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

Your chance to introduce yourself and your vehicle
ssmith
Posts: 91
Joined: Sun Oct 17, 2021 10:21 pm

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

Post by ssmith »

AndreyB wrote:
Sun Jan 02, 2022 6:19 pm
ssmith wrote:
Sun Jan 02, 2022 5:42 pm
exposing the dram locations for the hpfp control, adding code to set them, adding config variables, etc.
i need help with this part. Not sure if Chris @Noxz has noticed our progress yet :)
I think it's really easy actually - the addresses are in dram2.def; you have to offset them by 64 since it's channel 2, then add them to mc33816_memory_map.h. The rest is copy pasta. You'd want to double check with real hardware that the dac scaling is the same so 5 really means 5A and not something else. According to the schematic, the sense resistors have the same value.

Too bad dram2.def has semicolons in it, otherwise the syntax would be usable as a C header file.
ssmith
Posts: 91
Joined: Sun Oct 17, 2021 10:21 pm

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

Post by ssmith »

kb1gtt wrote:
Sun Jan 02, 2022 10:32 pm
In my scope capture, the vertical scale is questionable. I don't trust it because I didn't calibrate verify it first. It was a quick measure and seemed to be in the ball park. I think Matt's capture says 2 amps. The big thing is that I think a 4.5 amp driver will be ok.
The average current may be 2A but that doesn't mean if it's driven by a steady 12V supply it'll only draw 2A. The graph you linked to before clearly shows current controlled behavior. Can't do that with the standard MRE output unless you add a sense resistor and logic to read it.
User avatar
AndreyB
Site Admin
Posts: 12800
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:
Sun Jan 02, 2022 10:51 pm
I think it's really easy actually
please please please PR, any help is moving all this forward
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: 12800
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 »

We have progress at https://github.com/rusefi/rusefi/pull/3736

Does Passat B6 wiring diagram https://github.com/rusefi/rusefi/wiki/VolkswagenPassatB6 with pressure regulator wired into +12v tells us not peak-hold on OEM ecu?
image.png
image.png (26.03 KiB) Viewed 3964 times
image.png
image.png (23.39 KiB) Viewed 3964 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
mck1117
running engine in first post
running engine in first post
Posts: 1272
Joined: Mon Jan 30, 2017 2:05 am
Location: Seattle-ish

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

Post by mck1117 »

It could still be peak-hold. That just means it has no boost phase. The injector driver goes boost (high voltage) -> peak (low voltage, higher current setting) -> hold (low voltage, lower current setting). The pump could still do the latter two phases while skipping the first one.
User avatar
AndreyB
Site Admin
Posts: 12800
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:
Sun Jan 02, 2022 5:42 pm
2. Andrey needs to wire his HPFP + and - to the PT2001 board he has at the right spot (J12?)
Any way to avoid changes to OEM harness? I guess I can find a spare solenoid plug and bypass OEM harness in this specific area but OEM unmodified harness would be nicest if possible?
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: 12800
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 »

Another though: we have diagnostics code in TLE8888 driver maybe it's time to actually use it. firmware Fatal in case of overtemperature?
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: 91
Joined: Sun Oct 17, 2021 10:21 pm

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

Post by ssmith »

mck1117 wrote:
Mon Jan 03, 2022 7:22 am
It could still be peak-hold. That just means it has no boost phase. The injector driver goes boost (high voltage) -> peak (low voltage, higher current setting) -> hold (low voltage, lower current setting). The pump could still do the latter two phases while skipping the first one.
True, but if + is permanently tied to 12V, and - looks like the yellow line here: https://rusefi.com/forum/viewtopic.php?p=38300#p38300
then that says to me it isn't peak/hold.

Eyeballing the trigger, this looks like 775 RPM - 155ms for two cycles yields 1/(155/2. / 1000. / 60.)
The solenoid is on for around 6ms - that's 28 degrees (360/(155/2.) * 6)
So we're in the ballpark.

Andrey - I think you could just connect the - of the solenoid to the MC33816 board. It holds the negative pin to ground during the entire pulse. The mosfet on the GDI board has an Rdson of 25mohm, while the TLE8888 (is that what you're using) is 350mohm.

