UAEFI vs 407disco (actual) startup times difference - SOLVED

[gurov]

kind of comparing apple to rotten potatoes here, but would be neat to have a reason for why delay after 12v is applied is different on UAEFI compared to basically same 407 disco setup:

UAEFI startup sequence, 4 super bright LEDS are stepper motor coils https://rusefi.com/online/view.php?msq=1623



407 disco actual based https://rusefi.com/online/view.php?msq=1624



the delay between switch on (+12v) and when the 4 leds start flashing (stepper motor moving home) is significantly longer on the 407 disco.

what could be the cause of that ? sdcard logging is turned off on 407 disco ecu. i tried turning spi busses off as well, and that made no difference.

[AndreyB]

Try running uaEFI firmware on f407disco and vice versa?

A ton of compile time options are different between these two firmware one would be around https://github.com/rusefi/rusefi/commit/e8b1fc302f71a3df2162235dba02f0e3f91bf01f

One day someone should add some timestamps into start-up console message but for now maybe try once this change builds https://github.com/rusefi/rusefi/commit/05d744b8f29549af6a11e24fe2ffa3ab343cc63d

[gurov]

hmm, perhaps once i get better gate drivers for the ignition setup on there that doesn't need to start inverted.

the pinout isn't quite the same between these two setups, so maybe disco firmware on uaefi could work if i set just a basic proof of life sign up.

[AndreyB]

hmm, perhaps once i get better gate drivers for the ignition setup on there that doesn't need to start inverted.

Ohh that's what this is all about? Ignition setup depending on start-up time seems like a signal of danger ahead.

[gurov]

i'm using a low-side driver to drive the coils, uln2803a IC to pull down 5v low. ignition is in "inverted" mode in ignition setup.

so the startup time impacts that portion of the init sequence

in this video:

the LED array on the left is ignition coils (L2=cyl1, L3=cyl2, L4=cyl3, etc)

you can see they light up after switch on, and then go dark.

so a small quick blip would be fine, but charging for a few seconds before the "inverted" sets the right output to the pins:

image.png (6.15 KiB) Viewed 1025 times

[AndreyB]

It would be off "startup times difference" topic to start discussing pull-ups etc.

[gurov]

it's not, that's just a way to generate a +5v signal for coils using this kind of driver and using "inverted" ignition setting.

that part works fine, and it works fine once it's initialized.

this is how old megasquirt1/2 would set up smart coil ignition from the front LED that were driven by low-side drivers.

regardles of this particular hardware setup.

problem statement: "Inverted ignition" setting takes too long to apply to running pin state

expectation is that default pin state would be the very very super first thing configured after power on, maybe right after saying hello, because it's set to inverted for a reason.

[AndreyB]

At the moment that's DEFINITELY TOTALLY NOT THE CASE, PROCEED WITH CAUTION.

[Dron_Gus]

Good practice is to have HW protection against unwanted state during power on. So cpu can stay under reset/flashing state without any hw harm.
Default pins states are defined in STM32 manual.
Yes it takes some time for FW to start, init, load settings and apply this settings to pins.
If reset state of STM32 gpios cause some unwanted activity on ECU output - better to review HW, add inverters, pull-downs/pull-ups, EN signals and so on.

[gurov]

the startup time difference is solved by soldering sb10 on disco board bottom:

https://rusefi.com/forum/viewtopic.php?f=4&t=2747