[rusEfi] hello all - Mitsubishi #31

[stefanst]

My first suspicion would be the short dwell time. Just up it to 4ms or so and see if the issue goes away.
All "standard" Megasquirt are using DB-style connectors, lots of them with VR sensors. I don't remember any mention of that causing issues. So while it's definitely a possibility, it doesn't seem very likely to be the issue.

[matt]

i really forgotten about TS logging. since i booked dyno machine for 1 hour, me and dyno operator quite busy solve this issue. we have increase dwell, ignitions degree and change spark plug . one things in my mind maybe my mitsubishi cas is not working properly. when testing ecu on the table, i use another cas (4g93), and rpm can sync even i turn it by finger. this test is using Mitsu trigger in TS. at car i use another cas (4g63) it can't sync with mitsu triggger in TS, and working with miata na trigger.

[AndreyB]

Jared is right we need more data. Some logs?

Can you reproduce the issue on your driveway?

[matt]

ok, i will log it tomorrow. just now it can't start, fuel pump issue i guest.

[matt]

My first suspicion would be the short dwell time. Just up it to 4ms or so and see if the issue goes away.
All "standard" Megasquirt are using DB-style connectors, lots of them with VR sensors. I don't remember any mention of that causing issues. So while it's definitely a possibility, it doesn't seem very likely to be the issue.

i don't brave enough to up dwell to 4ms. me and dyno guy up to 2.5ms and issue still there.

he also got a burn coil (toyota vios/yaris) if dwell is set above 2.5ms.

[AndreyB]

We need data

Please tell me there is a video clip somewhere?

[matt]

yesterday i notice fuel pump not working. so today morning, i look at stm board and note the blue led is not blink as usual. i erase stm and load new firmware and it is working as usual.

what's cause this issue?

[AndreyB]

what's cause this issue?

Does not ring a bell, not sure. Next time please "read chip" and save to file for me to try reproducing this locally.

[matt]

quite weird.. i can pull my car up to 7000 rpm. i tried with 1st and 2nd gear, not enough brave to shift 3rd gear because of rainy day. anyway here is my log.

[AndreyB]

In your logs you have a lot of
/**
* This error happens if some configuration changes were applied but ECU was not reset afterwards.
*/
CUSTOM_OBD_ANALOG_INPUT_ERROR = 6038,


and a few

/**
* Commanded fuel exceeds your fuel injector flow
*/
CUSTOM_TOO_LONG_FUEL_INJECTION = 6050,

[matt]

yes, i do some tune while logging.
do you need a new log?

[AndreyB]

Let me clarify the error message
/**
* This error happens if some pinout configuration changes were applied but ECU was not reset afterwards.
*/

[matt]

i tune via micro usb on stm32disc. can it cause this issue?

got few other questions:

how do i wire input switch for launch control? it is just a ground switching to op-amp or going to do something like this:-

can i put capasitor at cam/crank signal to filter some noise? it seems my starter motor create some noise and interrupt rpm signal while cranking.

[kb1gtt]

On default Frankenso, input 8 and input 6 already have pull up resistors. On page 3 of the schematic, see R280 in zone B4, or R260 in zone D4. You can use the diodes if you prefer, but I expect they are not required.

Keep in mind the VR voltage at high RPM, a cap with low voltage might not survive a large voltages generated at higher RPM. Also keep in mind that a cap will become low ohms at higher frequencies. This can cause your cap to overheat. Also also keep in mind that caps add signal delays, so your detected angle is not good, as the signal is delayed form the real signal. Basically I suggest you avoid using caps on the VR to remove noise. Did you install a resistor across the leads at the harness connector? If you still have noise, perhaps you an lower that ohms.

Also take a close look at the 5V supply voltage. If your battery voltage drops to low during cranking, your 5V could suffer. I would verify that your 5V is good and solid.

