[mattt_damon]

The MAX992x chips use both zero-crossing reset and threshold triggering. What this means is that it uses a negative voltage (below zero-crossing) to "reset" and assume the next positive voltage is a real trigger pulse. The issue created is out-of-sync + or - "ghost" signals changing the shape or detection of one pulse or the other. This is what you see in the tooth logs.

Some have had success with MAX chips and coupled-ground sensor setups (@jonbill?), but many or most have not. Unfortunately, we have not found a consistent magic combination that works reliably. If anyone can add to the info base, it sure would help to solve more user issues. Because the sensitivity of MAX chips to this issue (and a couple others), the most popular alternative as @pazi88 mentioned is a brute-force conditioner, such as the DSC, which was designed to not be tricked by coupled-grounds, shallow gaps, and other OEM oddities on Toyotas, Subarus, etc.

Thanks PSIG that was a good explanation. I purchased a DSC from https://diy-efi.co.uk/product/speeduino ... onditioner and realised your logo is on it! leds are a nice touch.

It appears to have solved the ghosting you mentioned, but am not getting sync/rpm still. I've attached two csvs of cranking logs. Watching the gauges it flashes sync and flashes to 40ish rpm occasinoallu but thats about it. I've also attached my tune. If you can speculate what I'm doing wrong or next steps that would be very appreciated.

Thanks again

[PSIG]

Is this on simulator or working engine? We need more info to diagnose this. While it appears to be a decoder-to-signal mismatch, you need to verify. I suggest if on-sim, the patterns may not match each other, so either compare them or move to the engine. If on-engine, I would pull the FP fuse and spark plugs, and ground the plugs together to a solid ground for cranking tests. This will help save the battery, but also smooth crank rotation for best readings. Log all tests, and 'scope them if you can.

Reading the Subaru 36-2-2-2 decoder page again, there are some confusing issues that may be conflicting or confusing in your relatively unusual setup. For example, we don't know if the decoder was configured for VR or Hall, and if any conditioner inversion (no leading or trailing edge info), nor if you have the H4 or H6 variant. So, some suggestions, but try each one-at-a-time to check effects. Restore any changes (except the first two) before moving to the next test:

• Suggested Trigger Filter is Weak.
• Remove the cycle (cam) sensor connections temporarily, as that decoder does not need them and may be confusing it. While testing disconnected, use a 1k to 10k resistor on the cam input to any ground point to hold the cam signal LOW.
• Verify your pattern is 36-2-2-2, and TDC#1 is on the 4th tooth after the first gap (correct wheel and placement).
• The decoder states the common sensors for this pattern are Hall, which may output a reversed signal edge vs VR. Try Rising edge.
• Likewise, the polarity of your early VR may be reversed. Swap your sensor input wires as a separate test.
• Try an earlier code version (pre-202103, H4/H6).

Sorry I don't have a shorter or more specific list of suggestions, but at this point I can't tell if it's settings, hardware mismatch, or decoder mismatch yet. I hope others with 36-2-2-2 experience can help here, or older posts on the topic. If you can find a running example with specifics for components and settings, we may be able to reverse-engineer the setup to use your early components.

[mattt_damon]

Thank you @PSIG! I just achieved cranking. Your tips definitely helped. The issue was 6/7 pattern, and needed to reverse polarities.

Tips for anyone stumbling across this thread in future:
- Don't assume your trigger pattern
- SOHC/EJ201 uses 6/7 trigger pattern well into the 2000s (mine was 2004)
- Reverse the polarities if you can see both cam and crank signal but no rpm