Trackday Log Analysis Help

Hey guys, I just went to a trackday with my R32 GT-R and all in all had a great time!

There were a few issues with the car hitting cuts and the like, so I analysed the logs to see what was going on.

I have attached a log from towards the end of the track day where I tried high boost for a little bit, this made the issues worse.

I also attached the map the car was running at the time.

My analysis is that:

1) I am hitting oil pressure cut a fair bit, especially at high oil temperatures (120-140C)

Accusump is there to help, but obviously I need to resolve the oil control issues.

I might try going from a 10-40 to a 15-50, and add shrouding on the two oil coolers I have, as I don't think they are getting good airflow.

I have an enlarged sump with baffling, so if the above still doesn't help with oil control, dry sump might be the only way forward.

2) Voltage is dropping at high RPM high load, and this is occasionally causing fuel pressure drop.

I suspect that this is a bad alternator, so I have ordered a Franklin Performance LS alternator conversion kit.

The Deka ETX30LA battery is in the boot. The fuel pumps are wired in to the positive cable connecting the battery to the front of the car.

3) I have a PPG sequential, and the ECU was not showing correct gear position (-) when at high RPM.

I was also finding that it was not shifting well, especially on high boost. I guess the ECU not knowing which gear it's in wouldn't help?

The voltage offset is only at +/- 0.1V, so I suspect the voltage issues are skewing the gear position sensor reading. If the alternator doesn't resolve this I may also need to increase the offset slightly? 0.2V perhaps? Unfortunately I didn't have the ECU configured for any gear related logging which was an oversight on my part. I have configured this for next time, and I will be able to use the data to fine tune the gear position voltage values

Looking forward to any help you guys might be able to provide me so that I can take this thing to its full potential!

Your gear pos sensor not working wouldnt have made a lot of difference in this case since you have basically the same settings for every gear anyhow.

The gear "volt tolerance" on G4+ is a fine balance since it not only determines both the gear index for shift settings at the start of the shift but also it determines when the ecu considers the next gear is far enough engaged to cal the shift complete. Usually on a 6 speed with ~0.6V between gears the sweet spot is around ~0.15-0.20V.

A few other gear shift settings look a bit odd - your upshift and downshift force is quite a bit higher than normal - often more around 60N when using 100N/V. The start shift validation is usually more like 5ms. If you change these settings and do another log I can help you with the shift tuning from there. Log all the gear-related stuff at 100Hz - RPM, lever force, Ign cut, the raw voltage on AN Volt 2, gear shift status.

Is the ecu controlling the accusump valve? I would think you might get more stable pressure if you keep it open all the time. Increasing viscosity is going to help bring the overall pressure trend higher, but it wont improve the spikes low from oil surge. Looks like it is mostly under braking.

Thanks Adam!

Do these logging settings look good?

Accusump is on Ignition Drive 8.

DI 2 is starter.

So you are saying for VA 3, I should remove the oil pressure condition, and just make it always open above idle?

I did think that the shifting force felt quite high! I have pretty sore hands and was starting to get calluses!

Should I try settings more like this?

I also think the sensor is sitting at about -10N with no load, should I adjust the offset until this reads 0N?

For the Gear pos voltage "AN2 - Gear position" is the only one you need. You dont need ANV2 - Gear pos (status) - this is just whether or not the ANV is in fault mode etc. You also dont need ANV2 - DI or ANV10 - DI - these are only used when the analog input is set up to do a digital function. Log Percentage Ign cut @ 100Hz also.

"So you are saying for VA 3, I should remove the oil pressure condition, and just make it always open above idle?"

Yes, worth a try. I have always only ever worked on cars with the manual valve that you just open by hand before starting. They generally work very well. To me only opening the valve after oil pressure has already dropped significantly is a bit backwards - I have heard of many people doing it, but never seen any data to suggest if it works or not.

"I did think that the shifting force felt quite high! I have pretty sore hands and was starting to get calluses! Should I try settings more like this?"

Yes the high force and long validation time would have made it pretty difficult to initialise a shift. Try those settings I suggested and we will see from the logs if you are getting enough preload. 100Hz logging is actually a bit on the slow side to see enough detail for tuning the shift force but is the best we can do with G4+.

"I also think the sensor is sitting at about -10N with no load, should I adjust the offset until this reads 0N?"

Yes.

Thanks again!

These are my ECU logging settings now, hopefully capturing enough!

