If I remember right, you had done a test earlier where you shorted across the sensor. Absent a weak magnet, or one that has fallen out, I think this is pretty much the same test as trapping the float in a high position. But it "couldin' hoit", as they say across the river.

I remember you reported the float booster falling off at some point. Are you sure you get the current results with the booster in its correct place?

Try this one on: If the float were somehow saturated and so at just about the same density as the coolant, and so just barely able to float ... I can (by squeezing my eyes real tight) imagine a case where as the thermostat opened, the tank fluid for a time became warmer than the float itself (and so less dense) and for that time the float could go from just barely floating to not quite floating. As temperatures equalized, it would float again. A booster should fix this case though, so maybe you've already ruled it out.

You should get some sort of award if you solve this.

 

HA! I had forgotten about shorting across the terminals. Yes I did that to rule out some problem with the wiring elsewhere in the system.

I thought the booster float was a non-starter because the problem remained in the very first test drive after I installed it. However, it is curious that it worked itself loose soon after - something was certainly amiss there. To be prudent, I will retest.

At one point I had envisioned a thermal type problem with the float as you suggest but I'm not smart enough to really think it through as you did. The scenario you picture could easily be happening with my weak-floating float. Another good reason to retest the booster idea.

Has anyone replaced the tank/float and had the problem remain? I don't think so and this would most certainly point, again, to the saturated float idea.

Expanding on Gus' idea to paint the float, how about using crack filling foam-in-a-can to daub on the float. Maybe 3 blops evenly spaced on the sides of the float to try and keep it level and not bind but none on the top so it won't interfere with the top-stop. Time to test and see if it floats...

 

We are getting a little off track. Let me go through how the float works.

The float will only move off the bottom of the tankabout an inch or slightly less. At that point the magnet in the float closes the sensor switch and with the switch closed there is no warning light. The float will NOT go any higher no matter how much coolant is in the tank and it holds about a quart. When the float is floating as it should the tank has to be almost empty for the float to decend and cause the sensor switch to open turning on the light. With Spurlees' float it would not rise off the bottom of the tank until the cooland warmed up. His float has marginal floating capability as he determined and described in an earlier post. Floats better when the coolant is warmed up.

The only problem Spurlee has is getting the float to have more floatation. Nothing else is wrong with his system.

Remember, the float will not rise any further than about an inch, or slightly less, from the bottom of the tank. There is a physical stop on the post that it rides on.

 

Spurlee's symptoms, as I understand them, are a bit more complex: no warning with a cold engine, then a warning appears at about 150 degrees coolant temp, then no warning again after full warm up.

So either the coolant level in the tank really does go down during warmup, or the float is marginal and stops floating ... not when cold, not when hot, but only temporarily during warmup. The second one seems more plausible, but I'm fuzzy on how exactly this could happen.

Spurlee's float booster, if he can get it to stay put, may still be the fix.

 

It could be a thermal thing - easy to play with scenarios but harder to get to the truth.
It can't be a Curie point thing (magnet looses it's power as it gets hot)
It might be a weirdo expansion problem - float buoyancy is marginal, coolant density drops as it warms up, float sinks.
Heat finally gets into float foam, air in foam expands, buoyancy restored.
But then we know that any sufficiently advanced technology is indistinguishable from magic ...

 

Yeah, just about any of the above. I agree, If it were the Curie thing the magnet would not recover when it cooled. So that's out.

I favor "float buoyancy is marginal, coolant density drops as it warms up, float sinks" because its the least whacko of our ideas.

Maybe there is a stray Higgs boson in there weighing down the float?

 

Can't be a Higgs Bosun, I found all of those months ago and put them in the glove compartment.

Another advantage to the "marginal bouyancy" theory is that it explains why this only happens with floats that are old and presumably saturated. I'll bet a Poll would show that the problem is most common in older XK's, regardless of mileage.

Another point is that a new tank with a new float & sensor seems to cure the problem. The only operative element that I can see there is a fresh float since we have eliminated the sensor as a variable.

I will try the booster agan as soon as I get back to the car.

 

Whatever you do don't put Schrödinger's cat in there as well.

 

You're right, I'll collect them all up and put 'em in my Klein Bottle!

 

Be careful, the float will not behave predictably in there.

 

Indeed! It keeps falling out!

 

Not so, it can be completely reversible. This effect is used in many things, including temperature controlled soldering irons.

 

Stumpy,

I believe you are right.

It is possible to irreversibly destroy the magnetism in a ferromagnet by heating it ... i.e. it does not recover when cooled (see last paragraph under "Magnetic Domains" at Ferromagnetism - Wikipedia, the free encyclopedia).

But this it seems is different from heating to the Curie temperature (following paragraph, same URL).

I don't know how the annealing and Curie temperatures compare for different materials, but I suspect the Curie temperature is typically, if not always, lower.

Nice catch.

 

My point about the Curie point is that he gets a fault that goes away as the coolant heats further so it can't be a Curie thing since the magnet would remain weak until the coolant cooled.
We've got a woodburning stove and the ash door was held closed by a magnet.
When it got really hot there'd be a great crash and the door would fall open.
Once cold it was fine again.
Now there's a clip and no more Curie Crashes.

The more I think about it the more this is the least implausible explanation
It might be a weirdo expansion problem - float buoyancy is marginal, coolant density drops as it warms up, float sinks.
Heat finally gets into float foam, air in foam expands, buoyancy restored.

 

It's been a few days running now without the Low Coolant light. I'm pretty comfortable in saying it was the weak/saturated float all along. When I finally got the booster float to work properly the problem went away immediately and has not returned.

A new expansion tank will get you a new float - that's the $170+ solution. Fixing up some sort of booster to help your ailing float is free. I used a 1" plastic ring (tested to make sure it floats!) cut into a "C" shape and snapped onto the post, under the float. It took less than 15 minutes.

Thanks to all who helped along the way, it was an interesting experiment!

 

My foam float (found in a 50 pack of blank DVD's) is still working as well. Can't see replacing the tank unless it leaks.

 

This was a great thread that generally went well above my ability to comprehend...


For a few years now, when the ambient temp falls below the mid-40s, the "Low Coolant" light comes on in my car immediately after start up, and within a minute, the light goes out. The level of coolant is full, so I'm guessing that either my float has lost a little buoyancy or my coolant has lost some of its consistency where the float does not rise until the motor has run long enough to even out the coolant mix.