Coolant system.

[magnusfeuer]

So yesterday I took the Caterham out for a test drive, and it promptly overheated since I hadn't topped up the system prior to taking off.

Coolant loss has been a minor issue ever since I built the engine. The raceline water rail has a relief valve in its cap that opens at 20 psi to let excess water out from the system into an unpressurised catch tank. The hose enters the tank, which is always about half full, at the bottom. When the engine cools the coolant system sucks water back from the tank back into the main loop. Or so it is supposed to.

Every time when the car starts to heat up, I smell coolant-on-header for a while indicating that the system leaks during its heating phase. After a drive I have to add a bit of ounces of water. Yesterday I'd forgotten the adding part for the third time, and things went south.

Today I topped off the coolant, bled everything and fired up the car at the driveway while going over it with the IR thermometer. A drop or two of water from the thermostat housing gasket, but not enough to explain the loss. The radiator, once the thermostat opened, got evenly warm. The fan kicked in (in the correct direction) at 90-95 degrees and drove the temperature down immediately.

In conclusion the system seems to leak only while driving, although I cannot see how. Everything is sealed, and the catch tank vents at its top. I just have to drive regularly for this to happen, no Gs necessary.

So my question is, what happens if I eliminate the catch tank setup entirely and plug the relief valve hole and replace the cap with a regular radiator cap?
Alternative two is to plug the hole and setup a pressurised expansion tank.

Suggestions are most welcome.

/Magnus F.

[slomove]

I have the same system with Raceline rail (although mounted on the proper left hand side) with an 800 ml stainless drinking bottle as expansion tank and it has been working perfectly for the last couple of years. Long distance touring or track days over 100 degrees. I use a 16 psi cap but that should not matter much. It always sucks the expanding water back and usually I top off an ounce of coolant once or twice a year.

2 things you MUST do (never mind if you did already):

- drill one or two 1/8" holes into the thermostat flange to allow a minimal ongoing flow even if the thermostat is closed (easier to do than a small bypass hose). If you don't the remote thermostat will never see the engine temperature and and stay closed until it boils and the steam bursts through. Temperatures will fluctuate wildly and some water puke out.

- Look at the lower and especially the upper sealing surfaces of the thermostat housing to the cap. At some point I had the aluminum a bit corroded and the white salty crud prevented a good seal with the cap. Some sand paper (after removing the housing and rinsing afterwards) did the trick. If the cap seal is not perfect it will suck back air instead of the stored coolant.

[magnusfeuer]

Thanks Gert,

The thermostat has a bleed hole, and it works since the hose from the water rail to the radiator gets slightly warm before the thermostat opens.

I'll check cap seal tomorrow.

/Magnus F.

[Michael Murphy]

So my problem is that the coolant temperature fluctuates constantly while driving, once everything is warmed up. During a drive last weekend I watched the temperature gauge cylce from ~110 deg to ~85 deg consistently over ~40 second cycle. This went on for half an hour or so, long after the car was warmed up. It was a cool to cold day. I replaced the fan-switch-sensor thingie in the radiator several months ago. I think it is set to switch on at around 90 deg. Is the problem that the fan is too effective in cooling the fluid? Why does the temp go to 110 before things start to cool down? Should I just blame the English?

[slomove]

Michael, what is your engine and coolant plumbing configuration? Is the coolant sensor a direct switch to the fan or does it go through the ECU? Another dumb question: does the fan go off to start with or is the fluctuation just caused by the thermostat? Oh, and one more...did your car always do that or is it a recent behavior change?

[Doug Liedblad]

He doesn't have the bypass holes in his thermostat.

It's also possible his thermostat is bad. Mine behaved similar when it went.

There is air trapped in the system.

[Michael Murphy]

Gert, I'm off to drill some holes. That should help. Thanks.

[magnusfeuer]

I think I found my problem today.

I topped off the coolant system, which was under somewhat negative pressure when I removed the cap. This is good news since I at least don't suck in air through the cap as the system cools and the water contracts.

Out for a blat I went. Overheated I did.

What I did notice was that the fan kicked in way late. The temperature at the radiator entry was about 110 degrees when the fan kicked in. The fan is mounted at the coolest end of the radiator, where the temp was still pretty high. I forget the exact value since I was a bit stressed out about warped heads and blown gaskets.

