[AngryScotsman]

Issue:
Engine will idle for 30+ minutes, no problem.
If I drive the car, I can get maybe 15-20mins of run time before the car begins to hesitate at part throttle. Will stall coming to a stop. Won't start back up for ~30 minutes. If I let it cool down completely, works like nothing happened for another 15-20mins.

Datalogging shows nothing that would point to any kind of issue.

What was checked:

• Fuel pump replaced
• verified getting correct fuel pressure (55psi)
• New coilpacks, verified spark when engine won't restart.
• Wideband replaced
• Cleaned injectors and verified injectors were working when engine won't start
• Swapped relays around to see if car would start, nothing.
• Checked Crank sensor for voltage and 5v, both pass. Replacing sensor soon.
• Swapped maps, nothing
• Disabled Closed Loop, running in Open Loop, helped for a little, then would still fail.

What else could it possibly be? Why would the engine not want to stay running after being driven, but work without issue idling?

 

[Brotatochip]

Coolant temperature reading properly?

 

[Lotus]

what does the wideband read in the logs? Sounds like it starts to run lean or excessively rich.
Would you be able to post a log here as a cloud link?

What happens when you try to restart?
if you get fuel and spark, it should fire.
If timing somehow slips or valves a sticking, the logs should show in the CAM readings.

What car? What engine? What ECU?

 

[LotusElise]

Datalogging shows nothing that would point to any kind of issue.

You might sharing it via cloud? I would like to check calibration and a log of idle in "colder" and "heated" conditions.

Sounds like a thermal expansion (crack) or deceleration cut off setup issue. But this is just throwing ideas over the fence.

 

[MoHo]

What else could it possibly be?

Your low on gas? I would go with fuel pump myself. Please keep us posted.

 

[AngryScotsman]

Tune and Datalog here:

Shared - Google Drive

drive.google.com

Gas was full.
Coolant temp is reading fine. It was cold, but this issue started in august of last year where it was fairly warm.

Kpro V4.

Scenario 1:
idle car for 30 mins, idles fine, no issues, go for a test drive.... hesitation and stall within a mile or two. Struggle to restart.

Scenario 2:
idle car for 10 sec, drive, can get 10-15 miles before hesitation and stall. Struggle to restart.

When this happened over the past weekend, I had a friend tow it back to my house, where it still wouldnt start.
Pulled all plugs and tested for spark. Saw spark.
Pulled all injectors, verified getting fuel. Got fuel everywhere.
Fuel pressure was showing 55psi, which is normal.

Someone on the Hondata forum mentioned to update my dead times on my injectors, which I will try, but not sure how the dead times will cause a sudden change in reliability for the engine. Figured if the dead times were always off, then the engine wouldn't go from letting me drive it like a normal car... to having a range of 10-15mi.

 

[MoHo]

Gas was full.

I wasn't trying to be a smartass. But it seems to be acting like low on gas, bad fuel pump kind of thing. It would seem the problem is while it's under load. Have you checked the duty cycle on the injectors? Surprised your not getting more help on this. Your looking at all the usual suspects. Let me ponder.

 

[LotusElise]

Tune and Datalog here:

Thanks for sharing it. Following I want to be noted after skimming it:

• basics
  • ECT is around 58 to 59 °C or 66 to 67 °C according AN1 input and showed during 4 minutes part load operation no increase like I would expect on a small-big-circuit-coolant-layout. Is there only a big circuit available or is the thermostat out of order?
  • fuel pressure from AN2 signal does correlate to MAP in a lower intensity, but the signal calibration does not fit the sensor reading, goes from 0 psi to 200 psi, while MAP is only changing from 12 to 66 kPa. The factor between the bandwidth of the fuel pressure sensor and the MAP has no quantitativ match, only qualitatively. Beside that the oscillation frequency and amplitude does not match good enough, there is a constant overlay oscillation on AN2 of around 1500 ms period. Where is the fuel pressure sensor connected to and does the your setup calibration match what it is specified (100 [email protected] V)?
  • at 02:41 to 02:44 there is a paradox ratio happening of boost pressure (AN5) vs MAP. While MAP is done at around 20 kPa, boost pressure is almost atmospheric. Anywhere else, except the it correlates with MAP.
• noticeable
  • MAP sensor signal has a too big to tune oscillation. That signal need to be calmed, either mechanically or by SW. As this is a Honda ECU and the later does not work you need either change the sensor for test purpose or calm the pressure connection to the plenum by some obstacles (e.g. by fixed metal wool)
  • Lambda tuning at part load has some potential to improvement, as up to 25 % deviation is happening, especially around idle. But carefully, this can a consequence of an faulty sensor input too. So first would be checking MAP.
  • IGT is quite advanced at low part load, effective IGT's of up to 54° BTDC. When I compare IGT, VTC at 3500 rpm/75 kPa I see 41.5° BTDC and 20° VTC. The later tells me scavenging efficiency is not performing very well, but the first tells me it must be perfect or you lowered CR by around 1 point. On a 10:1 CR I would expect around 30° VTC and 36° BTDC and with a turbo around 10-20° VTC and around 30-35° BTDC, just as a thumb order. Point here is, even if it is E85 the MBT gives one a rational limit to advance. If your pistons would have no squish area with the head, maybe it is possible. Is that tuned on the dyno or street? Did you find MBT or just advanced till knock and added some IGT distance to knock margin?

