OK, it's shorted so we'll see is the message stays off as it should.

While I was at it, after I removed it, I connected a continuity tester to the sensor. When I pass a magnet up and down the 2" post the switch reliably opens and closes.

Interestingly the switch closes only when I pass the magnet just above the 1/2 point on the post. Below that, it doesn't close, above that, it won't close either. That tells me that the magnet needs to float at a relatively narrow height relatively to the post.

Also, I need to keep my magnet within about 1/2" of the switch post for it to activate the reed switch. Further away and it won't close.

Now I'm wondering if the magnet isn't coolant-soaked and weak.

 

I tried something similar. I used 2 different magnets, one was a weak refrigerator magnet (like the ones you get from RockAuto) and the other was a strong magnetic pickup tool. The weak magnet had no effect regardless of where I placed it.

 

Does any of this information help?

Link: JagRepair.com - Jaguar Repair Information Resource

I recently replaced my tank because it was leaking around the channel that housed the level probe and giving bad indications. I took it apart to see what was taking place and I agree with the float deteriorating over time. As for the tank cap it is one of the most misunderstood devices on the car.

 

Gus, I did check out your repair site. Thanks.

I jumpered the the connector and no message so far. I do believe that it is the float.

 

The magnet float will only rise about an inch off the bottom. Maybe even less. It will NOT rise all the way to the top with the coolant. Rising to the top is a total misconception. So for the float to drop enough to cause the switch to open you would have to be almost dry in that expansion tank. I know this because I drained my tank and was trying to replace the magnet that fell out. raising the float with a long rod I found it would only rise a small amount before hitting a stop. Those that say you have to keep the tank completely full to keep the light off have some other problem. The tank being a little down is not the problem in their case.

 

Are you sure the magnet is in the float?? Or is it maybe sticking out the bottom part way. If you take the coolant out of the tank, about a quart, you can see the float. You can also see the little magnet, about the size of a pencil lead. The magnet is stuck in vertical and should be flush with the top. There have been cases where the magnet has fallen out the bottom of the float and is laying on the bottom of the tank. Mine was one of those. Not a good design.

 

Scott,
Can you do your continuity check with the sensor in the tank, moving the float up + down.
If it goes O/C with the float at the top then you've got EZdriver's slipped magnet problem.

 

That procedure I described in #40 above tests for this and several other things all at once. If the car passes, we can rule out a slipped or weak magnet, a bad reed switch, or bad logic circuits. No disassembly required.

A variation: instead of reaching into the tank to move the float up/down, we could use a turkey baster to add/remove coolant, watching for where the low coolant warning shows up. The car does not have to be running but the ignition has to be on. This way we're also testing for a float that's sticking or not floating.

 

Sorry, Dennis, should have referred back to your post.
If it passes your test there could still be an intermittent temperature related bad earth or contact but I find it hard to believe it would be so consistent.
He's not getting the problem with the connector shorted so that pretty much rules this out here.

 

Steve,

Yeah, the damned intermittent stuff is always in play. It can frustrate the best of testing procedures.

One other thing I noticed in testing that I should have mentioned: The low coolant warning is self-clearing. It goes away (after 15 seconds or so) when the float rises without having to turn the car off and start over.

 

Thanks, all.

I shorted the circuit and,there is no message, so I conclude that the harness and communications are OK. I previously tested the sensor - it is OK. I have plenty of fluid as per the discussion. So, it is likely the magnet/float.

Thanks EZDriver for the description of the magnet - I had envisioned that the entire float was the magnet. I have no idea if the "pencil lead" magnet is in place or not. I cleaned the tank however, and shook the beans out of it and drained it several times, so it may be long gone... I remember someone else posting that their magnet fell out but I could not understand how it did that. Now I know.

The thought about the float rising only a small amount fit perfectly with the switch test I performed - indeed the tank needs only to be partitally filled to be "full". No need to overfill it.

I will pull the tank again and search for my magnet...

 

Well, I drained the expansion tank but left it in place.

I can clearly see the orange/yellow float and the pencil lead magnet within it. The float moves up and down it's post freely and hits a hard stop about 1" from the base. The little magnet is working, it will attach itself to a meal probe and lift the float all the way to the stop.

The jump-the-switch wiring test cleared the wiring harness from equation, and the air lock test was negative. The test of the reed switch itself shows that it is mechanically working. Thermostat is newish, as are the hoses. The tank is cleaned and the float moves freely when dry. RealGauge tells me the engine warms properly and maintains a proper operating temp.

