Zammish24

Title pretty much sums it up, if I drive nice and calm the temp stays in the middle, when boot it fast driving, boost kicks in the temp gauge starts to go up to the H.... when I slow down the temp starts to drop to normal again.
it's a 1989 eunos with a alloy radiator and the fan works perfectly.
I've bled it for a good two hours and seemed fine, took it back out again and it did the same :-/
what could I be doing wrong?

deezums

iTrader: (3)

If you cut the offending signal ground circuit like I told you to it would fix it.

Your logs look like this, right?





They will look like this, just cut the signal ground OFF the engine ground.

Zammish24

Yes my logs look like that, I cut all the grounds off, and wired in 4 grounds directly going from the engine block to the ecu. problem then is the engine wouldn't start so I had to connect the original sensor ground back up for the engine to start

Zammish24

hello again, i apologise for not listening to your advice in my previous wideband post.... i skimmed over it and not understood it properly. I get it now

what i noticed about the signal ground circuit that it comes out of the ecu as two black w/ green stripe wires, but further up the loom.... the two connect together to become one wire which is grounded near the cas by a 'push on' fitting.
if i cut this wire off you mensioned there should be no continuity between the framework and the wire, if there is does that mean i have a faulty sensor somewhere?
also i am going to move the wideband feed and ground wires directly to the battery (with a relay) hopefully this should fix my issue?

deezums

iTrader: (3)

Please see page 4 of this PDF. You can see the signal ground wires on ECU connector #2 pins C and D. See it leave the ECU, run straight down then splice over into the other signal ground wire? That's pretty much exactly what you are seeing on the car. I believe it's a redundancy loop or something, there should never be any large current running on the signal ground.

https://s3.amazonaws.com/miata/miata...g+Diagrams.pdf

For some reason Mazda grounded this loop by the CAS. It's putting lots and lots of noise into the megasquirt and it's causing all of your sensors to read funny.

As soon as you cut that one wire by the CAS (not all four!) you shouldn't see continuity between signal ground and the engine block any longer. If you do for some reason, I would diagnose by unplugging sensors. While the megasquirt is plugged into the ECU connector you will always read continuity between signal ground and ground, so make sure the ECU connectors are unplugged for the test.

Before you move the wideband I would try removing this wire first and see if it helps. It will make a massive difference to all of your sensor inputs, and you might not even need to change the wideband wiring any longer.

Zammish24

Thank you so much deezums, any reason why mazda grounded it in the first place?
hopefully this should fix my cutting out issue that's really getting annoying

deezums

iTrader: (3)

I'm not sure why they grounded it, but my guess is the ECU in the 1.6 cars is so primitive it doesn't bother it.

I guess it stands a chance of fixing a cutting out issue, but as I recall the CAS doesn't use a signal ground. If it doesn't I take that to mean it's capable of dealing with lots of noise. Still worth it to fix the coolant, TPS and other input noise.