Perhaps check your starter / engine block grounds. Perhaps even add a GND to your starter mounting bolt, such that starter currents do not run through the engine block. You can minimize radiated noise if your start GND wire is put directly next to your positive lead.

[matt]

On default Frankenso, input 8 and input 6 already have pull up resistors.

did you mean my crank/cam signal can wired directly to op-amp, no need to pull up? Btw I am using standart mitsubishi cas with hall effect.

[kb1gtt]

For the launch switch, you can use the input 6 or 8. There is a place on the PCB for the pull up resistor already. If it is not installed, you can install one there instead of proto area, or harness.

You can also use hall direct to input 9 or input 11. However it requires a small jumper wire to be installed. These have accommodates for pull up or pull down resistors as well. Typically people need a pull up resistor, but some need a pull down.

I thought you were using VR. If you have noise while starting and you have hall, then your problem is very likely a weak GND or a weak battery which is dropping the 12V to low and causing the 5V to go to low. Basically your GND voltage is bumping up and cause false trips of the input signal.

[matt]

currently my crank/cam are wired like this:-

campull.png (34.57 KiB) Viewed 29638 times

[kb1gtt]

That should work, but you don't need the external R970. The 1k internal to the Frankenso is sufficient.

[matt]

That should work, but you don't need the external R970. The 1k internal to the Frankenso is sufficient.

ahh, that is R470, sorry for my bad drawing. can this setup cause a trigger issue?
which should i remove? R1k or R470?

[kb1gtt]

I would say try removing the external 470 ohm and see what happens. Most hall's can handle 1ma to 5mA, Many can handle up to 20mA. In your case with the external resistor you drawing around 11mA, which might be a bit high, or it might not be high. When you start to saturate the hall's output yes funky things can happen. So I would say removing the external resistor puts you into a more safe range. However it is also likely that your not going to see a change.

Do you know your hall sensor? Does it have a datasheet? Many hall sensors have internal circuits that hunt for the optimal magnetic flux sensing. Your noise might be similar to the MAX chip with it's auto tuning and hunting. It's possible it's cranking up the internal gain and then triggering on small noise. A datasheet for your hall sensor would help show if that is a problem or not.

[matt]

done some testing today. i remove R470 and let 1K there, engine can"t sync and trigger error ligh up few times.
next test i will try replace 1K with R470.

[AndreyB]

How did you originally come up with R470 pull-up?

Assuming you had both would that be same as having a R319.7278 pull-up since http://www.sengpielaudio.com/calculator-paralresist.htm ?

[matt]

hmm.. maybe i read it here about R470.

[matt]

ok, found where the mistake is.
i populate ap amp channel base on r0.2 schematic. it seems pull up with R2700.
so i change R2700 to R1k and it can sync properly.

rusefi.png (162.94 KiB) Viewed 29624 times

[AndreyB]

Thank you for the update! Great news.

[matt]

one more thing. my engine still can"t start with mitsu decoders. it can start with miata na and trigger offset must change to 360.
so, what is different with miata na and mitsu trigger? it is use a same trigger calculation? i know i have aks it once, just need a comfirmation.

[AndreyB]

http://rusefi.com/wiki/index.php?title=Manual:Software:Trigger#Mitsubishi and http://rusefi.com/wiki/index.php?title=Manual:Software:Trigger#Mazda_Miata_NA shapes are pretty close

Trigger#Mitsubishi has no build-in offset, Trigger#Mazda_Miata_NA has build-in trigger offset of 294

So if you are using 360 on top of 294, try 654 for Mitsubishi ?

Have you confirmed TDC location with a timing gun while running fixed timing?

[matt]

Have you confirmed TDC location with a timing gun while running fixed timing?

Not yet, i dont have a timing light. I will test with Mitsubishi with 654 offset once I reach at home

[AndreyB]

Just created https://github.com/rusefi/rusefi/issues/3 to make documentation more user-friendly

[matt]

i found something. maybe a bit different between those trigger.
mitsu cam trigger slot looks bigger then miata n/a.