As an first assumption, I would call the MAP and the idle lambda the major suspicious to act on first (row of appearance).

 

[AngryScotsman]

The AN inputs are a bit... off. AN0 and AN1 should be good, but the others are a mixed bag as I've been using some of them to track fuel pressure, turbo line oil pressure, gearbox oil temp... but I didn't relabel them. The sensors also aren't calibrated for Kpro either, as I added them to use to push to my CD7L dash, which has a MUCH better conversion process than the Kpro.

MAP sensor, hadn't considered that. I could replace. the placement hasn't changed, but doesn't mean it can be placed in a better location. It's sitting in a throttle body adapter.

The tune was done on a dyno. Effective CR is around 10:1.

 

[LotusElise]

MAP sensor, hadn't considered that

Let me know if another works out smoother on the signal side. The jumping signal definitely causes a jumping fuel supply, which is in idle speed, which is a challenging balancing of fuel induced and friction torque, a disturbance which may cause the issues. When MAP is smoother, the fuel tune around idle should be revised. It should help to stabilize idle speed and prevent the engine from stalling. The deceleration cut off is the next point. The overrun cutoff table for low cam is set that there will be very very late a controlled run into idle mode, not to say there is just a wall. The overrun cut off at around idle speed need to give the idle controller more comfort, actually it show idle control just a finger and it is no thumb up finger.

The tune was done on a dyno. Effective CR is around 10:1.

So IGT was tuned according mean best torque?

 

[MoHo]

So IGT was tuned according mean best torque?

It seems like a heat soak type of thing to me? Closed loop, open loop? O2 sensor I just think it's something simple that's being over intelectionalized. I'd go back to basics but then again, I'm not a tuner and don't even have K-pro.

 

[Lotus]

“intelectionalized”
I would disagree with that. One needs to be aware of the failure mechanisms and what effects a failed or failing sensor has on the overall fueling in closed loop. That is if one takes care of their cars themselfs.

 

[MoHo]

if one takes care of their cars themselfs.

Nobody works on my car but me. I've owned it since 06'. An unfortuant hydro lock forced me to rebuild the engine. I will need K pro now? you must realize we don't all have degrees in quantom physics and what causes hydro-planning and how much fun it is at times? Your behond me. But I do try to follow along?

 

[Lotus]

@MoHo , may I suggest you open a separate thread for your topic?
That makes it easier to focus and distracts less from @AngryScotsman ‘s problem.

 

[MoHo]

@MoHo , may I suggest you open a separate thread for your topic?
That makes it easier to focus and distracts less from @AngryScotsman ‘s problem.

I'm just trying to understand? I hope he stay's in touch and lets us know what's what?

 

[Lotus]

@MoHo I was suggesting that as you asked: "I will need K pro now?"
Maybe I misunderstood you.

 

[Lotus]

I now also had a look at the log file.
you definitely have a MAP sensor or MAP sensor signal issue. The signal is way too noisy, reads way too low at times and at the end, when the engine dies shows only 170 mbar absolute pressure. This is lower than what you would read if you close the throttle at 8000 rpm.

The noise points towards a broken cable leading to intermittent contact. as the signal reads low, I would suspect it being the 5V MAP supply or the signal line itself.
Once the ECU reads MAP values that are too low, it will lean out the fuel accordingly and the engine will hesitate and then stop.
You can ignore the AFR and trim fluctuations as long as you don't have a smooth MAP signal.

It can happen that ECUs overheat and go haywire. But this is rather rare. Way more common for leads to brake near connectors from the engine vibrations or mishandling.
Where is your ECU mounted?

Is this car turbocharged or supercharged?
The map shows a 4 bar MAP sensor being installed. If you can limit boost to below 10 PSI, I'd go back to a stock sensor to confirm it is not the aftermarket MAP sensor failing.

 

[AngryScotsman]

Ordered a MAP sensor as well. Let's see what fixes it.

I'm almost certain it's a sensor of some sort... or a wire that is getting a bit loose -- was hoping the datalog would point to something more specific rather than me replacing all the things.

 

[Lotus]

map sensor is very specific

 

[MoHo]

map sensor is very specific

Now I'm starting to understand a little more? If it looks like a duck....It could be?