Dwell Time would be longer than the table

[Mach]

Dwell Time would be longer than the table, as if it were referring to a different table.
At low rpm (1500 rpm), there seems to be fine.
There are no triggering errors during the engine running.

Attaching the base map and an ECU log (battery voltage and dwell time at 500 Hz).
The point at which the voltage temporarily dropped was when the headlights were turned on.

https://drive.google.com/drive/folders/1toBwNBD62UktU0rE0SGxvHmGBRLTxYdm?usp=sharing

And I've done
- Replaced battery and ignition coils ; nothing change.
- Checked ignition timing with timing light ; no probrem
This was not in the previous log. What has changed since then is that the engine was replaced.

I've confirmed this in both Ver 6.21 and 6.22.
 

[Adamw]

Dwell scatter typically means there is a problem with the trigger.  If the ecu is not getting regularly spaced teeth events then it throws off the crank angle prediction so the ecu has to stretch or shrink the dwell aim to make the spark occur in the correct place.  Generally about +/-10% is ok, beyond that usually suggests there is a problem.  

What engine and trigger mode?  It looks like yours drifts most when the engine is slowing down.  Can you give us a couple of trigger scope captures when operating under whatever conditions gives the most error.   

[Vaughan]

Look into cam belt tension and inspect the trigger teeth on the cam gear for damaged edges

[Mach]

E07Z trigger mode.(https://forums.linkecu.com/topic/13923-trigger-setup-honda-3-cylinder-engine/)

Especially, it has most error (over 30%) between 1800~2500 rpm, and +10~20% dwell time at higher rpm.
At idling there is little error.

Attach trigger scope.
                                (Dwell time / Table)
- 2300rpm ; +30% (4.3ms / 3.3ms)Trigger_2300rpm_4.3s-3.3s_TPS6.llgx
- 4000rpm ; +13% (3.5ms / 3.1ms)Trigger_4000rpm_3.5s-3.1s.llgx
- 1100rpm(idle) ; little(3.4ms / 3.4ms)Trigger_idle_3.4s-3.4s.llgx

18 hours ago, Vaughan said:

Look into cam belt tension and inspect the trigger teeth on the cam gear for damaged edges

At least they are fine when I check 2weeks ago, but I'll check them again.

[Adamw]

I passed your files on to the firmware team today for advice.  We cant reproduce the problem when testing with a simulated cam signal on the test bench.  But one of the engineers noticed in your log right around 1800RPM there is a large step in dwell.  The ignition advance is around 16deg both times when this happens .

He asked if you can try holding the engine around 1800RPM, and move ignition timing up and down - say vary between 10 & 30deg to see if you can reproduce this big step change in dwell (Select all cells in ign table around that area and Shift+Pg Up/Pg down for large increment.

If you can make the dwell jump by doing this test, can you then also try the same advance/retard test at a different RPM - say idle speed to see if it related to advance or RPM.  

[Mach]

I followed your advice and did the test.
The dwell table was set to 2.5 ms for the entire engine speed.
As a result, a regularity appeares.
Change the ignition timing, the dwell jump RPM also moves.

PC Datalog - 2022-05-16 8;32;53 pm_dwelltest.llgx

(ignition angle / dwell jump point)
10deg : 2750rpm
15deg : 2400rpm
20deg : 2150rpm
25deg : 1850rpm
30deg : 1650rpm

[Vaughan]

On 5/13/2022 at 1:41 AM, Mach said:

This was not in the previous log. What has changed since then is that the engine was replaced.

 

Is the new engine completely factory or does it have changes like a different cam or a lightweight flywheel?

[Mach]

11 hours ago, Vaughan said:

Is the new engine completely factory or does it have changes like a different cam or a lightweight flywheel?

One of a change is cylinder head skimming (0.2mm), so the length of timing belt (from crank to cam) may be a little shorten.
But the amount of change is very small, and trigger sensor is only at the cam, so I think it less effect to trigger.
I've checked cam gear and sensor position.

Other changes are cam, valve spring, low-comp piston...
Flywheel is same as previous one.

[Adamw]

I discussed today with the firmware engineers.  It seems there is some variation in the timing of the edges received on the cam trigger.  We are not sure if this is a sensor/electrical issue or camshaft acceleration from loose belt or valve springs etc.  Of course 3Cyl with only 3 teeth on the cam is the worst case as the ecu is only getting a position update every 240degs of crankshaft rotation.  When the ecu receives an edge earlier or later than was predicted based on the time between the last 2 edges then it has to stretch or shrink dwell to ensure the spark occurs at the correct angle.

A couple of examples below.  Edges should be 0/240/480/600/720.  You can see at idle it is pretty close.  But at 2300RPM there is 5deg error on one edge.

Idle:

 

 

2300RPM:

 

Vaughan also noticed your older trigger scopes have a higher voltage on trigger 1 and a much sharper waveform.  Example below - old engine on left. 

From memory you may have had the OEM ecu piggy back?  Can you try adding a 1kohm pull-up resistor between 5V and trigger 1 to see if that makes the waveform more accurate.

[Mach]

The older log are with piggy back, but now it has been changed to standalone and use internal pull-up.

I understood the cause of dwell jump and that some of a hardware is causing the trigger error.
The belt tension is as per the manual, but I will try more tension or other solutions.
Ignition timing near the jump point is also almost correct, so for the time being, I'll deal with this by setting a smaller dwell time.

@Adamw @Vaughan
I appreciate your support.

[Adamw]

Try temporarily wiring in a 1K 5V pull-up to trigger 1, some hall sensors get slow when there is not enough current passing through them.  Im assuming it is the same cam sensor used on the old and new engine?