Okay, so I've gotten my board assembled, wired everything in, double/triple checked connections and I can't get a clean RPM and the engine runs incredibly poorly.

The CPU is MS1/Extra with the latest firmware. I followed Aaron Cake's writeup with the exception of the DB37 pinout.

Zeal Daughterboard Details:
--------------------------------
- I added a molex plug for the CAS inputs because the 2.2 board doesn't have the spare pins on the DB37.
* Red wire from CAS connected through my shielded 18/4 to VR1 and VR1OUT connected to bottom hole of R11.
* Green wire from CAS connected through my shielded 18/4 to VR2, VR2OUT connected to pin 11 of the CPU.
* Both CAS wheels have their grounds connected to the other 2 wires in the 18/4 shielded cable and are connected to their respective grounds on the Zeal board.
* The shield of the cable is grounded only one place - inside the case to the lower end of D1.
- Optional ground jumpers installed
- IN2 is connected to X2 (Fan), OUT2 is jumped to X14.
- IDLEIN is connected to left end of R16 and IDLEOUT is connected to lowest hole of Q5.

Ignition Outputs:
-------------------
- D14 (D17 on V2.2) is connected to X11 and wired to the leading coil trigger.
- D15 (D18 on V2.2) is connected to X12 and wired to the white wire on the trailing coil.
- D16 (D19 on V2.2) is connected to X13 and wired to the red wire on the trailing coil.

Injectors:
-----------
I had two high impedance 12 ohm injectors and two low impedance injectors.
I wired INJ1 to the high imps directly. They are installed as the primaries.
I wired 10 ohm 25W resistors inline with each of the low impedance injectors. They are installed as the secondaries.

Other Info:
------------
The car starts with some hesitancy, then it misfires and stumbles all over the place.

All of the signals seem all over the place, especially the tach. The factory tach also jumps all over the place in a similar manner to the gauge in megatune.

Everything is grounded to an eight gauge wire straight to the negative terminal of the battery.

The wideband is a PLX. It's wired with the included capacitor between analog output and ground but it still seems kind of spastic.

Attached is a log from tunerstudio and the msq I'm using.

Any ideas would be great.

 

 

Might not be the problem but your grounding method might be to blame. I have had the best results with the factory configuration.

MS should be grounded via multiple small wires to the engine. Engine should ground to the chassis with at least one ground strap. Battery should go to the driver's side strut tower, then from there to the starter.

Ken

 

Unfortunately this is the only RX7 I've ever worked on so I didn't know what to check for grounding configuration. The original wiring harness was a total loss due to the previous owner being an idiot.

I was doing some additional reading on Sunday morning and replaced the grounding strap from battery/chassis/starter with a new 4 gauge. I also ran a 6 gauge ground wire from the alternator chassis to the rotor housing. I did find the bolt on the starter end to be finger tight but even after cleaning the areas with sandpaper, installing new ground wires and snugging it all down I didn't get any better results.

I then ran a separate shielded 24/2 for each wheel on a different path back to the MS, grounding each shield at the ECU end and it actually got worse and now doesn't even start.

Tonight I'll run another 18/4 along that alternate path and I'll give the multiple ground wire to the engine tactic a try.

With the zeal board, I see that each VR input has one side tied right to ground... That is for the Ne- and G-, not for the shield, correct? I read 0 ohms between that solder pad and the other ground wires of the ECU.

 

Really the best way to do the grounding is the way the factory did it. You can use new wires/straps, but you should keep it in the same config that they did it in. You didn't mention whether or not you have a ground strap from the engine to the chassis (stock was to the firewall).

You don't want to haphazardly run grounds everywhere, that'll cause ground loops and make it so nothing works.

You want to run each VR ground back to the VR conditioner board. The shield should be grounded on one end at the ECU.

Ken

 

My new main grounds are run in the same configuration as the stock setup. Battery -> fender well -> Starter Bolt. I guess I don't have a ground strap from engine to firewall. There's a spade connector there but nothing was connected to it. Suppose I'll add that.

The only thing I added was the alternator one, just in case the alternator was introducing noise. No difference either way so I took it off again.

Less haphazard and more that I've never worked on an RX7 that wasn't butchered up so I just added one that seemed logical.

Okay good, that's what I did.

 

I guess backwards progress is still progress...

I took the ECU to work today and checked everything out again. I changed the ignition outputs to the molex plug and brought the CAS signal in as follows:
Ne+ on 24 -> VR1+
Ne- on 25 (X11) -> VR1- (gnd)
G+ on 27 (X12) -> VR2+
G- on 29 (X13) -> VR2- (gnd)
Shield on 2

Figuring that the ignition outputs care much less about noise and shielding.

BUT.... now I have no RPM signal. Regardless of whether cranking the car, or spinning the CAS gear by hand or by drill.

Here's what's weird though:
Checked continuity from inside the CAS itself all the way through CAS connector, wiring, DB37, and internal patches to the correct places on the card... Ne+ goes to VR1+, G+ goes to VR2+. Less than one ohm all the way.

Tooth log in tunerstudio shows one bar for each tooth that passes the Ne pickup, meanwhile trigger log in tunerstudio shows nothing at all.

I unplugged the coils/fuel pump fuse, turned on the "wheel stimulator" option and turned the CAS by hand a little bit. Every time a tooth passed the pickup I saw RPM spike in tunerstudio and heard audible clicks from the injectors.

Turned wheel stimulator option back off and no RPM and no injector noise again.

I just don't see what I could be missing here. I mean, I do this sort of troubleshooting for a living and this is a lot simpler than the multiplexed transit vehicles I work on but for some reason I just can't find the problem.

Next step is probably bringing my o-scope home from work and looking at the signal into the LM1815 and back out, although if the CPU sees something in 'wheel stim' mode doesn't that prove the LM1815 circuit inputs and outputs since I removed all the extra components specified in the build instructions ("On a V2.2 board, remove D5, R10, D8, C12, U4, C11, and R11.")?

 

If it works in stim for wheel mode and not when you turn it off, the 2nd trig (G) signal is probably not working for some reason. So my suggestion would be to check that one specifically.

Ken