Adamw

Post some actual info if you want help, my crystal ball isnt coming up with anything. Log, map, triggerscope, description of the problem would be a good start.

I dont know specifically for this model, but most of the Honda alternators will work ok without ECU control. Normally the ECU control is only used to switch it into low voltage mode to save fuel/emmisions when cruising at light load. Picture below taken from this document: https://www.giac.org/paper/gcia/9927/hacking-bus-basic-manipulation-modern-automobile-bus-reverse-engineering/133228 Shows some data from a 2011 civic. I have no idea if it matches your car. But if so then ID164, Byte 3 looks interesting (byte 2 in Link since byte numbering starts from 0).

The speed sensor shouldnt be outputing a frequency when it is not moving. You might have to look closer at the ground etc. The DI's on the G4X can all take at least 10KHz.

Use the AAC plug for the new idle valve. You cant use a MSD6A with coils that have built in ignitors. The MSD is only good for a single coil/distributor. The Expansion loom gives you the 6 coil triggers, the 12V will have to come from an ignition switched relay and the grounds should all go to a single point on the cylinder head. Yes trig 1 is crank and trig 2 is cam. You really need to look at a data sheet for the new sensors to determine what the colours mean, but typically for a hall senor with those colours I would expect Tan would be +12V, Black is ground and blue is the signal. I have attached a pic of the CAS pinout below. So both tan wires would go to the "CAS supply". Both black wires would go to the "Sensor ground". Crank blue goes to Trig 1 and Cam blue goes to trig 2.

"Knk Level Cyl*" is just the level of noise coming from that cylinder, it will increase with RPM, boost, ign timing and many other variables that effect the mechanical noise inside the engine. The level of noise above which you consider it to be knock is determined by the knock threshold table. The knock threshold would generally be determined by the tuner by ploting the typical noise levels when the engine isnt knocking, then purposely making it knock and ensure the knock level exceeds the threshold reliably when it knocks. In my opinion your threshold doesnt look well set especially around 4-5000RPM where knock would most commonly occur - you have a knock level on the noisiest cylinders only around 160 but the threshold is set at 615. A threshold around 300 would have a better chance of catching knock if the typical level is only 160. You should generally have knock threshold ploted on the same graph as your "Knk Level Cyl*" channels so you can see if any are getting close to or exceeding the threshold. Example below. Note the newer layouts supplied in PC Link - either 1920 x 1080.llf or Default_1366_x_768.llf have a knock page set up.

@Mach I will send you a beta version of PC Link to try which supports the E07Z trigger. Install this, then upgrade the ecu to the firmware included in this version. This has firmware has a new trigger mode in the list "Honda E07Z". Only the cam sensor needs to be connected to trig1, nothing connected to trigger 2. Please try running the engine, verify RPM is stable and realistic. Trigger offset will need to be checked/set with a timing light. If it all works ok, please report back what final trigger offset was required for accurate timing and we will add that into the trigger code so that the final publicly available version will have near 0 offset.

In thoery it would maybe be possible to use the two stock flywheel sensors for crank pos and a 3rd sensor on the cam wired to a DI for phase detection but it would need some fairly major trigger coding to make it work. I would also be a little concerned that the fixed arming threshold on the DI inputs would be too high for reliable sync at cranking speeds if it were a VR sensor.

Yeah Im gonna say probably a lack of fuel too. The fuel table shape looks a bit flat for a TP based map, but just bumping up the fuel master as Vaughan suggests will probably allow it to run. I would also turn off the 4D fuel table that adds fuel based on battery voltage, not sure what that was for but it will possibly make tuning difficult.

I answered this earlier on your facebook post so I will past the same answer here: You will need to add ignitors into the wiring or change the coils to something with built in ignitors. The Ford ecu had ignitors inside the case, the Link ecu doesn’t.

Basic set up from the help file below. From memory they have a cam sensor on each side of the engine and the trigger offset is 360deg different depending which side you use. So trigger offset may be around -4 as below or 356 for the other cam (I dont know which cam our data was based on). There is a 1UZ Atom base map in G4X PC Link. It is only a basic start up map and not from a tuned engine. If using traditional fuel equation you dont really need to know injector size, just adjust the master fuel value until it starts/runs. I would expect if using the stock 251cc injectors the master fuel setting will need to be up around 20-30ms.

The 5V supply is rated for 400mA at its specified accuracy. So typically enough for 30-50 sensors.

Is it a G4 or G4+? Black or blue case?

The flywheel turns twice for 1 revolution of the cam. So when the flywheel "TDC" tooth passes the sensor there is no way for the ECU to know whether it is TDC on the compression stroke or TDC on the exhaust stroke. For direct spark and/or sequential injection you need to know which stroke your are on which typically means a cam sensor. The other option would be to change to a more conventional missing tooth crank wheel and use "MAP sync" instead of having to add a cam sensor as well (G4X can use a MAP sensor connected to 1 intake runner to detect the intake stroke).

@Mach The original osciloscope of the trigger you posted above, can you confirm that this engine only has just the one sensor on the cam? Am I right the signal you show on trigger 2 is being generated by the factory ecu and would not be present if the factory ecu wasnt in place? Also, can you confirm there is no VVT on this engine?

Then I cant think of any reason why it wouldnt work then. The Aux output should produce something very similar to a hall effect sensor. If you connect the speed sensor to DI1-4 you could do a trigger scope to capture the duty cycle of the original signal if you wanted to duplicate it as close as poss, but usually speedo's are just looking for edges and the DC doesnt really matter. Or you could just try say 30%, 50% & 70% to cover the bases.

Yes, it works fine, but is only capable of wasted spark or distributor ignition. If you want sequential and direct spark you will have to fit aftermarket triggers as Sterling suggested above.

Swap the dip switches and check you have realistic RPM displayed when cranking. If you do then move on to confirming timing. If you dont then they were probably correct the first time.

Is his V3.1 hardware? You can see in the ECU information screen.

I suspect you have the trigger dip switches set wrong. Did you try them in the opposite configuration as per the manual?

No, case does not need to be grounded. The main ECU power grounds are B4, B69,B79 & B80. Not all of these may be connected in all cars, But at least a couple of them should be and they should have very low resitance to the engine block.

I've got the pinout below for the cam sensor, you will probably find the crank is the same. "Out" connects to the ecu trigger 1/2 pin. "Gnd" connects to the ground wire. "S" connects to the shield.

If you cant find information from some other engineer that has done it himself and publicly shared the information, then the only way is to connect a CAN sniffer to a car with the stock ecu and reverse engineer the CAN messages.

Did the speedo work when connected directly to the sensor as per factory?

I have never tested a R35 coil, but many modern coils switch at "TTL" level. This means they only need to see a signal of about 2V on the trigger pin to start charging, and they will spark when that trigger pin voltage drops below 0.7V. So you only need a relativly minor ground problem before you can end up with a offset between the ground that the ECU uses and the ground that the coils use and run in to behaviour like this. I just meant wiring issues in general. The crank and cam sensor are very simple devices, only 2 wires each yet they were both wired wrong from the start. In contrast, coils are relatively complex high energy devices with 18 wires between them, so I think the chances of a wiring issue in this dept would be something I wouldnt rule out.

This ecu is the latest hardware revision and has the new ignition circuit which shouldnt be able to cause a spark at power up. Given your other wiring issues, I suspect you may have an odd ground offset or something between ecu and engine/chassis causing the coils to spark on power up. A quick fix might be to change the trigger wire on your coil relay so that it is switched on with the fuel pump.