Constant throttle stumble.

[AustrianVespaGuy]

Step 1: Check plugs

Step 2: Run in wasted-spark COP mode instead of sequential COP.

Step 3: Post a datalog.

[ray_]

9 hours ago, jimk said:

Not difficult to give a try.  Run a manual disconnect wire from battery power direct to the B11 ECU power in or alternately to the switched power into the rely box  The ignition switch may have intermittent power that is not noticed when you have your foot in it.

 

Do you still have to turn your heater fan on to get the engine to shut down? 

 

 

I noticed my turn signals on the white car were quite dim. Turns out they were getting maybe 7 volts. Issue was dirty copper in the ignition switch.

 

Dave, am I hearing you are powering your ECU thru that switch?

 

Edited October 2, 2021 by ray_

[irdave]

1 hour ago, AustrianVespaGuy said:

Step 1: Check plugs

Step 2: Run in wasted-spark COP mode instead of sequential COP.

Step 3: Post a datalog.

 

Log of what?

Plugs are the most beautiful light brown I've ever seen in my entire life.  Ever.  They'd make you cry.  All of them.  But I'll look again.

 

 

9 minutes ago, ray_ said:

 

 

I noticed my turn signals on the white car were quite dim. Turns out they were getting maybe 7 volts. Issue was dirty copper in the ignition switch.

 

Dave, am I hearing you are powering your ECU thru that switch?

 

 

No, that's just the switched 12v to turn the ECU on.  Main power comes from the end of the e30 fat cable run up front.  The harness is the Haltech premium harness, which has everything, including relays, already put together- along with shielded wire for crank and cam triggers.  It also has a central ground for everything, central 5v, 8v, and 12v sources- all from the ECU.

 

But this line of thinking had me looking at some stuff.  My 12v switched for the ECU comes from the green double wires from the ignition switch- which is where the previous people tapped in the stock ECU.  One line goes to regular stuff, but the other green line is supposed to go to the coil apparently.  My car hasn't had a coil in decades.  Where does the other line go? (Wife has the weekend off, so we'll have to wait until tomorrow morning to find out...)  So no, I'm not really thinking Jim is onto something with that- like I said, it's just the switch to turn the relay on which powers the ECU and everything therein and therewith and therethere.    And I will look into whatever else is connected to the same place at the end of that battery cable.  I think it's just the relays for the rad fans; could be the main fuse box, too, for the car.

 

Thanks y'all.  I really appreciate it.

[irdave]

3 hours ago, SydneyTii said:

I’d reinstall the distributor just to start from a known base, see if it goes away and move forward from there?

 

Dude.  I was 75% there last week.  Except that I can't get the cap over the rotor, because I installed the MIshimoto radiator with simpler hose routing to the upper part of the rad, and the dizzy wants that same space.  Plan is to trim one end of the hose a little to see if that'll make enough room.  (Dizzy for the s14 is on the front end of the exhaust cam, not off the back, so it's a little tight up there.  But I'm totally with you.  Hence the plea for help. 

[ray_]

OK, so anything powering the electronics is via a relay? 

 

Well, good then. Maybe. 

 

Look at grounds

[John76]

3 hours ago, irdave said:

Where does the other line go?

OK Dave.

Here's your Dr. Zeus scribble for the day.

Best,

John

 

 

[AustrianVespaGuy]

12 hours ago, irdave said:

Log of what?

Whatever your Haltech can take a log of.  Being able to see crank/cam signals, commanded ignition timing, pulsewidths, and AFR are probably the most useful things to look at to start.  But if there's more that's OK too; usually you can filter down what to display pretty easily, so better too much data than too little!

[irdave]

Ok.

 

Installed dizzy yesterday.  Plugs were beautiful, really.  Stumble was better, but still there.  Means COP conversion probably isn't the problem, just better able to reveal the problem.

 

Jim's thinking my new injectors (FIC) aren't dealing with the low pulse width very well.  Pulled out old injectors and made new loom to plug them in.  Find out Haltech has a function to help with this, try that first...  Add a bunch of data channels, go for a drive.  Lots of stumble with the dizzy.  Get home, pull data.  ECU is weird, has a bunch of old data- lots of redline pulls- which I didn't do this morning.  Pull IAT trace to verify; it's all over the place with drop outs everywhere.

 

For the umpteenth time, pessimistically optimistic.  Going to get new sensor now.

[tzei]

Oh and btw how is your TPS doing? New, old, wires ok? Dunno how your Hatech is set up in that regard but worth an examination?

[irdave]

24 minutes ago, tzei said:

Oh and btw how is your TPS doing? New, old, wires ok? Dunno how your Hatech is set up in that regard but worth an examination?

 

Original TPS I had was a little rough.  This one is a used s54 / e46m3 TPS- not to say it's any better, but it sure seems to be, both by the pedal and by the trace on the data.  I'll be honest, I've fixed so many other little things, which I've been led to by this problem, that it's great, but...

