My ol' Mazda 626 (2.0I 1992, manual no AC) has been running for +280 000 km and is currently having an overheat problem. Something is wrong with the cooling system.

The long story:
1½ years ago I made a major overhaul and lifted the top to replace an exhaust valve in cylinder #2. Cleaned the gunk in the cylinder chambers too ofcourse.

The car has been running fine since that day but suddenly the car started to overheat. Coolant leaked out from the camshaft side of the engine and temperature peaked.
I pulled off the top once again and was surprised how loose the 10 bolts (holding the top in place) were. Coolant/oil had leaked into the combustion chambers. Cleaned it up, replaced the gasket and made sure the top bolts were tightened more than before.

Replaced the following parts to brand new some probably unnecessary but at some point these parts will wear out:
- radiator (had a leak)
- temperature sensor (looked quite OK)
- water pump (no major problem here I think)
- thermostat (didn't look too bad)
- lambda sensor (between engine and catalytic converter, quite dirty but what could one expect?)
- oil and coolant (ofcourse)

This has prevented the engine from immediate overheating and there are no leaks. Still the temperature needle is a little notch too high for my taste (given all the improvements). I sense something is amiss.

The problem:
- The cooling fan does not start when the engine heats up.
- The radiator feels hot on the driver side and cool on the timing belt side.
- Rubber hose from the engine is hot on the driver side but cool at the thermostat.

It feels like the coolant isn't flowing as supposed in the engine.


How is the cooling system supposed to work?
Here comes 1000 questions...
1. In which direction should the coolant flow through the radiator - from the water pump side (I guess!) or from the ignition generator side (driver side)?

2. Is it only the water pump which drives the coolant flow?

3. The thermostat, apart from that the rubber packing around it feels a bit "loose" (cannot be entirely waterproof), should it be mounted with the spring side inside the engine block? Is the placement of the "hole" in the metal dish vital, that is is it sensitive if it has been rotated?


The fan motor itself seems to work ok (connected it to the battery directly, the fan blades starts to rotate). Cannot see that any fuse in the fuse box close to the battery is broken. I have bought a second-hand relay for the fan, there is only one relay in my fuse box. Still to be tested.

Anything else besides the temperature sensor (green plastic thingie which is on driver side of engine block) and the relay which could be the culprit?


Any hint is greatly appreciated!

 

Car is once again fully assembled.
Ran engine warm and successfully adjusted the timing to +12 degrees. The engine now sounds a lot better and idle speed is running normal.

Temperature needle is a few degrees higher compared to normal engine temperature but not alaming. However I have not run the engine for a long time yet...

I have switched to the "new" relay KA30 (light blue in the engine fuse box) but the fan does not start. Cooling system has same symptoms as before. Rubber hoses are hot on driver side and cool on passanger side (close to thermostat). It eventually started to get warm.

Fan problem persists...

 

Assisted with the Haynes manual I made a first trial this evening.
Text in bold are headings, important conclusions and things which I have not figured out yet.

The car has no A/C and manual transaxle. I used the sections concerning cars made in 1993. There is only one cooling fan relay in the main fuse box, I guess #2 and #3 are for cars with A/C? As well as the AIR COND RELAY which is also gone. Further the 15A (R. FOG) and 20A (FOG) fuses are missing...

The cooling fan is single speed with only two wires. Connected them to the 12V battery, black wire to - and Yellow/Blue to +. Dunno of polarity matters here... It immediately started.
Fan motor is OK for sure.

The Haynes manual states that the 40A cooling fan fuse and the 15A engine fuse inside the car should be checked.
I tested all the fuses in the main "FUSE BOX" in the engine room, and all the fuses inside the car using an Ohm meter. All showed 0 Ohm resistance.
Fuses are OK (in the main "FUSE BOX" and at driver's seat). Although some of the fog light fuses _might_ be missing...

Measured the resistance of the ECT (Engine Coolant Temperature) sensor. It showed more than 2,7 kOhm I think, but it is also +5C inside the garage... I replaced this sensor just recently and I believe it is not faulty.
ECT sensor is OK.


That leaves me with five options?

1. The second hand KA30 Cooling fan relay I just bought is also broken?
How should one test the KA30 Cooling Fan Relay properly?
It has four pins, COIL, COIL, N.O, COM.
I understand that the COIL pins are for the control circuit and the other two are the for the current being controlled.
The resistance between COIL pins are less than 100 Ohm and infinitive between N.O and COM (=OK).
Does polarity on the COIL pins matters when testing the KA30 relay?
Which current should be applied to COIL? 12V or less?
I guess the N.O and COM will lead 12V when activated?
The Haynes manual is very vague in this point, but stating that one could switch COIL polarity when testing in case it does not work. Other Internet sources claims that the relay could be blown if connecting polarity wrong...