Are you sure the HPFP solenoid is 2.2ohm? I know the injectors were around 2.2, seems odd the solenoid would be that low too if it's driven in saturated mode.
ssmith
Posts: 91
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 03, 2022 8:22 am
Another though: we have diagnostics code in TLE8888 driver maybe it's time to actually use it. firmware Fatal in case of overtemperature?
Never fatal. Try to cope, throw warnings, go into limp mode, etc. Never fatal.

As Matthew has pointed out many times before, you can run a GDI motor without HPFP with less power. Your injection cycle is limited to 3/4 a stroke (can't inject during the last half of the compression stroke because there isn't enough fuel pressure to overcome cylinder pressure), and the flow rate is reduced by 6x (sqrt of (200 bar / 5 bar)), but you can still limp along.
ssmith
Posts: 91
Joined: Sun Oct 17, 2021 10:21 pm

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

Post by ssmith »

Andrey - maybe you can bench test the driver without running the motor. Set the output pin that drives the HPFP to PWM, 12.5% duty cycle, 20Hz. That mimics approximately 30 degree on time every 240 degrees at 800RPM. Watch it on the oscilloscope (while driving the HPFP) and see if it shuts off at some point. Maybe point an IR gun at the TLE8888 chip while doing this.
User avatar
AndreyB
Site Admin
Posts: 12800
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 do not like this result so I would like to declare it inconclusive for now :) Because I can totally hear things clicking.

I should get MRE on my Miata for reference and bench test injectors.
EngineState: confirmation_hpfpbench:9
EngineState: Running bench: ON_TIME=20000 us OFF_TIME=500000 us Counter=3
EngineState: output on ext:TLE8888.OUT6
EngineState: Diag says short_to_bat overtemp
EngineState: Diag says short_to_bat overtemp
EngineState: Diag says short_to_bat overtemp
EngineState: Done!
but overall build-in diagnostics is the way to go I wish we could have smart drivers on more of our boards.
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: 91
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 04, 2022 2:47 am
I do not like this result so I would like to declare it inconclusive for now :) Because I can totally hear things clicking.

I should get MRE on my Miata for reference and bench test injectors.
EngineState: confirmation_hpfpbench:9
EngineState: Running bench: ON_TIME=20000 us OFF_TIME=500000 us Counter=3
EngineState: output on ext:TLE8888.OUT6
EngineState: Diag says short_to_bat overtemp
EngineState: Diag says short_to_bat overtemp
EngineState: Diag says short_to_bat overtemp
EngineState: Done!
but overall build-in diagnostics is the way to go I wish we could have smart drivers on more of our boards.
How long did you leave it running? I wonder if at some point it shuts off due to overtemp, while at first it just runs in current limit mode.
Also you should set ON_TIME=6000us and OFF_TIME=44000us to mimic real load at idle. Counter = ... ?? maybe 1000?
User avatar
Dron_Gus
contributor
contributor
Posts: 327
Joined: Wed Nov 13, 2013 1:11 pm
Location: S-Pb
Github Username: dron0gus

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

Post by Dron_Gus »

Short to bat means your load is too heavy for this output.
Overtemp is it time to add thermal pad from pcb to bottom half of case. And from 8888 to top part?
ssmith
Posts: 91
Joined: Sun Oct 17, 2021 10:21 pm

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

Post by ssmith »

Dron_Gus wrote:
Tue Jan 04, 2022 7:51 am
Short to bat means your load is too heavy for this output.
Overtemp is it time to add thermal pad from pcb to bottom half of case. And from 8888 to top part?
It might be that the thermal design is fine but we're expecting too much from it.

Overtemp may be a function of the protection mode of the output. Is it set to current limit or shut off? If it's set to current limit, then the device itself will generate a lot more heat as it dissipates extra power to bring the current down to the acceptable range.

There's also the heat loss for switch on/off time, but maybe that's minor even at 20Hz (up to 7us each way).
mk e
Posts: 474
Joined: Tue Dec 06, 2016 7:32 pm

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

Post by mk e »

Ok, enough whining for 1 thread....please just add an external solid state relay and get on with it already.... hell.... I'll send you 1 if it will stop the whining! :roll:


seriously though, looking forward to see it it work...so add an external driver already :lol:
User avatar
AndreyB
Site Admin
Posts: 12800
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 »

Going through the wheel well seems to be the better way to replace starter. Ops this car has a security wheel lug nut on each wheel. Ops first butt-welded bolt has snapped trying to unscrew the bolt?!

