[*sean*]

ive yet to see ckp show signs of a 'missing tooth' style of reluctor wheel.. it could be counting teeth for RPM data and NEP could be phasing the ignition timing as the trigger??
really wish i was more intimate with my g1 sensors right now. i'll be staring at my motor until my feet get cold out there.
more data = more better

edit: crank sensor is simple 2 wire sensor counting teeth..
and looking at graphs provided so far, the nep is feeding the ignition coil, i would consider it to be ignition advance at this point in time, this is where a comparison to obd scanner could quickly verify.. hmm

I went for a drive and collected data with both the Saleae and my OBD2 reader, mostly the same data. The OBD2 stream (which I have not extracted yet from the reader) has a sample rate of about once every 1-3 seconds per data element for this much data. The Saleae, on the other hand, which I slowed down to 31K SPS (samples per second) is just a tad faster So I have the scanner data, just have not compared the two yet.

By the way, it would have been far easier to find or fabricate pins to wedge into the backshells of the ECM connectors and connect the Saleae directly to those, and with some tape to secure it, they might not fall out. This would have avoided days of work to build this contraption if one just needs to confirm a suspicion or learn more before going deeper.

At any rate, the EGR position data is showing classic signs of an intermittent EGR valve wiper. I built a tool to test this, but it's in pieces at the moment.

This screenshot captured conditions when a slight jerkiness was felt during cruise (not in lean burn). Notable is that the EGR data is a mess. Based on my earlier testing (see image in post linked to above) the valve may actually be a lot more open or closed than the values suggest. Also, looking at changes to the "EGR drive" PWM signal, you can see the ECM appears to be chasing the valve position and it is underdamped. Finally, looking at fuel injector pulse width I can see instances where an 8.4 ms pulse appears between a 9.8 ms pulse and 10.4 ms pulse. That's quite a big difference in fueling! So it's time to look more closely at this EGR valve.

I wonder if the EGR valve position is used by the ECM to determine injector pulse width.

I am curious to look at AFR. I am wondering if its role is for fuel trimming, to maintain a target AFR on average, with things like manifold air pressure and maybe EGR valve position having an immediate effect on injector pulse width?

 

[Gasoline Fumes]

ive yet to see ckp show signs of a 'missing tooth' style of reluctor wheel..

12 teeth according to a drawing in the FSM. No missing tooth.

 

[trickle]

im familiar with your past tests with egr. still wrapping my brain around the 3-cylinder cycle.. lol
my thoughts on AFR: at any given rate its really not as simple and/or fast acting as one would think, port injection is a dynamic mess, some fuel hits the port walls and puddles up until the valve opens again, some vapes off the valve, some shoots right in the hole. this all happens at different pumping air velocities because the intake manifold runners are shared, there is some reversion pumping to boot. This car uses VE tables, because it has no mass airflow sensor. Basically, there is different look up tables for the various sensors' vs load

At any given rate.. the egr job is to add inert volumes of gas into the incoming charge, theory behind it is fill the void for a leaner mixture while reducing Nox
the spark can only burn oxygen and gas, the egr reduces oxygen (by adding Inert gases) therefore the ecu can add less fuel for given qualifiers.
If egr is fluttering around all willy nilly like that, it's commanding erroneous injection pulse widths, allowing for erroneous 'actual mix ratios' to enter the cylinder, by the time all that info gets processed by the Laf it's too late.. the bad combustion event happened and continues to happen in a way that the ecu can't correct.
on a well-balanced four cylinder or a v8 it's probably much less pronounced than it is in the G1

The inherent out of balance three-cylinder four stroke cycle 'magnifies' the jerking motioin everyone speaks about

i hope this makes sense and helps paint a picture

 

[trickle]

I don't know what your next move is sean but i would gladly share the steps i would take after sorting out that epileptic egr valve. there could be a multitude of other hinderances effecting one another. I'm sorry for my 'loose' explanation above, I'm capable of being more thorough, but i very much enjoy learning how other people 'fix stuff'

 

[trickle]

12 teeth according to a drawing in the FSM. No missing tooth.

