New radiator, overheating
 

I've read years of posts and not found enlightenment. I'll try to make a long story short:
Blew out radiator end tank on highway far from home. Shut down immediately. Found replacement radiator locally (huge surprise) and installed it. Discovered that overflow tank also leaked. Fashioned emergency overflow out of Arizona iced tea can to get home.

Worked fine on highway (50 miles), but would overheat and illuminate 'low coolant' on back roads every 20 miles or so. When this happens, overflow can has 16+ ounces of coolant in it. If I bleed pressure and pour coolant back in, all is fine for many miles. I shut down and allowed to cool overnight - all coolant from can is drawn back into engine.

Since extended highway driving doesn't cause overflow or problem AND because cooling engine draws coolant back in, I'm assuming that it's not a blown engine seal - system seems to be pressure-tight. Cooling fans work.

I did notice that the replacement radiator seemed to be copper rather than aluminum and had much lower fin density than the stock radiator. Could it be that I have a cheap knock-off radiator that fits mechanically but doesn't provide adequate cooling if you don't have a 70mph airflow through it? Any other reasonable theories?

Did you check for a air pocket ?

That's the problem with making a long story short - you end up leaving out important stuff :-(

Yes - I very carefully purged air, several times. In all cases, once it started it and drove, it would go at least 20 miles before any problem, and often 60 or 70.

One other symptom: In most (and maybe all) cases, there would be a very brief temp spike on the temp gauge. It would then return to normal for a few minutes, then climb and trigger low coolant alarm at the same time. The engine at that point is VERY hot, and it takes several minutes to bleed pressure with continuous boiling as pressure is reduced.

The temp gauge is the original non-linear OEM gauge, so it's pretty useless at detecting small changes.

In all cases the engine was shut down as soon as the gauge started to climb.

Temperature spikes usually have to do with the thermostat.... my other thought process being (with coolant NOT being sucked into the radiator from overflow with the system under pressure) is the radiator cap. How deep is it comparatively with the radiator overflow hose? See where I'm going?

My other thought is to get your cooling system checked for exhaust gases. It's specialized equipment, but most repair shops can do it pretty cheaply and quickly- and it will eliminate the possibility of your coolant seals being blown and pushing combustion gas into the cooling system (rather than the other way around.)

During this whole process I discovered that the original cap had failed due to the rubber gasket developing cracks. That led to boil-over within a few miles of spirited driving or a few minutes of idling. It has a new but non-Mazda cap on it now.

I'd assumed that it couldn't be exhaust gases being forced into the coolant since it exhibits no problem during prolonged highway driving, and has enough suction to draw coolant back in as it cools. Wouldn't that preclude exhaust in the cooling system? It seems to me like it's running too hot rather than pushing coolant out at normal operating temps.

It's on the lift in my garage now getting a new coolant overflow tank. I'll probably do the temp gauge linearization and calibration at the same time. I'm OK with getting a better radiator, but I'd like to understand the mechanics of what's happening. It really doesn't make sense to me.

I can see if there's a shop that can detect exhaust gases in the coolant. Would a pressure test accomplish the same thing? I can do that here, but it would be on a cold engine.

Look at AST removal or replacement.
The stock Plastic AST is Crap and prone to problems.
I'm not an FD guy but that may be what you are up against.

I'll expand on 'near stock': I removed the AST and the pre-cat 13 years ago, and added a boost gauge and poor man's cold air inlet along with a silicone hose jib.

The core of the mystery is this:

It's pushing coolant into the overflow tank, followed by a 'low coolant' alarm and rapidly climbing temperature.

It's not losing coolant in any significant amount, and the cooling system appears to be pressure-tight.

Could it just be overheating due to an inadequate radiator or a cap with too low a pressure setting?

Are you running a 50/50 coolant/distilled water mixture ?

Yes. Actually, since the replacement was done in a Uhaul parking lot and UHaul didn't have distilled water, it was 50/50 coolant/city water after the radiator replacement.

Radiator cap was OEM 0.9 bar when all of this started. I assume the replacement is the same, though I didn't see any pressure rating on it.

I'm cursing myself for not having done the temp sensor linearization mod. I really can't say definitively that it's running hotter up to the point when it spikes, though it seems to me like it is.

Here's a picture of the new radiator. Seems like WAY fewer fins, though I no longer have the original.