LukeAPEX

I have an 2005 6 port that's having issues.

Anywhere above around 5000 rpm the power curve becomes chaotic.
Best way I can describe it is that it feels like fuel surge. While trying to accelerate to redline the power will dive for a second and then kick back in and just be jerky in general. The duration of each state (bogged down / normal power) seems to be no shorter than half a second and can be as long as a full second.

• 5000 - 6500rpm it's noticeably down on power.
  
• 6500 - 8000rpm it will almost completely lose power for a second.
  
• 8000 - 9000rpm if I keep my foot down and make it here then the power kicks back in for a second.
  
• Then if I shift up it will be really jerky for a second or two.
  

Those rpms are just a rough guide as it isn't strictly linked to those ranges.
It just helps visualise the time frame that things happen when accelerating hard in 2nd gear.

What I've tried so far:

• 3 different fuel pumps across 2 different fuel hanger assemblies.
  
• Pulled the plugs and gave them a clean. The plugs have done less than 1000kms anyway.
  
• Cleaned MAF sensor with MAF cleaner.
  
• Changed all coils and leads. I had D585 coils fitted and swapped back to OEM to check. No difference.
  
• Scanned for codes. Only code I get is p0410 because I've deleted the air pump.
  
• My Short and Long fuel trims seem to be around +3% while revving. That could indicate something, I'm not sure what though.
  

Potentially important recent events:

• I recently rebuilt the engine when it's compression become unusable. My first big mechanical project, I took my time and spent a few months opening the engine, cleaning everything, checking specs, replacing stuff and then spent a day reassembling.
  In the end I replaced the front rotor as it had spun the bearing, replaced all rotor and stat gear bearings, all new seals and springs. Long story short, it freaking worked and after a few days of gentle run in it was able to hot start in under 2 seconds. It's only gotten better since then.
  
• The weekend after the rebuild I took it to a drift event that was going great despite some clutch slipping. Power felt great, no fuel starvation, no misfiring. So anyway the pressure plate exploded...During the carnage it completely chopped the wiring loom for the gearbox speed sensor (I assume) and post cat o2 sensor. (I can post the detailed version of what happened if interested)
  
• Since then I've repaired the damaged loom, patched up the tunnel, machined the flywheel, fitted a new clutch and pressure plate, replaced the gearbox and also replaced the muffler with another OEM one. (I cut off the last one for the event because last time spectators said they couldn't hear it)
  

My last remaining ideas:

• Maybe the ECU was damaged when the sensor loom was damaged? I have a spare ECU but I'm pretty sure it would need to be programmed to work in a different RX8.
  
• Maybe it's intake/vacuum leak related? I definitely don't think it's the APV as the change in power is so instant that I don't think the APV motor could rotate the tubes that fast.
  Maybe it's the SSV or the other one? But I don't know why it would only be an issue now when it was fine at the event.
  I've ordered a vacuum line tester that should be delivered in the next week or so.
  

This is starting to get me down as I'm spending so much time ruling out the possibilities and not enough time doing my part for the great tyre genocide.

Loki

iTrader: (1)

One thing that doesn't appear on the list of things you've checked is the cat. The previous engine died of loss of compression, which is usually caused by poor ignition and/or failing cat. Have a look inside it, these signs aren't uncommon with a clogged cat.

Fuel trim while revving is not useful, what are the short and long term trims when the car is warmed up and idling?

A couple of other things to check: reset your ESS profile/NVRAM and see if the clutch safety switch is fully disengaging (it's on the pedal). The gearbox sensor that was severed is either the reverse or neutral switch (there is not trans speed sensor). Neutral switch could definitely play into this if it's not providing a consistent signal. How did you splice back the wires? AFAIK it's a resistance sensor so soldering can alter the resistance value of the circuit and mess with the ECU.

LukeAPEX

Sorry, I forgot to mention that my cat is completely gutted.
Fuel trims at warm idle are about 3.5% Short and 0% Long.

In regards to the sensor wire resistance.
Wow okay, that's the first potential cause I've heard that actually lines up with something I've changed since it was running fine.
I did happen to solder the wires, I couldn't get a confident crimp that didn't pull apart.
Some more info on the wires, it's the loom that comes from the top of the engine and runs to the reverse and neutral switches and also the post cat sensor. 8 wires total.
It was cut off right at the top of the tunnel with about 8cm extra hanging over the engine end plate. So it was a bit of a pain to work with while under the car.
Instead of soldering the original wires directly back onto each other, I extended them all with about 6cm of higher gauge wire.

This is starting to look like it's going to require me to pull the engine out again and replace the whole loom. As the part that was damaged doesn't have a plug to separate it from the rest.
Is there some hacky why I can't avoid that?

If it's going to require that much work I might as well stop putting off my dream of building a GSL SE hybrid turbo engine.

Loki

iTrader: (1)

Is the higher gauge wire you used the same metal as the original? You don't want to mix copper and aluminum. If do have to add an extension I would it use connectors with the extension.

You could also sound out the switches with an ohmmeter to make sure they're OK. I don't have a lot of info on their expected values but perhaps you can find that at foxed.ca rx7 manuals (including rx8).

I don't know if I would jump on replacing the loom just yet, let's see if we can confirm that's the issue.

Loki

iTrader: (1)

Ome other thing: see if your calculated load via OBD is stable when the car is maintaining speed on level ground. It should be but if it's not it's another clue.

And not sure if you confirmed but reset the ESS profile.

LukeAPEX

Firstly, thank you for taking the time to help me try and figure this out. I really appreciate it.

