug the PT2001 board?) and see how long it takes to build pressure. That will help isolate whether it's a pump problem, though if it builds pressure right away that doesn't rule out pump control. If it doesn't build pressure right away it does mean pump control.

There are of course an infinite number of problems it could be, but two spring to mind:
1. Peak angle is still wrong. It seems unlikely this is the major contributor though; the car doesn't build pressure when only requesting 80 degrees of pump, where 120 should be the max. If you wanted a quick hack to test this out, you could change the Controllers/HPPF Cam Configuration so that the max pump angle is, say, 50-60 degrees. I really don't think this is the problem though.
2. VVT is getting in the way. I don't know much about how it works in all cases, but it may be that the cam is changing position at rest, and only going to the right position when there's enough oil pressure. You could try setting the cam under HPFP Cam Configuration (I think it's Intake Bank 1) and then update Angle of First Lobe to 140*, BUT... IMO the RPM and VVT data from RusEFI is not very good. I notice the RPM updates infrequently even though you're on 60-2 - you should be getting more updates per second with that many teeth. The VVT angle has a similar issue - it updates infrequently, which may be a syncing issue in the code, or maybe it isn't designed to update more than once every cam cycle, or ...??? That's beyond my knowledge of the codebase.

Also, there may be a bug in the HPFP code handling VVT where it really should subtract the VVT position instead of adding it, but I can't tell because I'm not sure how VVT position is reported (degrees advance or retard?), and I don't really have faith in the VVT numbers I'm seeing in the data (BMW N20 cam can adjust up to around 70 degrees from fully advanced to fully retarded, yet I'm seeing swings of around 320 degrees in RusEFI).

Code: Select all

        return engineConfiguration->hpfpPeakPos + vvt + next_index * 720 / lobes;
Also, not related to the pump issue but related to starting - the normal fuel calculations take into account the pressure differential when deciding how long to run the injectors. This is important when the difference is small (like at start without pressure build up yet). You have two potential issues:
1. Your MAP sensor doesn't seem to work well.
2. I don't think your fuel high pressure sensor is calibrated correctly. In the logs its common to see the low pressure reading be higher than the high pressure reading, even without the motor running. And if your low pressure sensor is gauge pressure instead of absolute pressure, then it's off by even more.

First rule when trying to get a car tuned is to get the sensors working. You need to do that first. BE CAREFUL WITH HIGH PRESSURE FUEL, it can cause severe injury or death. But if you believe the pressure has bled down (maybe wait until the next morning, the pressure should have subsided by then) then you can check if the sensors read 1 bar / 100kPa or 0. Also get the MAP sensor working - the 2nd log is full of oddities like no change in the value then a bunch of changes then no change. It also goes from 100kPa to 0 in 0.03 seconds then recovers at one point early in the log (around the 4 second mark). Is TunerStudio having trouble keeping up with the logging rate?

Anyways I feel like this is going from a HPFP problem to a "what else is wrong" problem which is beyond my ability to help.

* the angle of first lobe needs to be updated for VVT because the "default" position of the cam seems to be -360/360 degrees. Since there are 3 lobes per cam cycle, that means they are 240 engine degrees apart. If you enable VVT following for HPFP, then it'll add 360 (or -360) degrees to the position, which will immediately make it 120 degrees off from where the code calculated it before. Thus, add 120 to the existing angle if you enable VVT.
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Let me at see what I can do with VVT. The jumps from -360 to +360 are clearly asking for VVT offset, and firmware improvements on top of that to get VVT to show something less noisy.
image.png
image.png (37.68 KiB) Viewed 44230 times
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Engine sniffer tells me that for some reason we are not attempting to engage HPFP while cranking with injection and spark disabled.

tune https://rusefi.com/online/view.php?msq=828
text log attached
Attachments
2022-01-17_23_59_35_399rpm_0_maf_0.0.png
2022-01-17_23_59_35_399rpm_0_maf_0.0.png (36.09 KiB) Viewed 44117 times
2022-01-17_23.55.41.zip
(2.3 MiB) Downloaded 469 times
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
ssmith
Posts: 92
Joined: Sun Oct 17, 2021 10:21 pm

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by ssmith »

Probably because scheduleOrQueue events don't run when ignition is off.