 

As it's not the COP issue, I'm going to uninstall the dizzy and put the COP set up back on- it just takes up so much less space, and I don't have to fight for space between the dizzy and the upper rad hose.

 

I tried the function on the haltech to add some pulsewidth- didn't resolve anything.  Tomorrow will try pulling the new injectors and installing the old.  Currently tracking down some random wires from when the swap was originally done, stuff that was just cut and left.  And I did add a diode to the extra / new blue wire added from the alternator to the dash, so now the car turns off with the key- no more having to turn the heater blower motor on to turn the car off.    Did not resolve the stumbling issue.

 

I also moved the crank and cam position sensor wire bundles from under the intake, by the alternator, to over the intake, just sitting there, but about split between the dizzy and the alternator; no change.  I mean, they are shielded wires- but I figured it was an easy test.

 

Was pulling some data; found that my IAT was varying with rpm.  Yep.  And I often get data drop outs- and it's across all recorded channels.  Could just be random while recording with the ECU.  I need to check with data recorded on the laptop vs on the ECU.  Maybe vary sampling rate, too.  Oh, and went into town to get new temp sensor and groceries- left my wallet at home.  No new sensor, yet.  But why would it increase with rpm?  Something not having a good reference sounds like.  Is it possible to wire the IAT backwards?  It's only 2 wires; signal and ground...

[irdave]

More looking at the data; looks like IAT vary with long idle times, which makes sense.  Gaps in IAT readings are maybe due to sampling rate.

[tzei]

Prolly have checked ECU ground already? Have you logged voltage? Make sample rate as high as possible to see does data have spikes that lower Hz masks away - is TPS still okay?. Can you try different ECU unit?

 

If IAT shows potatoes does ECU have some failsafe number for that situation? Unplug it or set compensations to zero for a test?

[irdave]

1 hour ago, tzei said:

Prolly have checked ECU ground already? Have you logged voltage? Make sample rate as high as possible to see does data have spikes that lower Hz masks away - is TPS still okay?. Can you try different ECU unit?

 

If IAT shows potatoes does ECU have some failsafe number for that situation? Unplug it or set compensations to zero for a test?

 

Looks like I can log 5v sensor voltage.  I'll do that tomorrow.  Thanks.

[AustrianVespaGuy]

On 10/4/2021 at 6:47 PM, irdave said:

Looks like I can log 5v sensor voltage.  I'll do that tomorrow.  Thanks.

Can you post/share any of the logs? Would love to help more but really need to see the data in order to try to find something useful!

[irdave]

3 hours ago, AustrianVespaGuy said:

Can you post/share any of the logs? Would love to help more but really need to see the data in order to try to find something useful!

 

Ok, CSV file for sure.  I'll try to get a link to the haltech data viewer software as well, should be able to get a link to the tuning software, too, so you can look at the tune.  If not, I'll see if I can zip it and send it to you.

 

Spoke today with another guy at haltech.  He suggested changing my injector firing timing BTDC- he suggested simplifying the table (rpm only) and making the timing advance as the revs went up- made sense to me.  The table I had had more complexity and would retard the timing at about 4k- my guess at about torque peak.  It was also most advanced at lower rpms, then would retard a little as it went up.  So now I start at 375 at 0 rpm, 450 at 9k (whatever), with linear regression across.  This seems a little better.

 

Also, turned the idle screw out a bit to bump the idle to about 1100.  This helps hide the problem.  Had Jim's carb tune set up (vacuum ish tubes) a couple weeks ago.  Went through the procedure to adjust and balance the throttles.  Onboard MAP sensor plumbed directly to where the fuel pressure regulator is.

 

Also, guy suggested turning off all the tuning / self learning for now, as well as most / all of the corrections- just the fuel table and tune it by hand.  Which I did for a bit today driving around in the sticks.  I have to say, while the stumble is still there, it feels better not correcting everything itself all the time.  Throttle response feels better, too, even without any transient functions being active.

 

Also, I've completely removed the ICV and blocked off the openings- it just wasn't being consistent enough to solve problems.  Still left the inlet with a little filter so it can pass air to the air bleed screw so I can adjust the idle- just no ICV.

 

The dizzy didn't make it any better, so I reinstalled the COP set up.

 

The stumble seems to not be so bad the first couple of minutes of driving, then gets worse when it gets warm.  The stumble is available at idle, just sitting there, as well as it noticeable  at most any constant throttle setting.  Transient response, even with that function turned off, is not bad at all- this is not a transient result- at least I can't see how it could be.

 

Attached log is constant throttle for some seconds, then a bump in revs, then more constant throttle.  Stumble was noticeable at both constant throttle positions.

5oct21-stumble.csv