You mentioned ignition cut, but the cut mode is set to Fuel % Cut, so I've logged both.

Is it okay to be using Fuel cut for gear shift? Ignition cut would make it go bang between shifts, right? 😁

It's using Ignition cut for the oil and fuel cuts.

Fuel cut for RPM and MAP limits.

Thanks Adam, this is how it's meant to shift! So much better!

I've attached an ECU log.

As you can see below sometimes the voltage is dropping and this is causing the sensor reading to fall outside the +/- 0.15 volt tolerance.

I figure that I should fix the voltage drop issue (new alternator ordered) before I change the Volt X values too much, as that wouldn't be the correct fix?

Keen to hear your thoughts!

Based on my analysis of the log I'm looking to change the volt tolerance to 0.2V, and I tweaked a couple of the gear voltages.

The difference between each gear is precisely 0.56V. Is this a correct assumption to make? Or could the voltage gaps be different?

With these settings it would not have incorrectly detected any gear in that log based on the AN 2 trace.

My only concern is that the torque introduction could be too early and damage the dogs?

That said, would the 20ms torque reintroduction delay provide enough of a window to ensure the gear is selected correctly?

Is it okay to be using Fuel cut for gear shift? Ignition cut would make it go bang between shifts, right?

Ign cut is generally much better. Fuel cut is much slower to remove torque and slower to reintroduce torque. You can often reduce the backfires a lot by adding a positive fuel trim for the shift - say +20%. It would be worth trying ign cut, it makes for a much nicer shift. You can often use less cut in the higher gears - say 100% in 1st reducing to say 80% in top which will help reduce backfires as well.

As you can see below sometimes the voltage is dropping and this is causing the sensor reading to fall outside the +/- 0.15 volt tolerance.

Yeah, you have an usually large voltage variation for a stationary barrel. Is there significant backlash in the drive tang for the position sensor or something? I wouldnt expect it is related to batt voltage as that sensor is supplied with regulated 5V. It could be a ground offset but it looks more mechanical to me since it is most unstable when you are giving it some.

Based on my analysis of the log I'm looking to change the volt tolerance to 0.2V, and I tweaked a couple of the gear voltages. The difference between each gear is precisely 0.56V. Is this a correct assumption to make? Or could the voltage gaps be different?

Yes you can try opening up the tolerance a bit more - 0.2V is more than I normally need when the voltage is more stable. If you go too much it will reintroduce the torque before the dogs are far enough engaged.

Awesome I'll try ignition cut with driven upshift fuel trim of 20%.

I'll leave the tolerance at 0.15 as I definitely don't want to break this box!

I'm not sure how to check the backlash so I'll have to investigate that.

Doesn't this screenshot seem to indicate that the low gear position reading correlates with the low voltage spikes, despite this being at 100% TPS for the whole duration? I guess I'll find out for sure when I replace the alternator, as it would be good to rule that out for other reasons (low fuel pressure cut).

My adjusted values with 0.15V tolerance should precisely consider these lows as the correct gear, but no lower, so I think that should be a good improvement.

Looking at that trace, there is a lot of creep in the logged voltage being sent by the Gear Position Sensor, more than I would be happy with. It has moved enough in couple of points that the voltage has dropped outside of the tolerance window and the Gear Position has gone to the default value. As long as the sensor is getting a clean 5V and 0V signal, this sort of movement should not be happening. Either there is slop in the mechanism driving the sensor, or the sensor is starting to fail. The typical cause of gear sensor failure that I have seen is oil contamination of the sensor.

Thanks BlackRex.

We do have oil leaks which I am looking to resolve, so oil getting into the sensor is a possibility.

What do you think about the Gear Position voltage drop correlating with the battery voltage drop (dark green line).

Does that lend weight to the sensor not getting a clean 5V or ground signal?

Okay, I think I know what is going on.

The PPG comes with a display which is what is providing 5V and ground to the sensor.

The display is powered by battery voltage.

So the issue here is that the ECU is reading a sensor which is not fed by it's own 5V rail, it's piggy backed off the PPG provided display and associated harness.

I suspect the 5V voltage regulation in this display can't keep up with a voltage drop down to ~11.5V and this is dragging the sensor signal down.

Hopefully replacing the alternator keeps the voltage more stable around 14-14.5V, allowing the gear display to drive the position sensor more consistently.

I don't really need the display, so I could also consider bypassing the display and having the ECU drive the sensor directly?