accurate reference Fumes.. 1 tooth = 30* ... 12teeth = 360* of crankshaft, compare that to seans 'mad scope shot above'

finally got my brain thinking 3 cylinder.. took a minute
i counted 24 ckp signals between tdc events, 720* of crank = 360* of cam = 1 four-stroke cycle

from there the ecu needs to keep track of 3 cylinders somehow, reference tdc1 and tdc2, the two cam sensors

firing order is 1,3,2 where;
cyl 1 = tdc1 and tdc 2 signals simultaneously
cyl 3 = tdc1 signal only
cyl 2 = tdc2 signal only
..than repeats

imagine this;
cyl1 fires at 0* ... 720* ...
cyl 3 fires at 240* ... 960* ...
cyl 2 fires at 480* ... 1200* ...
...240* between firing events.. 60* of which has no power stroke applied.. its why the car is gutless at low rpm without ima and contributes to the rocking motion that retires rotten rear motor mounts'

NEP signal switches high/low every 120*
360* / 120* = 3 cylinders
my intuition says input reference to ecu (like a tach signal)
ecu could use this as baseline for top dead center and convert it to different outputs like injection timing, spark advance, tachometer

 

[Honda hybrid442]

Just buy a used egr swap it out. Come up the garage and ill throw one in for free .Merry Christmas Sean.

 

[Gasoline Fumes]

NEP is an ECM tach signal output. I don't know what the letters NEP stand for, but the FSM says it's "engine speed pulse." The signal (A19 on the ECM) is sent to the MCM (C15), gauge cluster (A21), EPS (B14), and test tachometer connector (near brake master cylinder).

 

[*sean*]

[the] ecu could use this [NEP] as baseline for top dead center and convert it to different outputs like injection timing, spark advance, tachometer

I agree! That might be the reason for its existence internally to the ECM! A quick check of my SM pdf (I did buy a copy of the Bishko-sourced PDF) found that it is consumed by the MCM, TCM (CVT), and EPS computers.

Just buy a used egr swap it out. Come up the garage and ill throw one in for free .Merry Christmas Sean.

I definitely owe you a visit. I might take you up on the used EGR valve once I finish V2 of my EGR valve tester as test sample. The used EGR valve I am currently using seemed to work great in my 2004 but <edit: I think it was the one that was in the car when I got it> is definitely showing that it's got an intermittent wiper based on the capture above.

So, it's time to build a new EGR valve tester, since I broke my last one by stepping on it.

 

[trickle]

NEP is an ECM tach signal output. I don't know what the letters NEP stand for, but the FSM says it's "engine speed pulse." The signal (A19 on the ECM) is sent to the MCM (C15), gauge cluster (A21), EPS (B14), and test tachometer connector (near brake master cylinder).

back to back excellent references Fumes.. were you a librarian in a former life?
Must be nice to have a FSM for this car beings its such an odd/pioneering car from honda. The EPS must need an on/off switch for autostop conditions?..

 

[Gasoline Fumes]

TCM

Yeah, that too! I have the 2000 ETM, so no CVT info.

 

[Gasoline Fumes]

back to back excellent references Fumes.. were you a librarian in a former life?
Must be nice to have a FSM for this car beings its such an odd/pioneering car from honda. The EPS must need an on/off switch for autostop conditions?..

Never been a librarian as far as I know, but I keep my FSM & ETM handy at all times! They can be found on eBay, if you want a paper copy.

EPS is on and active whenever the ignition switch is on. Except if you start rolling with the engine off, it won't activate until a few MPH. Annoying if you want to coast out of a parking spot!

 

[*sean*]

NEP is an ECM tach signal output. I don't know what the letters NEP stand for, but the FSM says it's "engine speed pulse." The signal (A19 on the ECM) is sent to the MCM (C15), gauge cluster (A21), EPS (B14), and test tachometer connector (near brake master cylinder).

I was composing my post when you sent yours! I didn't find the connection to the instrument cluster or test connector which I'd forgotten about! Thanks!

I keep my FSM & ETM handy at all times! They can be found on eBay, if you want a paper copy.

Yep, got mine from there too. The paper ETM is currently buried under some project... so much easier than pulling up the PDFs. But "pdfgrep" - Linux app, available via Homebrew on Mac and probably under WSL on Windows is a great way to search all the PDFs for keywords. "pdfgrep -Ri EPS" for example will look in all PDF files in the current directory and below for the word "EPS" or "eps". -R for recursive, -i for case insensitive.

The other document which I find incredibly useful and essential is Service Bulletin SB 03-020 which has technical details about what exactly triggers many DTCs. Really helpful for understanding how things work. @mudder has a copy somewhere at doppelhub - Overview.

