Auto-tune does not produce good results

Tuning, troubleshooting and the nitty gritty of using rusEFI to make your engine run nicely!
Post Reply
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Auto-tune does not produce good results

Post by AndreyB »

Nick is asking me how come auto-tune just keeps adding until 180% VE?

I know that we are confident in engine displaced. I believe we trust the MAP sensor. So it has to be about injected fuel amount? Is it reasonable to question the 630cc/min flow rate of some aftermarket injectors sourced in Canada?

Is it time to send those injectors somewhere to get flow benched? Any US shops we like for injector flowing test?

https://rusefi.com/online/view.php?msq=654 note 1.3 CLT correction
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
mk e
Posts: 486
Joined: Tue Dec 06, 2016 7:32 pm

Re: Auto-tune does not produce good results

Post by mk e »

Is auto tune looking at O2 readings? If so then the tune is right and as you say there injector flow is probably wrong.....so do you care what is in the VE table is if its running good? If so do you know where the VE should be on this particular engine? 90% so doubling the injector flow rate would put it just about perfect?
'
What I'm saying is I'm not sure money spent flow testing injectors is money well spent after the fact if you're already up and running and so know the answer from the O2 data.
Abricos
contributor
contributor
Posts: 849
Joined: Mon Aug 18, 2014 12:32 am
Location: Carteret, NJ 07008

Re: Auto-tune does not produce good results

Post by Abricos »

It's only imagination that fuel injector dead time is wrong
At same point you need test all sensor
Idle on this injectors is perfect stable
This is specs of injectors
630cc BOSCH EV14
New Bitmap Image (1).bmp
New Bitmap Image (1).bmp (2.45 MiB) Viewed 15327 times
b08e922fec2c6f6cd83d437edd344778.jpg
b08e922fec2c6f6cd83d437edd344778.jpg (40.83 KiB) Viewed 15327 times
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: Auto-tune does not produce good results

Post by AndreyB »

Nick, I've discussed all our progress with DaveB over the phone since we both did not want to go to be early.

Dave's first suggestion is to add a second MAP sensor definitely with a different calibration ideally different physical layout/shape and confirm match at idle and under boost. Dave say that he has seen physical damage issues which are not displaying themselves under vacuum/idle but only faulting only under boost.

Dave's second suggestion is fuel pressure sensor. Ideally between fuel rail and fuel regulator.

Nick, please post an image of your fuel rail/regulator setup.
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: 849
Joined: Mon Aug 18, 2014 12:32 am
Location: Carteret, NJ 07008

Re: Auto-tune does not produce good results

Post by Abricos »

VW ABA fuel rail
3 bar regulator
Screenshot_20210927-071005__01.jpg
Screenshot_20210927-071005__01.jpg (183.02 KiB) Viewed 15307 times
Screenshot_20210927-070957__01.jpg
Screenshot_20210927-070957__01.jpg (199.8 KiB) Viewed 15307 times
Simon@FutureProof
contributor
contributor
Posts: 413
Joined: Tue Jul 24, 2018 8:55 pm
Github Username: Orchardperformance
Slack: Orchardperformance

Re: Auto-tune does not produce good results

Post by Simon@FutureProof »

oooh, I strongly second the fuel pressure sensor.

If your fuel pressure is dropping due to a lack of flow under boost then you will see the VE wind up as it tries to give more fuel but cannot as the flow rate of the injectors is dropping with the pressure.

This might also explain your 6000rpm "rev limit" as that is a symptom often associated with a bad pump of fuel pressure, you hit a certain amount of flow and then it just feels like you hit the rev limit as the pressure falls off.

What fuel pump do you have installed?
Now keeping MRE in stock in the UK - https://www.FutureProofPerformance.com
Simon@FutureProof
contributor
contributor
Posts: 413
Joined: Tue Jul 24, 2018 8:55 pm
Github Username: Orchardperformance
Slack: Orchardperformance

Re: Auto-tune does not produce good results

Post by Simon@FutureProof »

Are we looking at these injectors:
https://www.bosch-motorsport.com/content/downloads/Raceparts/Resources/pdf/Data%20Sheet_67797771_Injection_Valve_EV_14.pdf

If so, is that 630cc the bosch rating or one the supplier has given?
I ask because bosch rate their injectors using N-heptane which has a different density to IsoOctane/gasoline. This means they may not be exactly what you expect.

Can we get the part number off those injectors?

Why N-heptane -> https://en.wikipedia.org/wiki/Heptane
It is available pure and thus a fixed density for testing and is basically the zero point for RON/MON testing. Gasoline is a mix and not really pure or reliable to rate different injectors over a long period of time.