2. Wiring from the PCM to the COIL pins of the KA30 relay is broken?
Do not know how to test this properly. The wires are hidden under the main fuse box. Not tested.

3. Wiring between the ECT sensor and the PCM (Powertrain Control Module) is broken?
Looking at the wiring diagrams and under the hood, it seems like a big bunch of sensors are sending back signals to the PCM (or is it "ECM - Engine Control Module" as the Haynes diagrams shows?).
The ECT Sensor has a Black/Red and Blue/Yellow wire. As the sensor measures resistance a bad wire here will probably show incorrect values for the PCM, the wires themselves should show close to 0 Ohm resistance?
- Black/Red seems to go all the way into the PCM, to test if it is broken will require very long arms or some help wires... Not tested, but it does not show infinite resistance when I measure this wire against the two pins of the IAT (Intake Air Temperature) sensor. The resistance of the PCM (2.7 - 10.4 KOhm) ? Then the Black/red wire is probably not snapped off?
- Blue/Yellow wire is connected to ground according to the diagram. Just as the Blue/Yellow wire of the IAT sensor. The resistance between these two wires are 0.
Blue/Yellow wire (ground) is OK, the wire to PCM is probably OK.

4. Wiring to the cooling fan is broken?
Removed the cooling fan relay #1 from fuse box, pins mentioned below are the contacts of the fuse box. Also removed the contact to the cooling fan.
- Resistance between the 40A cooling fan fuse pin (closest to the front) and the the cooling fan relay pin "N.O" is OK, 0 Ohm.
- Resistance between the "COM" pin of cooling relay #1 and Yellow/Blue contact to the fan is OK as well.
Wire between 40A cooling fan fuse and cooling fan is OK.
- Resistance between the black wire of the cooling fan to one of the cooling relay "COIL" pins showed 0 Ohm. Dunno if this means that the black wire has ground (I guess it should?).
It is likely that wires for cooling fan is ok.
If removing the cooling fan relay #1, should it be possible just to connect battery + to the "COM" contact of the cooling relay and get the fan going? This could be a last resort, to have a switch between + and "COM" to control the fan in case it won't engage automatically...

5. The PCM is broken?
Well, how to test this? Saw a notice about removing the PCM, open it up and look for leaky capacitators visually... Anyone who have done this before?

To be continued...

 

Did you get the head machined when you replaced the head gasket?

 

I had the top machined properly at the episode 1½ years back when one of the broken exhaust valves were replaced. However the "bottom part" was not machined as I did not remove the whole engine from the car, just the top. Just cleaned it well.
I guess it will make it harder to get a good seal but I hope the surface is not too warped...

 

The last head gasket replacement - no machining...
I think overheating in combination with badly tightened head bolts could have caused the leak problem (which comes first?).

 

Did you replace your headbolts? Some are torque to yield (they stretch) and will not hold the proper clamping force if used again.

 

Yep, during the first major overhaul (1½ years ago) I used new head bolts. This time I reused them, I think it should be OK? Next time though I will discard them...
- Cleaned the bolt holes better this time
- Did not apply any copper paste on the bolts (might have done it last time, perhaps copper paste shouldn't be used on every bolt...)
- Used higher torque.

Thanks for your input BTW!

 

Good question. A.: either the chicken or the egg.

I am afraid you have to do the head job again. Have the head machined and use new bolts or studs. Do not guestimate on torque or anything else. You put this thing together under ambient temperature. But the assembly has to withstand high temperatures and high pressures. The torque value has been engineered to account for that. The only way you get the torque right is with new studs and new nuts if they used "yield to torque studs". (I don't think there are bolts). If they used hardened non-yield studs you must thoroughly clean the the threads and chase them with a thread chaser die. A thread chaser die is different from a thread cutting die! If there is friction between the nut and the stud it throws off your reading on the torque wrench. It will not be tight enough. Even worse, each stud will be tightened to a different value. Make sure your torque wrench is good! They go off calibration if they are not handled properly.

If you cleaned the block without causing gauges you should be fine there. Use a straight edge and a light to check if the deck is good.

 

And when you use the new bolts just use some light oil on the threads instead of copper paste, a nice 10W or hydraulic oil should do just fine.