 

[*sean*]

So I'm going to try to build the new version of the EGR tester from a Seeed xiao esp32c3, which I happen to have in my junque box: Getting Started with Seeed Studio XIAO ESP32C3 - Seeed Wiki

Tiny cheap BLE + SAMD21 microcontroller available for <$6 from Digikey or Mouser or $25 for 3, same day, Amazon.

I am going to try to do a schematic using Fritzing. I was going to build it from Github source but took a look at the dependencies and decided to pay the devs $9 for a pre-built version. One minute later, installed and running. Done. Download Fritzing

Found the Seeed part for Fritzing at Seeeduino Xiao - Anyone got the part?.

 

[*sean*]

Looking at leveraging phyphox.org to use the phone to provide the display for the Arduino. This is promising. The Arduino speaks!

 

[*sean*]

Spent two hours trying to figure out why analogRead was not working on A0. Changed to A4 but it was coming in on A2. Only some digital outputs working. ?????

Spoiler

Regarding the mapping of the analog inputs, it appeared to be due to problems with the definition file for the ESP32C3! It finally appeared in Arduino 2.0.3 which released recently. See
ESP32-C3 cant read all analog values with analogRead() · Issue #5502 · espressif/arduino-esp32. Additionally, I was using a pinMode(pin, INPUT) on the analog pin and that seemed to cause isses with the digital pins, perhaps on the same port. Removing that brought the digital pins back.

 

[*sean*]

PWM output working. It was really easy to change PWM resolution to 14 bits and still reach over 2000 Hz PWM frequency with the ESP32 "LED Control" library.

 

[*sean*]

Wow - just found this article about detecting cylinder imbalance using the A/F ratio sensor. I got too excited to do more than skim it because it suggests that one might be able to detect a cylinder with a lower output due to fueling, ignition, or other issues by looking at the LAF sensor output compared to the others.

One has to wonder if this could also be used to test a LAF sensor's performance against a known good sensor or even just standalone.

I need to finish my EGR tester and get that sorted out. Then looking into a tool that can help diagnose cylinder imbalance is high on the list. If it works, it could be an essential tune-up and diagnosis tool. It is apparently a feature of advanced diagnostic code readers, and a monitoring requirement for newer cars, according to the article.

https://www.autonerdz.com/yabbfiles/Attachments/Individual_Fuel_Control.pdf

 

[*sean*]

Happy New Year!!

The EGR valve driver in the test circuit works. Using a 1N5822 for a flyback diode, but the current draw is well under an amp so I may switch to a 1N5817 (smaller; leads fit into breadboard).

 

[*sean*]

Unlike my last tester, which used the default PWM frequency of the Arduino UNO, I can easily set the PWM frequency. I found, from my last capture, that the car uses a PWM frequency of 108.5 Hz. That's pretty low.

I grabbed my box of old EGR valves and discovered that most of the old Insight valves were fairly quiet while others (particularly the non-Insight valves) were quite noisy, to the point that I could feel them vibrate.

Dug out the scope, hooked up one of the louder ones, set it to be open about halfway, and voila, there it is, a sine wave in phase with the PWM signal. To confirm, I pressed the valve shut with my finger. As it was closing and the force against my finger grew, the amplitude of the sine wave increased, until I held it shut and immobile, at which point it became a flat line.

This is interesting:

(1) These vibrations may be accelerating the wear of the valves.
(2) Perhaps the original valves were tuned to this frequency to minimize shaking???? this may be a stretch.
(3) Or, perhaps the PWM frequency is by design. The low frequency may actually help position the valve as it might help break static friction that could cause positioning error as the bearing friction increases with age and contamination.
(4) You can make a (crappy) speaker from an EGR valve.

There is a healthy amount of noise. I'm not sure of it's source, but I might need to find and fix it. The waveform changes a little if you turn the valve sideways which increases the bearing friction.

 

[*sean*]

It occurred to me to apply a 30 Hz signal to the valve. This shakes it plenty, to the point where I'm now wondering how I could modify one and attach it to the bottom of the seat as a butt-woofer!

But seriously, a low frequency signal could help quickly diagnose a stuck valve. If it doesn't make a racket, it is stuck. I have one that was throwing a code in my HCH which is now one of the test samples.