I just checked and the wire I used is 1.13mm² which works out to slightly thicker than 17AWG. It's "High quality oxygen free copper" if the packaging can be believed.
Pretty sure the solder I used was 40% lead / 60% tin.
I did the 20x brake peddle reset and confirmed the "oil pressure" needle did it's dance. Unfortunately it didn't make any difference when I took it for a drive to test.

dannobre

iTrader: (3)

Considering that the rear O2 sensor is mostly an emissions thing it shouldn't cause you issues. If you disconnect it that could rule out it giving The ECU bad info....

The reverse switch and neutral switch shouldn't cause issues either.

You sure the revs when damage to the clutch happened didn't hurt the motor?

Still stock flywheel? You didn't go to a different flywheel when you repaired things?
an out of balance engine can act pretty strange at certain rpms... and not as bad at others

LukeAPEX

Here's some more detail on what happen when the clutch failed.
The minor clutch slippage nearly cost myself or my passenger a leg. Towards the end of the day I was initiating a 360 entry into a corner. I was sideways building up some inertia and then I partially depressed the clutch to allow it to slip enough for the rear to regain traction and whip me around into a fast spin. Instead what happen is as soon as I touched the clutch peddle the pressure plate super heated and exploded. The transmission bellhousing now only exists in my dreams and the heavy pressure plate surface has broken into 4 large projectiles. One palm size piece shot straight up in behind the radio in the cabin. If it went to either side instead it would have gotten someone's calf.
This all happened at 9000rpm. Insert 'yikes' face.
The engine continued to run and idle while I coasted my way trying to get a close to the pits as possible. Once I'd come to a stop and processed what had happened I turned the car off after checking that the coolant temps were below what I'd normal leave the car on to keep water pumping.

The flywheel is an aftermarket light weight with separate counter weight.
I've been using this same flywheel since before the incident and had it resurfaced since the incident.
When I rebuilt the engine (a week before the incident) I also installed some extra stiff poly engine mounts. After the rebuild there is a decent amount of low (800 - 1800) rpm vibration but I just attributed it to the new mounts and crossed my fingers that it wasn't engine related.

LukeAPEX

Last night I used the track recorder plugin in Torque to capture the issue as it happens.
I added both the Engine Load and (Absolute) Engine Load to the screen in case there was a difference between the two.
I have the throttle gauge on screen but when it reads 78% that actually means 100%. It's always been like that, even before this issue.
Towards the end of the video I realised it'd be good to see the relative throttle position as a graph to compare it with engine load.
The relative throttle position (R THR %) matches up with my right foot.

TeamRX8

iTrader: (25)

that end of the harness on the S1 is just the neutral switch, reverse switch, and rear O2 sensor. Reverse and neutral switches are simple on/off signals. As Dan stated, rear O2 is a read only emissions sensor that isn’t likely to impact anything. The speed readout comes from the ABS system based on wheel rpm and an assumed tire diameter. Some of what was replied back to you is a bit of nonsense and unnecessarily distracting.

Instantaneous power cuts like that could be electrical wiring or ground related though. Are you sure there wasn’t any wiring damage under the dash? A flaky CAN signal might cause the ECU to hiccup. The Cobb AP hand-held tuner device was notorious for doing that on occasion when plugged in. Swapping the engine harness would be a way to eliminate it from the equation though, but a but of a pain doing that with the engine installed in the chassis.

so the question is; what clutch was that? that it was slipping some after a recent rebuilt should have given you pause for thought. was it the prior used clutch or not? any chance it might have been overlooked for final bolt torquing?

LukeAPEX

The clutch that blew was the cheapest ebay spec ceramic puck I could find. Brilliant idea right?
XTD was the brand. The friction disc itself actually wasn't too bad. But the pressure plate might as well have been made out of sticks and leaves.
It only lasted around 2500kms before the clamping force was a joke.

It's now got an Exedy friction disc and a Direct Clutch (Australian brand name) heavy duty pressure plate.

I'm honestly considering biting the bullet and ordering an Adaptronic standalone ecu. If that fixes it I'll stay N/A and maybe get a 5.8 ring gear.
If it doesn't fix it then I'll be pulling the engine out again and might as well do something stupid like a turbo GSL-SE hybrid with fresh air anti-lag. Just to try and fit as many buzz words in as possible.

I have a spare stock ECU but I'm not sure what is needed to swap them over. I assume it'd require a professional version of MazdaEdit to allow multiple VINs. I haven't done anything with ECUs before.

LukeAPEX

Well I fixed it.

Thanks Team for suggesting grounding issues.
It didn't end up being the problem, but checking it lead me to finding the real issue.

Turns out a piece of pressure plate shot up on top of the engine taking out all 3 vacuum solenoids on its way.
I replaced the solenoids and I've got my high RPM power back!
I also swapped back to my D585 coils and leads while the upper intake was off.

Thanks everyone for chiming in. I'm heading to the track tomorrow!

Loki

iTrader: (1)

Wow. Did not see that coming. So it tore through the bellhousing, ricocheted off the firewall and still had enough energy to wreck the solenoids. Good thing it's just solenoids and not a leg. Maybe invest in a ballistic blanket.

LukeAPEX

Yeah, a lot of energy was released because it happened at 9000rpm.
The bellhousing was pretty impressive, but understandable being aluminum and relatively brittle based on that it shattered without deforming.
But to go through multiple layers of firewall steel like it was a minor inconvenience. That's kinda scary.

Very lucky to come out of it with just a sore left foot from the clutch peddle kicking back as the slave cylinder got owned.
Lesson learned. Avoid Chinesium when choosing your lumps of metal that spin at crazy speeds.