I am going to assume these are the 429g/min injector as that corresponds to 630cc/min N-heptane.
N-heptane is 684g/litre while gasoline is 743g/litre for E10 (https://www.epa.gov/sites/default/files/2017-02/documents/exhibit-h.pdf)

If your supplier has not done the calculation for CC/min using the N-heptane density then these could be the 503g/min injectors which would obviously cock up the dead time estimates and the flow rate would actually be higher.
Now keeping MRE in stock in the UK - https://www.FutureProofPerformance.com
Abricos
contributor
contributor
Posts: 849
Joined: Mon Aug 18, 2014 12:32 am
Location: Carteret, NJ 07008

Re: Auto-tune does not produce good results

Post by Abricos »

How many HP engine used to rotate crank shaft at 6000 rpm ?
Car not move ...
blundar
contributor
contributor
Posts: 141
Joined: Tue Jan 07, 2014 4:38 am
Location: Cincinnati, Ohio
Github Username: blundar
Slack: Dave B.
Contact:

Re: Auto-tune does not produce good results

Post by blundar »

If I understand you correctly, Nik, you're asking how much energy is needed to make the crankshaft spin before the engine actually starts doing the work of spinning the transmission? Frictional losses, accessory drive, lots of stuff go into this. I am going to go out on a limb and say "single digit" (i.e. 1-9) HP is the answer.
Simon@FutureProof
contributor
contributor
Posts: 413
Joined: Tue Jul 24, 2018 8:55 pm
Github Username: Orchardperformance
Slack: Orchardperformance

Re: Auto-tune does not produce good results

Post by Simon@FutureProof »

As a bench mark it is about 15kw to spin a 600cc engine to 6k. But that is steady state and ignoring what we are seeing here.

You need very little power to maintain the speed of the crank, you need a lot more power to accelerate the crap out the crank in a short period of time.

By the same argument vehicles only need about 13hp as that is roughly all it takes to maintain 70mph.

Take a look at your log below, you hit 83% VE at 5800 rpm with a lambda target of 12.2, to fuel that the log is saying your demanding 55mg/injection, that's roughly 890cc/min total flow.

Also note that the AFR begins to lean out as the RPM begins to drop and then the AFR really falls off as the ECU demands more fuel.
image.png
image.png (81.77 KiB) Viewed 15286 times
Now keeping MRE in stock in the UK - https://www.FutureProofPerformance.com
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: Auto-tune does not produce good results

Post by AndreyB »

OrchardPerformance wrote:
Mon Sep 27, 2021 1:11 pm
Take a look at your log below, you hit 83% VE at 5800 rpm with a lambda target of 12.2, to fuel that the log is saying your demanding 55mg/injection, that's roughly 890cc/min total flow.
What do those numbers mean? 890cc/min reasonable to spin up the crank? 890cc/min completely unrealistic to spin up? Too little? Too much?

Reminder that https://rusefi.com/online/view.php?msq=654&dialog=cltFuelCorrCurve has 1.3 CLT correction across the board are you accounting for that?
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
Simon@FutureProof
contributor
contributor
Posts: 413
Joined: Tue Jul 24, 2018 8:55 pm
Github Username: Orchardperformance
Slack: Orchardperformance

Re: Auto-tune does not produce good results

Post by Simon@FutureProof »

Numbers taken from the log so accounting for all corrections before injection.

It is a medium amount of fuel, if you were traveling at 60 mph it would be 4.5mpg.

It is enough to make it worth checking fuel pressure for drops and for me to look at other logs and see if there is a pattern above that fuel flow level.
Now keeping MRE in stock in the UK - https://www.FutureProofPerformance.com
mck1117
running engine in first post
running engine in first post
Posts: 1493
Joined: Mon Jan 30, 2017 2:05 am
Location: Seattle-ish

Re: Auto-tune does not produce good results

Post by mck1117 »

AndreyB wrote:
Mon Sep 27, 2021 2:16 pm
reasonable to spin up the crank?
Spinning up the crank is no different to flooring it on the highway. Floored is floored, whether you're accelerating just the crankshaft/flywheel/clutch, or the whole car. it would be easier to see a reproduction of this on the street. Hold at 5k rpm in 3rd gear or something, use left foot brake to hold RPM, and slowly add throttle, maybe from 0->100 over a few seconds. A sharp transient like flooring it at idle is pushing the limits of the O2 sensor's response time, so it's hard to see what's a real problem, and what's an insane transient that we don't expect valid lambda information from.
mk e
Posts: 486
Joined: Tue Dec 06, 2016 7:32 pm

Re: Auto-tune does not produce good results

Post by mk e »

mck1117 wrote:
Tue Sep 28, 2021 3:56 am

A sharp transient like flooring it at idle is pushing the limits of the O2 sensor's response time, so it's hard to see what's a real problem, and what's an insane transient that we don't expect valid lambda information from.
and is as much or more a measure of wall wetting effects as actual tune. I don't think you even can autotune with sharp transients like that can you? Left foot dyno (in short 5-10sec runs with lot of cool off time) or the biggest hill you can find to get near steady state. Personally I would turn off autotune and do 2 back to back runs with the lambda changed 10% and see if AFR changed 10%, that would either confirm or close the fuel flow/pressure questions right now with no more wondering. If its still drive way only I don't think there is any way to tune as you just can't separate transient effects....put it on jack stands facing tthe road and try the brake dyno trick, I've seen it done but I hate that idea as dangerously unsafe.
Post Reply