Adam will be able to answer that question better than I can, as he works for Link and knows the characteristics of the regulated 5V rail vis a vis Battery Voltage into the ECU. Typically the 5V rail output should not be affected by changes in the battery voltage (hence them being a regulated 5V output) until the battery voltage gets down below a certain point that the internal regulator cannot operate at. If the battery voltage is getting below this limit then I can see this being an issue, but it would also be affecting all of the other sensors that are powered by that output.

If it was battery related, then you wouldn't have the gradual drop in the voltage being read by the sensor that shows in the 4th gear run.

I have attached a run from my car, as can be seen, the Gear Position voltages remain stable (that's a 10second window between the datum points) even though the battery voltage fluctuates. The drops in battery voltage are when the compressor for the paddle shift system activates.

It is always advisable to have the device that is reading the sensor also powering it, so that you do not have to deal with supply or ground offsets.

I've asked my wiring guy about bypassing that display and running the sensor directly from the Link, as that will eliminate any electrical issues and then it will be clear to see if there are any mechanical problems with the sensor.

Any thoughts on why the battery voltage could be dropping?

My fuel pumps (2x 044) are always on, not staged, the engine fan didn't come on, and I'm told coils and injectors don't have a significant load.

It leaves me thinking alternator regulator is bad? RPM/temp/load related?

Belt tension looked okay so I would be surprised if it's slipping?

Yeah, Im with Stephen here, you want the gos pos sensor 5V and ground to come from the ECU. You will most likely be able to share the signal between ecu and PPG display without affecting accuracy. The Link 5V output is rock solid down to around 6.5V battery.

I suspect your gear pos voltage correlation to batt voltage is probably more related to a variable ground offset between ECU and gear display as chassis electrical load increases - rather than the 5V reg in the gear display, but that's just assumptions.

Thanks Adam, so it could just be as easy as grounding the display in the same location as the ECU ground?

I'll take a look and hopefully it's an easy fix!

No, you want to power the gear position sensor using the ecu 5v supply and sensor ground.

Okay I won't be lazy and I'll get my wiring guy to take a look at doing that 😇

It's shifting so much better now! This is what a sequential is meant to be like!

There was one incident below where the ECU ended the cut prematurely because it lost track that it was in 3rd gear.

I don't want to keep playing with voltages and offsets until I get the sensor voltage stable, as I really don't want to damage the box!

I've attached an ECU log in case anyone would like to take a look to see how the shifting is performing?

Thanks so much for all the help, this has really transformed the car and made it much easier to drive!

That Gear Position Voltage is still an issue, it should remain stable once the gearbox is in gear. That it is dropping as much as it is is either movement of the shaft that is driving the sensor. or a fault in the sensor. The reason that it has dropped back to the default gear is that the voltage for third has exceeded the tolerance range for that position.

I called PPG and they are happy for me to wire the sensor directly to the ECU, so I'll get that done and keep their one-fits-all provided loom in safe keeping 😊

I had a quick look over your log. Agree with Stephen, gear pos signal is still far from acceptable. There does definitely appear to be a relationship to battery voltage, where bat voltage is reasonably stable your gear pos voltage is much more stable - so Im hopeful that probably means the sensor and sensor engagement is mecahnically ok and it is just the bad idea of having the display box power it that is most of the problem.

In general shift tuning it looks like the upshift force is in about the right ballpark now. Upshift Cut is possibly ending just a little early but that is because you are having to run too much tolerance. When you have more stable voltage you can bring the tolerance down.

Some of your downshifts are being incorrectly catagorised as "driven down shift" when they are actually over-run downshifts. This is due to you doing a manual throttle blip that is going above the drive/over-run TP. This means you get a cut which will prevent the blip. Since you dont have E-throttle you want to set this drive/over-run TP to something like 50% so all your downshifts with blips are considered over-run downshifts. Driven downshifts are pretty rare in general.

Awesome, thanks for that Adam!

I had noticed the driven downshifts being detected, so I'll adjust the drive/over-run as you have suggested :)

Hmm, if I set the Drive/Overrun TP to 50%, it would mean part throttle upshifts would not work correctly, right?

This thing is insanely torquey, and even at 20% TP it accelerates very quickly and has enough load to upshift through the gears.

Could I just zero out the Driven Down Shift Cut table?

How about the other tables? Would zeroing out these be okay as well?