E-Throttle Min %DC Limit and a couple of others.

[ayjayef]

Hi Clever people!

I was on the dyno today and ran into 3 problems.  Background: Link Fury, Supercharged bike running E85

 

First one:

ECU Fault Code 70 E-Throttle Min %DC Limit
ECU Fault Code 73 Aux 9/10 Supply Error
Engine speed limited to 1800

I got these when rolling off after WOT power-run.  Key cycle fixed it but it came up again a few runs later too.  I don't understand the Min %DC settings, can someone please shine some light?

 

Second problem:

Using a pressure regulated fuel pump (on AN 2) it shows idle pressure at 394kpa and it ramps up with the plenum pressure to 495kpa at 9900rpm then even though the MAP keeps increasing the fuel noses-over and drops pressure to 350kpa (obviously also gets lean) and then returns to "normal" when I roll off.  You can see the AN2 dip in the middle of the attached picture.

Tested the pump separate using a syringe (to replicate positive pressure on the diaphragm) and it works fine from 250-500kpa so I don't think it's the pump.

Battery voltage is fine before, during and after the pressure dip.

I can't find a MAP limit or any other limit that might do this but maybe I've missed something?

 

Third problem:

It's hard to start, sometimes.  From cold it's pretty reliable but with some temperature in the engine it's often hard to start and I've even had to use some throttle to get it to fire (which is weird on EFI).  Once warmed-up and the enrichment have ended the idle is neat but I've got something wrong in there.  The fact it will start with a bit of throttle might be a clue but I don't know if its starting because the throttle adds more fuel or more air.  I bet it's embarrassingly simple.

 

The bike is making good power,  285hp at 10,000 and will make plenty more if we can get the pump to give it some fuel in the higher range.

Thanks for reading!

 

Andy.

2017 Kawasaki H2-028.pclr

Log 2019-02-14 2;48;29 pm.llg

[cj]

Your throttle faults will be caused because of target->tps actual mismatch errors. Looking in your log we can see that every time you roll off the throttle, its really slow to bring the TPS actual back down. Either your PID numbers for ethrottle control need to be a bit more aggressive, or you need to check that return springs etc are ok and that the throttle blade is not running into some sludge etc in the last few %.

 

Once this counter hits 100 (or 255?) the ECU will kill the ethrottle relay and the rest of the codes all pop up.

 

The fuel pressure drop seems to be directly related to fuel flow so could be any of 

1) lines too small

2) pump too small

3) voltage drop at the pump so its not flowing what it should

[ayjayef]

Wow, thanks CJ.

I didn't know about the TP error accumulator, makes more sense now.  The throttle bodies are working fine, clean with no sludge and return is strong so PID settings makes sense

Current PID is 7.00, 0.145, 60.00

What's a bit more aggressive? 

 

 

Fuel pressure: 

1) lines might be too small, but the logged pressure is at the pump.

2) pump is 320/lph, should be plenty for a 1 litre motor.

3) battery voltage is logged at the ECU okay, I'll see if I can get a voltage reading off the pump when I can get back on the dyno.

[cj]

decrease D by 5-10 or increase I by 0.02 or so. More details here. You want to set up some logging and change the AP targets from 0-100 down to 20-80 while tuning it so you can see if you are overshooting badly or if you are getting oscilations.

 

[ayjayef]

Awesome help, thanks.

[Adamw]

I dont think it is PID related, I would say it is something more mechanical.  You can see it the pic below after you lift off the ECU is almost instantly commanding -90%DC to the motor for about half a second trying to get it to close the last little bit and its just creeping closed slowly.  So the ECU's response is already maxed out - changing PID is not going to give anything more.  You can bump the min clamp up to -100% which will give you a little more juice but it is unusual to need that much. 

So basically the ECU doesnt have enough grunt to push it closed the last little bit.  I dont know if this is something odd with the design - like maybe vacuum or boost on one side of the throttle blade making it difficult or as CJ says something sticky or a spring or cable working against it?

 

[ayjayef]

There is plenty of boost prior (300+kpa MAP) but on roll-off the blow-off valve dumps that to 50kpa near instantly (as seen in the first pic).  This error comes shortly after.

I just got the bike back here (8:30pm) so will have a fiddle tomorrow.  Will let you both know what if anything improves it.

Thanks for the help.

[ayjayef]

So, now I know to look at the TP Error Accumulator% I just logged a pile of 0-100% throttle actions (ignition on, bike not running) and cannot get a single error with snap-shut or slower roll-offs.

Which should rule-out any mechanical return action problem in the throttle body.

e-throttle-test1.llg

[Adamw]

You can see now without the engine running it needs much less DC to pull the throttle closed.  This suggests the pressure differential thing I mentioned earlier is at play.  Try setting up your E-throttle target table something like below, this will hold the throttle open a little more at high RPM over-run which should reduce the pressure differential effect.

[ayjayef]

I would have thought we would see the pressure on the MAP logging but this is an easy fix if it's it.

I've put that into the target table, gotcha is I can't test without dyno time (not sure I can get on the dyno again before racing 2 weeks).  Fingers crossed.  At least I have a few things to experiment with if it happens at the track.

Thanks heaps Adam and CJ. 

[Adamw]

23 minutes ago, ayjayef said:

I would have thought we would see the pressure on the MAP logging but this is an easy fix if it's it.

