Speeduino

========================================================================================
Rev.A - 4 ignitions / 4 injectors - source on Github --> https://github.com/dvjcodec/ECUTeen_3.5
Rev.B - 8 ignitions / 8 injectors - source on Github --> https://github.com/dvjcodec/ECU_Teen_3.5_rev.B
SCG-ECU1.0 + STM32 to Teensy adapter - 8ign / 8inj - source on Github --> https://github.com/dvjcodec/SCG-ECU-1.0
========================================================================================

After reading many topics, I decided to build my own board, based on informations from this Forum.

Nice. Curious a couple things about your MOSFET outputs; the injector circuits have LEDs on the signal (not output?) which along with the gate resistor (?) would allow just over 4V gate drive, and also reduce capacitance discharge rate. Any thoughts on coupling the LED from Vbat to indicate actual MOSFET/circuit operation, increasing voltage and current to the gate for better switching and lower Rds? Alternatively, use a little MOSFET driver and no gate resistor for quicker solid switching either way? Perhaps diagnostic LEDs on the other MOSFET outputs also? I am also a bit concerned by the low-ish 30V rating for inductive loads that flyback much higher. Any calc's for the added package heat from avalanche or slower operation of (unkown) inductors with a 30V avalanche?

David

This design of Mosfet's chanel I was test a lot of times in 3D printers (PWM for heaters, fans, etc.). IRLR8743 work good, without overheating up to 12A current via 12V . I have many IRLR's from other projects, so want to use. In waterproof enclosure led's are invisible, so don't need to solder in future.

I also think you'll need a linear regulator in there to smooth the power supply, i did what you have there and the analog inputs would jitter from the switching of the buck power supply .

You should have a filter on the input of the switcher, it will put a lot of noise on Vbat otherwise. You should also read up on automotive transient protection, you need to improve the input protection.
The output FETs are too low a voltage really, it should be at least 45V or you will have all your outputs on during load dump. Also, in many vehicles the ECU is the central clamp in the event of load dump, without that you run the risk of destroying other electronics as well as your ECU.
Consider adding reverse voltage protection to all the circuitry rather than just the regulator.

Thank you. Work to do

I think it would be worth putting 8ch of spark and injector on here as the teensy will offer the ability to run full sequential on an 8cyl or more .
I know the code doesn't support that now but if a board was made that could , then more chance the firmware can evolve

Hi

Nice design!
As JHolland said, you need to take really care about buck psu, because these components are "noisy".
What about 3.3v PSU, not yet into the schematic? This PSU need to be clean too, because you"ll feed the analog lever shifter with.

Just one detail about 10K serial input resistors for the MAX9926. VR signal could be in the worst worst case around 300V, so, for a good heat dissipation, it seems the best way is to go for 2 serial resistors of 4.99K, with a minimum size 0805 package.

Will change to linear 5V and add 3V3 supply. Thnx for help.

don't change to linear, just add to the final output . the buck is great for bringing down the 13+ volts, its too much for the linear to drop 13+volts to 5 volts.
i set my buck to output 7v to feed the linear regulator, no heat generated