Code: Select all

void TriggerScheduler::scheduleEventsUntilNextTriggerTooth(int rpm,                                 
                                                           uint32_t trgEventIndex,                  
                                                           efitick_t edgeTimestamp) {               
                                                                                                    
        // Not sure why we check ignitionEnabled; it's left over from when this code lived in       
        // spark_logic.cpp, but I think it should be removed.                                       
        if (!isValidRpm(rpm) || !engineConfiguration->isIgnitionEnabled) {                          
                 // this might happen for instance in case of a single trigger event after a pause  
                return;                                                                             
        }                                                                                           
Just unplug the connectors, don't bother disabling in software.
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Sounds like it's really time to remove that condition
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
ssmith
Posts: 92
Joined: Sun Oct 17, 2021 10:21 pm

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by ssmith »

AndreyB wrote:
Tue Jan 18, 2022 6:07 pm
Sounds like it's really time to remove that condition
Probably, so long as the check exists somewhere in the ignition logic. Alternatively, just disable injection, but leave ignition on. The ultimate wasted spark mode!
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Thank you Scott I've removed it https://github.com/rusefi/rusefi/issues/3808

See attach the interesting log in binary format operator error sorry.

PS: those high-frequency logs are sweeeeeeeeeeeeeeeeeeeeet thank you @mck1117
image.png
image.png (48.4 KiB) Viewed 44074 times
Attachments
2022-01-18_17.21.03.zip
(1.38 MiB) Downloaded 438 times
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
ssmith
Posts: 92
Joined: Sun Oct 17, 2021 10:21 pm

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by ssmith »

Yay for mlg-converter: https://github.com/karniv00l/mlg-converter/releases/tag/0.50
I'm able to convert binary logs to csv, since I don't want to shell out $$$ for megalogviewer.
https://datazap.me/u/scottsmith/2022-01-18-172103?log=0&data=1-2-4-5-8

Andrey - what changes did you make to the VVT code in the interim? I'm seeing the VVT angle hover around 190-195 now during crank. Before I would see it at +/-360 when the engine was idling. That's a 170 degree swing which is pretty unusual for VVT. This makes me think the VVT logic isn't quite right yet, OR maybe you made some changes so that now an idling engine will have a different VVT reading.

For those that may have to diagnose in the future, here's what I see: For the two main times that pressure built, the pressure didn't really start changing until DI:PumpAngle reached around 100 or so (and the I Term was 30+%, preferably 40%). On the third time, pressure built immediately with a relatively low pump angle, but VVT also swung from +190 to -170 degrees.

Naively, based on the first two attempts, I would say either change peak pump angle, or change the HPFP Pump Compensation table to be 40% at 0 RPM (and not change any other RPM). But given this engine idles fine once it is going with the same peak pump angle and relatively low pump activation angles, so that tells me the problem is somewhere else.

I really think it's related to VVT. I don't think it's just a coincidence that the one time pressure built up really quickly also happened to be the time the VVT went from +190 to -170. Which, BTW, is a 360 degree swing. Cams don't swing 360 degrees without engines going boom.

Please look into VVT support for this motor. Get that working, even at crank. It would be nice to see RPM and VVT position update more than once per engine cycle, since that's 0.4s at 300RPM. The cam looks like it has 4 uneven teeth, that's 8 edges per cam rotation where you have a chance to update cam position (once you've established initial sync, which typically you can do as soon as the crank missing teeth pass by). Looking on ebay, it looks like there is approximately 2 edges per 90 degrees of cam, though it isn't always the rising or always the falling edge that lies approx every 90 degrees: https://www.ebay.com/itm/393682754025?hash=item5ba95219e9:g:oQwAAOSw9CNe9~sK

How does RusEFI deal with finding whether #1 is on the intake or combustion phase when syncing with the crank? I use the "cam level" mode in Link which basically means "sample the cam right after the missing tooth" to determine which phase. As it happens, for the full range of VVT adjustment on my engine, this relationship always holds true. That's why, at least with my motor, it's possible to recompute VVT on each tooth edge. Does RusEFI work similarly? Or maybe it's just a matter of VVT is not well configured on this motor yet?