The MAP sensor is only indicating pressure on one side of the butterfly though - My thinking is since these have a crankshaft driven supercharger and RPM is still high, even with the dump valve open there is likely a large pressure difference inside vs outside the butterfly. 

I have seen a similar but opposite effect when trying to open the throttle.  Some BMW's have butterfly shaft slightly offset from center, so on those if you let them close too much when returning to idle they are sucked shut so hard you cant even move it with a screwdriver or similar.

[cj]

adjusting for boost pressure seems like a fix for the problem you initially reported, but looking at that last log with the engine off, I think you may have tuned the ethrottle a little too aggressively as far as "how much force do I apply to match big changes" and a little too soft on "if i'm close to target, how hard do I try to get that last little bit". You're over shooting the target TPS a bit, and then staying a couple % off target for at least the half a second visible in the log. If this is a race only bike, you'll just have to be aware of this and if it tries to stall when returning to idle, give it a bit of gas (its also not great for throttle motor life span to drive it open past 100% like it looks like its trying to do). If its street driven you'll want to tune it a bit better as I suspect any small (<5% change in throttle will fell a bit lazy)

 

In this screenshot (about 0.5 seconds from the red line where TPS hits 0% to where the yellow line is) you can see the target dropped to 1% when you got off the throttle, but the TPS % overshot this and went all the way to 0%. It then stays at 0% for at least that half second before you give it more gas. If this was trying to return to idle at 1-2% TPS, it would stall.

The second thing i've highlighted is on the TPS main and sub voltages, that the voltage spikes past the 100% level for about 0.25 of a second when you go full throttle then drops back a bit. The ECU sees this whole thing at 100% TPS, so it means you are exceeding the "100%" you saw on your TPS calibration for a bit, but its not likely a calibration error. Bascially your throttle blade/motor is buidling up some inertia and hitting the stops on the TPS while the motor is still applying force. This will shorten the life of the throttle/motor/sensor, but I cant say by how much as its entirely model specific.

The proper fix would be to tune the PID numbers a little more carefully with the engine off until you iron out both these problems, but the quick and dirty fix is just set you AP->TP table to max out at 95-98% instead of 100, and either live with the inconsistent return to idle or add 1-2% to the entire "0" row of the AP-TP table and let it idle a bit higher.

[ayjayef]

It's a race bike, registered for the road but I don't know where you could hope to ride it!

I've previously seen and looked closely at the 0.7 seconds that it takes TPS to go from 0 to E-throttle target of 1 but put it down to "must be just the way the ECU does it"

Earlier in the week the zero row was set to 1.5 but the idle was over 2000rpm so I dropped the target in the 0-2000 to 1.0 to help keep the idle in check.

[cj]

That makes sense. The problem wont be "1.5 is too much", the problem will be the PID settings you have bascially mean "anywhere within 1% is near enough". I bet depending on how slowly you approach 0% target you can have it either hold 1-2% high or shoot 1-2% lower than target.

It only take about 30 min to tune it to pretty much dead on if you are methodical about it and have a reasonable understanding of what each value does. See my post above or the wikipedia article for PID tuning.

[edit] one of either I or D (cant remember right now) is essentially "how long do I tolerate errors", and so this is what you tweak to make it respond to "small errors held for a long time", in a shorter amount of time.

[ayjayef]

That Wiki is tough reading, I did the best I could.

Setting I and D to zero and adjusting P till it flutters, then halve it and add I to stabilise then D to clean it up.

Resulted in P2.0  I.078  D5 but that wasn't good.

With some more tweaking I have ended with P 4.2   I.102  D 35 which has stopped the bounce at return to zero but TPS and APS are a fair way apart.

 

p4.2i0.102p35.llg

p2i078d5.llg

[cj]

Yeah its a bit maths-ey. I do like the GIF they have of the effects of each parameter change on the output graph. That always helps me visualise what i'm doing.

that p=4.2 log looks pretty good to me. Remember to compare the Ethrottle target % against TPS rather than APS -> TPS. the P value being 4.2 is a bit lower than most of the reccomended settings for various ethrottle's in the help file so you may be able to make it a bit more sensitive by increasing P by 1 or 2 and then tweaking the other 2 if needed but it certainly looks usable as is.

"Ethrottle target" is what the ECU is trying to hit, not the AP%, and it seems to follow this ok. The APS -> ethrottle target correlation you are seeing includes our AP->TP table and the offsets that are added through idle base table and similar idle controls. Ethrottle target is after the ECU puts all of this together and decides where it wants the TPS to be.

One thing you can do to check is to temporarily change your AP->TP table so 0%AP =  20% TP and 100% AP = 80% TP. That way you can easily see if you are overshooting the target. a tiny bit of overshoot is ok, and at 0/100% you wont see it, but it will matter when you go 0%->10% as it may jump up to say 20% for a bit then drop back down. Also try moving the throttle slowly and check it follows quickly and doesnt lag, and also run some tests to only part throttle. 

[ayjayef]

That all makes sense.

"temporarily change your AP->TP table" 

Do you mean the "E-Throttle 1 Target" table?

[cj]

Yep thats the one

[ayjayef]

(I was just making sure we were talking about the same thing!)

That was a good tip setting 20% and 80% in that table, with some fine tweaking I have zero bounce on normal throttle roll-off and snapped shut I sometimes get a poofteenth overrun which is corrected in 0.1-0.2 seconds. 

PID settings are now

P4.20

I 0.102

D 30%

Looks like it's going to work much better now.  That's heaps for your time.