I have a '03 NBFL miata (or mk2.5 in Europe) with a pnp speeduino having random rpm spikes up to 13k+ (mostly during high load) causing frustrating hiccups. With the developer of the board we tried every software trick there is. Nothing. Seems to come back to the ckp/cmp triggers of a NBFL (4tooth on the crank and 3 uneven on the cam) being a noisy and sensitive solution foor the arduino.
So i started wondering with all those pnp speeduino mx5 kits around. Is everyone running a aftermarket crank wheels and other means to cure this? Or is everyone running fine on the stock triggers and is my pnp board faulty by design? This board works perfectly everywhere other then the rpm spikes, but i would hate to buy another and have the same issue again.
So my question is: are there people out there with a NBFL running perfectly fine on stock triggers/wheels (ckp/cmp) and wich pnp speeduino are you running? Or what have you done to make the stock triggers working?
As far as I know the NB2 and NB1 (or however you want to refer to them) use the same trigger setup. IIRC in the Speeduino, that's referred to as Mazda 4+3 Trigger or something similar.
I ran a speeduino, specifically the SpeedyEFI PNP, on my NB1 with stock triggers for 6 months or so without issues. I actually ran 2 different physical ECU's and they were both fine. I swapped to a 36-2 trigger wheel after that, and that became a big headache that required custom firmware to support. There was supposed to be a firmware update to support that, but I'm not sure if that's happened.
Posting your tune here might help. Have you tried adjusting the trigger filter settings?
As far as I know the NB2 and NB1 (or however you want to refer to them) use the same trigger setup. IIRC in the Speeduino, that's referred to as Mazda 4+3 Trigger or something similar.
I ran a speeduino, specifically the SpeedyEFI PNP, on my NB1 with stock triggers for 6 months or so without issues. I actually ran 2 different physical ECU's and they were both fine. I swapped to a 36-2 trigger wheel after that, and that became a big headache that required custom firmware to support. There was supposed to be a firmware update to support that, but I'm not sure if that's happened.
Posting your tune here might help. Have you tried adjusting the trigger filter settings?
Goor to hear the speedyefi worked solid. It's indeed the 4 (ckp) + 3(cmp) trigger. What made you want to switch to a 36-2 wheel?
I tried every trigger filter setting i could find (from off to aggressive) and all combinations of trigger edge settings (rising/falling). I'll go get my tune and some datalogs
This old thread https://www.msextra.com/forums/viewt...=28944#p192331 (from post 3 onwards) seems to point out the problem for megasquirt users back in 2008? Is this a miata specific trigger setup incorporated in most pnp boards or is there already a software fix for it in speeduino? The board i have had this setup for the ckp/cmp inputs (see the picture)
The 36-2 trigger wheel is a pretty common upgrade. It just gives you a bit more resolution the crank position. I think the wheel actually came with an aftermarket damper I ordered so I threw it on. I didn't even consider that Speeduino might not be able to control it because it's such a common upgrade for these cars.
Are you in the UK? Is this a 1.6 NB?
Looking at the logs, there's no sync loss that I'm seeing, so I don't think your trigger settings are the issue. What version of the Speeduino firmware are you running?
I know it's a pretty tired suggestion at this point, but I would probably be doing a sanity check on my grounds and wiring just to be safe. ETA - I noticed your AFR seems to have some significant spikes at times as well which is why I'd be looking over wiring/electrical health.
ETA - I've also found a couple threads on old Speeduino forums which suggest going to semi-sequential for an issue that sounds identical to what you've described. I know you've already tried semi-seq based on the datalog names, but I'd probably be trying that for now.
Looking at the logs, there's no sync loss that I'm seeing, so I don't think your trigger settings are the issue. What version of the Speeduino firmware are you running?
I know it's a pretty tired suggestion at this point, but I would probably be doing a sanity check on my grounds and wiring just to be safe.
I'm in holland and yes it's a 1.6.
Yes it's very strange to not see sync losses. But seeing the random 13k spikes (random in high load and not there when cruising) having tried different firmwares (currently the latest) and all filtering options.
I will be testing grounds, but seeing it's an pristine unmolested car with untouched wiring i did not expect anything. For piece of mind tough i will
Gotcha, that's a pretty important detail. I haven't done anything with the 1.6L NBs. Just based on what I've seen on forums and such I'd probably play around with semi-sequential more and see if that fixes it. @AlexE seems to pop up in a lot of places where issues like this are mentioned as well, so maybe he has some ideas.
the 1.6 in the NBFL is almost identical to the 1.8 VVT, all the same sensors/wiring etc. minus the VVT and displacement ;-)
i found the semi-sequential suggestion of AlexE and tried it in multiple variants (see datalogs), not a single difference. The producer of my board is now making a small daughterboard with the suggestions of the above old thread for filtering out the miata specific noise present in our beloved cars/wiring.
Anyone else having good succes with PnP boards on a NBFL miata?
yes, it seems schmitt trigger on CKP and CMP inputs solves problem.
I based on circuit by Abe Mira, can be found in Megasquirt threads. This has some drawbacks, for instance using TL082 is not a good idea, LMV358 is better. It also inverts signal phase making engine terribly hard to start (but in case of mk2 NB mk2.5 NBfl you can switch trigger edges from rising to falling, then it works).
also I changed pullup resistors to 1k, and filtering to 1k / 4.7nF
on oscilloscope it still has some flaws but engine works well