I think this also points to the (eventual) need to key the HPFP off the intake cam. We knew we'd have to do that, but when we do we'll have to adjust the peak pump position accordingly, and there's no point in doing either until VVT is outputting sensible values.
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Zero changes to VVT logic, just added random offset in order to avoid the 360 degree swing which is caused by

new setting

Code: Select all

	// random number just to take position away from zero
	engineConfiguration->vvtOffsets[0] = 180;
old code just for reference

Code: Select all

static angle_t wrapVvt(angle_t vvtPosition, int period) {
	// Wrap VVT position in to the range [-360, 360)
	while (vvtPosition < -period / 2) {
		vvtPosition += period;
	}
	while (vvtPosition >= period / 2) {
		vvtPosition -= period;
	}
	return vvtPosition;
}
Nick our VAG guy says VVT does not really more if we do not engage the solenoid. Let me look into the weird position on third go, let me see if we can start reading position on every tooth not just on sync tooth.
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Wheel shape as we consume it

Image
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

ssmith wrote:
Wed Jan 19, 2022 4:38 am
I really think it's related to VVT. I don't think it's just a coincidence that the one time pressure built up really quickly also happened to be the time the VVT went from +190 to -170. Which, BTW, is a 360 degree swing. Cams don't swing 360 degrees without engines going boom.
That looks like rusEFI phasing/VVT logic being confused and believing to be in the opposite phase does it?
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

With code fix 360 degree VVT swing is gone. 50 degree artifact remains and I will attempt addressing that with longer sync recognition pattern not today.

First log cranking with injection/ignition disabled, second log cold start. Would not stay at first after a few attempts got warm enough. This battery is bad and so far I do not control alternator that's not helping, battery swap still on the TODO list.
image.png
image.png (52.46 KiB) Viewed 44038 times
image.png
image.png (54.22 KiB) Viewed 44038 times
Attachments
2022-01-19_01.10.42.zip
(4.6 MiB) Downloaded 458 times
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Shit Passat got towed away :( It was parked a bit iffy and luck has run out :(
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Red old, green new. Green does not have injector wiring yet but has a couple of extra wires. If anyone spots a weird difference please let me know :)

All wiring info at https://github.com/rusefi/rusefi/wiki/VolkswagenPassatB6
Attachments
old-mre-adapter.jpg
old-mre-adapter.jpg (680.62 KiB) Viewed 41023 times
new-proteus-adapter.jpg
new-proteus-adapter.jpg (621.59 KiB) Viewed 41023 times
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

notes to myself
gdi-pin 90 orange - 3neg - VAG17
gdi-pin 94 green - 3pos - VAG32

gdi-pin 92 blue - 4neg - VAG16
gdi-pin 96 yellow - 4pos - VAG48

gdi-pin 114 white - 1neg - VAG2
gdi-pin 118 blue - 1pos - VAG33

gdi-pin 119 red - 2neg - VAG1
gdi-pin 121 black - 2pos - VAG47
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

cold
BMW injector box to begin with
https://rusefi.com/online/view.php?msq=1055
https://rusefi.com/online/view.php?log=870

does MAP not moving look right? do I have cold engine tune problem or MAP problem or else?
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

Added some cold fuel and it runs with BMW injector box
image.png
image.png (61.18 KiB) Viewed 40897 times
https://rusefi.com/online/view.php?log=871
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

I have two wrong wiring diagrams for my WVWAK73C16P101397 test mule at https://github.com/rusefi/rusefi/wiki/VolkswagenPassatB6 and zero matching diagrams.

In real life I have alternator wire black/brown one pin away from CAN wires, CAN wires are at same position on both versions.

"Early" version shows green wire 64 which I do not have and alternator wire one pin away from CAN like I have.
image.png
image.png (45.45 KiB) Viewed 40600 times
"late" version does not have green which I do not have but alternator wire is two pins away from CAN unlike my reality.
image.png
image.png (49.01 KiB) Viewed 40600 times
Attachments
WVWAK73C16P101397-big-plug.jpg
WVWAK73C16P101397-big-plug.jpg (616.64 KiB) Viewed 40600 times
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2006 Volkswagen Passat 2.0 Turbo GDI test mule #64

Post by AndreyB »

rusEFI power off is messed up, Lua and CAN might actually help me a bit
image.png
image.png (5.97 KiB) Viewed 40596 times
no idea what is 050h packet but it seems to only be alive when ignition key is inserted, let me try that.

Code: Select all

canRxAdd(0x050)

shallSleep = Timer.new();

-- we want to turn on with hardware switch while ignition key is off
hadIgnitionEvent = false;

function onCanRx(bus, id, dlc, data)
	print('got CAN id=' ..id ..' dlc=' ..dlc)
	id11 = id % 2048
	if id11 == 0x050 then
		-- looks like we have ignition key do not sleep!
		shallSleep:reset();
		hadIgnitionEvent = true;
	end
end

function onTick()

   if hadIgnitionEvent and shallSleep:getElapsedSeconds() > 3 then
     -- looks like ignition key was removed
     mcu_standby()
   end
end

Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
Post Reply