Second bolt welded to stud of first bolt did the trick, I've removed the wheel.
Attachments
IMG_20220108_150725.jpg
IMG_20220108_150725.jpg (559.75 KiB) Viewed 3674 times
IMG_20220108_144209.jpg
IMG_20220108_144209.jpg (525.68 KiB) Viewed 3674 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
Abricos
contributor
contributor
Posts: 843
Joined: Mon Aug 18, 2014 12:32 am
Location: Carteret, NJ 07008

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

Post by Abricos »

security wheel lug socket in glove box ...
User avatar
AndreyB
Site Admin
Posts: 12800
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 »

Abricos wrote:
Sat Jan 08, 2022 10:28 pm
security wheel lug socket in glove box ...
I am sure it is! Can you please help me locate it?
Attachments
20220108_174157.jpg
20220108_174157.jpg (914 KiB) Viewed 3664 times
20220108_174206.jpg
20220108_174206.jpg (1.17 MiB) Viewed 3664 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: 12800
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 »

great news! i have external vnld5090 clicking HPFP that's a 13A mosfet

bad news I have RPM, VVT position and no fuel injection or dwell events whatsoever on master

https://rusefi.com/online/view.php?msq=815
https://rusefi.com/online/view.php?log=561
image.png
image.png (71.57 KiB) Viewed 3647 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: 12800
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 think rusEFI has GDI as of today.

Jared, Chris, Nick, Matt & Scott - all that work got is to https://rusefi.com/online/view.php?msq=816 and https://rusefi.com/online/view.php?log=562

That's with external vnld5090 driving HPFP

Attachments
IMG_20220108_203723.jpg
IMG_20220108_203723.jpg (843.84 KiB) Viewed 3629 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: 91
Joined: Sun Oct 17, 2021 10:21 pm

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

Post by ssmith »

AndreyB wrote:
Sun Jan 09, 2022 4:20 am
I think rusEFI has GDI as of today.
Awesome! If you're willing to rewire the hpfp to some random external mosfet board, you should be able to rewire it to the pt2001 board. From what I remember, if you don't want current control you should be able to just rewire the negative lead, and leave the positive one connected as is to the fuse.

I think the pressure isn't very stable at low RPM because you don't want to run the pump every lobe. There's a picture in this thread with the stock pump only running once per cam revolution: https://rusefi.com/forum/viewtopic.php?p=38300#p38300
The current algorithm actually runs often enough (200Hz = once/lobe at 8000RPM) to control each lobe, but I think for smoother control it maybe should be slowed down, and instead the driver code should understand "X% every Y lobes" instead of just "X% every lobe" like it is now. That might mean the pressure signal needs a low pass filter on it (similar problem to MAP sensor where you only want to capture the value right before you activate the pump?)

EDIT: ah I see the log, I'll take a look
User avatar
AndreyB
Site Admin
Posts: 12800
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:
Sun Jan 09, 2022 5:17 am
There's a picture in this thread with the stock pump only running once per cam revolution
I remember that picture, I still have that oscilloscope and still can entertain you with similar pictures if that would help :)

Now that we have some high pressure control it's really the question of how much more effort would you be willing to put this in what direction. I am available to be the welding-bolts-to-replace-starter but for this GDI effort I have no brain to spare.
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: 12800
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:
Sun Jan 09, 2022 5:17 am
EDIT: ah I see the log, I'll take a look
Please note that's with the wrong https://rusefi.com/online/view.php?msq=816&dialog=hpfpCamDialog

Just hard-coded desired curve and I think next step is to fix that CLT sensor :) https://github.com/rusefi/rusefi/commit/bd69941ec4121f360284ce934ffd8004c5c69b0a
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: 91
Joined: Sun Oct 17, 2021 10:21 pm

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

Post by ssmith »

1. Change the tune; under Controllers/HPFP Pump Configuration, set
Proportional gain = 0.01
Integral gain = 0.0003 (30x the current value)

2. Change the logging to only record RPM, TPS, all the DI: variables, and Fuel Pressure High. That should increase the data logging rate.

