Speeduino is now on Github Sponsors (Rather than Patreon): https://github.com/sponsors/noisymime
Help with building your Speeduino, installing it, getting it to run etc.
By kettlekev
I've spent the last few days hunting down noise issues and have come to this test setup and conclusion.

I am running simple Alpha N with ignition only.

I have the speeduino running from it's own independent power supply to remove PSU noise issues.
I have tunerstudio running with the 5v USB line disconnected again to reduce noise issues
My sensors logs look OK with IAT and CLT resistor grounded to give good readings.

The engine starts and runs for a short period. As far as I can see on a short log, the trigger log looks good.

Engine dies, tuner studio looses comms and needs to reset port to start again.

Speeduino needs power cycle then engine starts again.

If I run my speeduino on speedy sim it works forever.

There is still probably an ignition noise issue getting back in and it seems to lock the arduino up somehow and/or my VR conditioner is having an issue for some reason (it is a falling edge trigger from DIY-EFI). If I have a polarity issue there, what would be the symptom, given my trigger log looks good and why would it take out the serial line to tunerstudio?

If I disconnect the PC, the behavior is the same. Power reset and the engine runs again for a short period.

Below is my basic setup...
overall.PNG (201.72 KiB) Viewed 998 times
By kettlekev
Its a brand new one that came with the speeduino and it works for hours on speedysim.

It must be something fundamental to drop out so quickly......

I have a scope on D18 which is the VR conditioner input to the Mega and that is fine and still runs even after the ignition has stopped and needs the power reset to go again so all the way to the Mega the trigger signal is good.

I have also scoped the 5v in speeduino line during engine running and after it stops and that is good as well so nothing is shutting the 5v regulator down.

Something is causing ignition to stop and the serial port to drop out.....

By theonewithin
It's not ignition stopping and serial dropping out.

I would say the CPU is crashing.

Post tune and log as always when it happens.

Oh and why separate power supply?

I hope you have connected grounds between car and Speeduino at least.
By kettlekev
Separate PSU was an attempt to keep any noise of the speeduino power lines and yes earths are linked.

I agree the CPU is crashing and therefore takes out the serial. I am struggling to get a log because of this but actually made an improvement today.

If I add additional screening (and earth one end) of the cables to and from the ignition module, it seems to help (runs for longer) but still not perfect.

If I scope the speeduino internal 5v I can see noise on it and if I scope the coil 12v feed I can see the same noise there so I think the noise is coming back across the coil from the HT side. Coils are well earthed.

Looking at the V4 circuit, should the TC4424EPA require a decoupling (bypass) capacitor? Would that help reduce interference being transmitted back down the LT lines? As my additional screening today was in this area, is that a clue?

I managed to capture on short log (rename as .mlg) . The engine runs worse with tunerstudio attached. It clearly shows sync loss but I just cant capture a trigger log because the serial line/cpu keeps dropping out. If I do one without the coils firing ( just on starter), the trigger log looks OK.
(60.64 KiB) Downloaded 32 times
Last edited by kettlekev on Sun Feb 14, 2021 10:59 am, edited 1 time in total.
By kettlekev
ignmod2.PNG (31.82 KiB) Viewed 888 times
Standard setup. Speeduino feeds primary via the FET driver

and here's the trigger detail.
igntrig1.PNG (98.19 KiB) Viewed 888 times
Its definitely noise crashing the CPU somehow. If I just run it on the starter motor (no coils power) the mega stays up and the serial comms is solid and I can grab logs. With the coils firing all that gets intermittent and there is noise on the speeduiino 5v rail.
By kettlekev
I see on the data sheet that they recommend a 0.1uf cap close to VDD and ground on the trigger chip.
(338.98 KiB) Downloaded 35 times
I will try some ferrites on the power feed as well.

Does anyone know if all spare inputs have the pullups attached? That would help.

Any other pointers? I've just noticed the circuit I posted is V4.3 and I have a 4.4c. is there a link to a new diagram anywhere, cant seem to find one in Google...
User avatar
Hardware references are always in your code download from GitHub, such as here: https://github.com/noisymime/speeduino/ ... 0.4/0.4.4b Download it or look at it online

The latest code is only updated to v0.4.4b schematics, but that may be of help to you, and the labeling indicates there are no major changes to the basic v0.4.4 design.
By kettlekev

Brilliant. Thanks.

Looks like the additional decoupling cap has been added to the igniter circuit now so I am going to start with ferrite coils in all the trigger and power lines to reduce the noise.

Its definitely the noise that is screwing the CPU, as I add/takeaway screening I can see it become more/less reliable and without the coils firing, everything stays locked.

I am wondering if more decoupling nearer the Arduino 5v will help as well?

I guess need to go through the code and look for pin pull_ups being configured as well, although I cant believe they wouldn't have been.....
Boost By Gear

well good to know , I haven't seen it on github or[…]

what I can't understand is what you mean by &quo[…]

In a general sense, the reliability of things like[…]

As far as I know all NA 1.8's are sequential. Does[…]

Still can't find what you're looking for?