Tomorrow I'll remove the thermostat and check that it works by heating it in water on the stove (with a thermometer). If it is ok, I'll move it to the end of the water rail where it measures the water at its warmest. That should get the fan to kick in earlier.

I'll keep you posted.

/Magnus F.

[slomove]

Yes that is weird.

I suppose the fan is controlled by the ECU but then the position of the thermostat should not matter for the fan to turn on but rather the ECU coolant temp sensor.

Or are you using the electromechanical fan switch on the rail?

FWIW, on my car the fan does not turn on at all as long as the car is moving at least 20-25 mph. The natural air flow is more than sufficient to cool it down.

Another wild guess: do you happen to have an unconventional serpentine belt configuration that makes your water pump spin opposite direction? I have that (to save a bunch of pulleys) and accordingly I need the water pump from a European Escort instead of the Contour where the engine is from.

[magnusfeuer]

As soon as I'm moving, the coolant system goes down to 90 degrees.

The fan is directly controlled by the fan thermo switch mounted on the cool side of the radiator. The ECU has nothing to do with it.

Since the water is at its hottest when it exits then engine at the rear, I believe that the water pump rotates in the right direction.

I'll move the fan thermo switch to the hottest spot (at the thermostat) tomorrow and see what happens.

/Magnus F.

[slomove]

Sounds like a plan.

But why don't you control the fan with the ECU? Then you can program the fan on and off temperatures with the other map parameters (at least on my Emerald ECU I can do that). The ECU coolant sensor sits anyway on the hot engine exit side in the water rail. And you can shed a few race deciding ounces by ditching the separate thermo switch :rolleyes:

BTW, when the coolant pump rotates the wrong way, the coolant flow still goes the same direction because it is a centrifugal pump but the pump efficiency goes way down. But that should only matter when driving really hard.

[magnusfeuer]

I'll dig out the schematics for the T6 to see if I can wire in the fan switch.

Since the temperature comes down so fast when I start moving, I think the pump is rotating the correct way.

This is what the belt route looks like:


Page 5 of Cosworth's Duratec catalog at:

http://25.enable.com/fusion/getfile.asp?k:A3:5F:24:D0:11:8Eg:DC

shows their belt route. The pump rotates the same way in both configurations.

/Magnus F.

[slomove]

Yes, same direction so you should be O.K.

I just mentioned because I got stumped when I bought a replacement pump here that the impeller was curved the other way and ended up buying a new one in Germany.

[magnusfeuer]

Moving the fan thermo switch to the water rail seems to help a lot.
The fan now kicks in 15 degrees earlier, my coolant temp sensor calibration is off, so I don't know the exact value, and it drives down the temperature about ten degrees before turning off again.

The fan may be a bit anaemic, or sit too far from the radiator (1 cm gap), because it takes a while for the temp to come down. I'll bleed the system tomorrow and see if I can get the fan closer to the rad.

At least I don't have to worry about overheating any longer.

I checked my schematics, and I can use either the variable cam or idle air controller PWM to drive the fan relay. In order to extract this wire from the loom, I need help from Adi, who built my harness. I'll get around to that some day.

Thanks for your advice, it really helped.

/Magnus F.

/Magnus F.

[JohnCh]

Magnus, what fan do you have now? My car initially had a Perma Cool 12" fan that was rated at 1650 cfm and 7.7 amps, but which struggled a bit bringing the temps down at stoplights on really hot days. When investigating replacement options, I spoke with the folks at Flyin Miata who had bench tested the Perma Cool and Spal fans and discovered the manufacturer’s flow ratings didn’t accurately reflect the real world differences. Their results were:

Spal 30100178, rated 1390 CFM. Tested 1975 CFM, 12.5 A draw
Spal 30101500, rated 970 CFM. Tested 1610 CFM, 8.3 A draw
Permacool ??, rated 1650. Tested at 1520, 7.5 A draw

Based on this information and the fact that the Spal fans have a much bigger amp draw per rated cfm, I picked up a 12†Spal rated at 1630 cfm and 14.4 amps and then performed my own comparison tests. Each fan took a turn attached to the back of the radiator and hooked up directly to the battery. Flow (actually suction) was measured with a Synchrometer held at specific points on the front of the radiator. The Synchrometer readings bore out what I could feel with my hand; the Perma Cool read 9 on the Synchrometer’s internal scale, and the Spal read 12. This difference has been born out on the road.

-John