Right now my theory is that the float is sautrated (like old carburetor floats used to do) and not rising until the coolant warms in the expansion tank and thins. Until that happens, I get a low coolant message, after it warms a bit the float rises and the circuit closes erasing the message.

So...I filled the expansion tank with clear water so I could see what is going on. Guess what - the float doesn't float! But water isn't antifreeze, I bet the specific gravity of coolant different and it could effect the float's floating ability.

To test this I partially filled the tank with properly diluted coolant, just enough to cover the float a bit. Sure enough it is kind of bouyant but not what I would call really floating strong. I put in a few more ounces to really cover the float. Now if I push it up and down it's post it just stays where I leave it. It seems like the viscosity of the antifreeze solution lightly goops the float to the post. Without help it didn't rise to it's top stop like I think, logically, it would. I could easily see it dropping to the bottom over time and simply not having the float-power to rise in cool anitfreeze. Does antifreeze become less viscous when it is warm?

Maybe somone with a known working unit could verify if the float sits at the top of its travel with tenacity?

Suggestions are welcome!

 

I really, really admire your tenacity and thoroughness.

On my particular car, the float floats "with tenacity" as you said. Not like it's anywhere near wanting to sink. 50/50 coolant mix.

Temperature effect: According to our friend Archimedes, it's the density (specific gravity) of the fluid, rather than its viscosity, that will determine sink or float. The coolant will be more dense when cold than when warm, so the float will not be able to float at higher temperature if it sinks at lower temperature. But the reverse could happen: it could float at lower temp but sink at higher temp. All this goes out the window of course if the float somehow gets stuck on the post.

edit: I should qualify the temperature thing a bit. The float itself will expand ... become less dense ... with increasing temperature. So what I said above about a temperature effect holds only if the coolant expands more than the float does as temperature goes up. I think this is a safe assumption.

 

OK to summarize the 50/50 coolant/water has to be at the correct level and specific gravity in order for the magnetic float to reach the top of its travel in the tank.

From what I've read, different types of coolants may have different specific gravities. I have been adding distilled water when doing my testing. I don't know how much, maybe a quart.

Maybe a drain and refill with the correct coolant might be in order.

BTW I'm a retired banker. This is rocket science to me!

 

I really think you have solved a long existing mystery. The float will become "water logged" with time and won't work properly. That make sense in your case and in other cases as well. Like some that say they have to have the tank completely full to get rid of the message. They are instead just running into a logged float. That could also explain why the magnet falls out. The magnet falling out could be due to the float slowly disentegrating. This could be explained by the fact it is built by Jaguar and they didn't test it and didn't use the same float they used on earlier cars. My 95 XJ6 works perfectly with the same sensor and with 119,000 miles on it.

You done good Kid.

 

THATS' IT! My float is just barely floating (in the 50/50 mix). After 12 years it has become saturated and has thus lost it's calibration, so to speak. Unable to float properly on its own, it relies on the changing specific gravity of the antifreeze to eke it's way up to the proper operating level. As long as the fluid stays warm - and less dense - it will operate OK. Once the fluid cools and becomes more dense the float struggles and the problem reappears. This fits the facts of my experience exactly.

I didn't hang with Archimedes but understand what you said. In proof of your observations regarding temperature and specific gravity - the on/off cycling of my Low Coolant message is uncanny in it's consistency. At each cold start this weekend, for example, it came on within 1/4 mile of leaving my house and turns off just as I get to the end of the road. There was practically no variation in this.

I attribute this repeating pattern to the consistent heating of the fluid in the tank and it's effect on the float.

So, it looks like a new tank, float and sensor are the solution...

 

Thanks, but it was a group effort. This is a great forum.

 

Why not make a new float out of expanded polystyrene or similar (the sort of stuff that's used for insulation under concrete floor slabs)?
My guess is that the float is air filled and has leaked or is foam and got soggy.
In either event when it gets hot the remaining air will expand to make it more buoyant.
The SG of 50%% glycol is less than 1.05 and many people in hot countries use plain water as coolant (better heat capacity) so there's no way the float was designed as a hydrometer to measure the coolant SG.
It's just, as EZdriver said, a weak design.
I'm going to mark this thread as 'Resolved' to help others but please let us know the final outcome

 

Yes! Or maybe fabricate a "life preserver" for the current float ... some sort of donut made of very bouyant material to help the old thingy stay afloat.

(Note to Rtucker: I think your issue sort of got lost in this big discussion. Maybe start a new thread?)

 

Nice bit of investigative work and analysis guys!!!