3. Log in ASCII

4. Run a test at idle and 2k-3k ish RPM.

btw why does the MAP sensor read 100kPa at idle? It also reads 100kPa when the car is off. Maybe the input is wired wrong? Kind of amazing it runs at all.
ssmith
Posts: 91
Joined: Sun Oct 17, 2021 10:21 pm

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

Post by ssmith »

AndreyB wrote:
Sun Jan 09, 2022 6:31 am
ssmith wrote:
Sun Jan 09, 2022 5:17 am
EDIT: ah I see the log, I'll take a look
Please note that's with the wrong https://rusefi.com/online/view.php?msq=816&dialog=hpfpCamDialog

Just hard-coded desired curve and I think next step is to fix that CLT sensor :) https://github.com/rusefi/rusefi/commit/bd69941ec4121f360284ce934ffd8004c5c69b0a
Ah that will have a huge impact on low load stability. Please change that too before running the next test.
mck1117
running engine in first post
running engine in first post
Posts: 1272
Joined: Mon Jan 30, 2017 2:05 am
Location: Seattle-ish

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

Post by mck1117 »

ssmith wrote:
Sun Jan 09, 2022 7:23 am
2. Change the logging to only record RPM, TPS, all the DI: variables, and Fuel Pressure High. That should increase the data logging rate.

3. Log in ASCII
Nope, these don't make a difference. Use the beta version of Tunerstudio and you can log every single channel at >500hz.
User avatar
kb1gtt
contributor
contributor
Posts: 3732
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

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

Post by kb1gtt »

A tip, if you put the security lug back on, put all of them on one wheel, like the rear passenger side. Then you mostly don't care about the stupid adapter.

Cool to see progress. Thanks Scott for the recent kick to get it moving along again.
Welcome to the friendlier side of internet crazy :)
User avatar
AndreyB
Site Admin
Posts: 12800
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 »

Today was garage cleanup day tons of thanks to @stefanst for stopping by and forcing me to throw out a lot of crap!

In light of some GDI progress here is a snippet of OEM powertrain CAN bus traffic - next step is at least having tachometer working.

Same as 760 test mule this vehicle has multiple can buses with powertrain traffic not reaching OBD-II port. I've hacked into harness at the ECU connector while attached image helps to hack under the dashboard if so desired.

full trace log https://github.com/rusefi/rusefi_documentation/blob/master/OEM-Docs/VAG/2006-Passat-B6/passat-b6-stock-ecu-ecu-ptcan-idling.trc
image.png
image.png (268.46 KiB) Viewed 3451 times
Attachments
photo_2022-01-09_12-26-14.jpg
photo_2022-01-09_12-26-14.jpg (123.09 KiB) Viewed 3452 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: 12800
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 »

This car has MED9.1 ECU and Funktionsrahmen is one google query away http://nefariousmotorsports.com/forum/index.php?topic=15018.0title=

A lot of overlap with https://github.com/commaai/opendbc/blob/master/vw_golf_mk4.dbc and https://mdac.com.au/2021/04/11/dsg-control-with-rabbit-ecu/ is amazing for transmission integration overview even while mine is not DSG.
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: 12800
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:
Sun Jan 09, 2022 7:23 am
1. Change the tune; under Controllers/HPFP Pump Configuration, set
Proportional gain = 0.01
Integral gain = 0.0003 (30x the current value)

2. Change the logging to only record RPM, TPS, all the DI: variables, and Fuel Pressure High. That should increase the data logging rate.

3. Log in ASCII

4. Run a test at idle and 2k-3k ish RPM.

btw why does the MAP sensor read 100kPa at idle? It also reads 100kPa when the car is off. Maybe the input is wired wrong? Kind of amazing it runs at all.
CLT & IAT added
Proportional gain = 0.01
Integral gain = 0.0003
high-frequency logging

note that sometimes I am failing to start this thing
https://rusefi.com/online/view.php?msq=817
ascii log attached
Attachments
2022-01-09_21.54.59.zip
(8.92 MiB) Downloaded 22 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
Post Reply