Originally Posted by firstgear

Had a thought about this tonight while I was driving home....GM says they want just a small square inch bubble in the reservoir.....then I was reading about some of the other peoples experience where the small bubble disappeared when the fluid heated up (expanded).

So I started thinking about the testing and it made me wonder if the stock pump set up needs that very small bubble to facilitate expansion and for the most part force feed the pump......if the bubble is too big the pressure never builds enough in the system on the feed side to the feed side of the pump that it cavitates.....

Cavitation is caused by air in the system where the pump gulps air slugs (very bad) but it also happens when the fluid cant get to the pump fast enough (also very bad). If the bubble doesnt enable the fluid to build pressure enough in the system to feed the pump, then bad things will happen.

You dont see this problem before it is warm because the fluid isnt warm enough yet to expand but likewise the fluid still isnt up to critical temp and everything is fine because IATs are still happy.....when the temp goes up but the pump isnt feed enough all hell breaks loose and the pump goes into hybernation for 3 minutes only to repeat the cycle again as temps arent kept in check.......

just my wild *** thoughts.....

Originally Posted by GSpeed

Windshield time... and shower time... ALways good times to think. Amen to that.

Lets disregard what GM says for a minute- and get back to basics. In your line of work, would that square inch fly? NEVER.....no air is best...none, zero!

From a closed atmosphere system, that sees large temp swings. how do you feel about that? Air is bad no matter what. The expansion rate of air is far cry from fluid....that air at best acts like a spring. But the pressure within the container will never build as long as the air bubble is sufficient size. When the air bubble compresses which is what happens when the fluid expands it will get very small to the point that it is still there, but pressure has built in the system.


Lets also liken this to a water pump on an engine- Same scenario. However, we all know what would happen if THAT system had a critical shut down. That is because the pump can't pump air, it is meant to pump fluid.....

Same environment, different hardware. We all know there is air in that system.

Now that I think about it as I type this up- the Engine waterpump draws coolant from the bottom of the radiator, just like GM OEM Blower pump, draws from the heat exchanger. Same theory, and execution.

We have reversed it, giving it a large tank with head pressure, and pump flows through the heat exchangers, as opposed to drawing through one half, and pushing through the other half (OEM setup)


We hope to validate this tomorrow, or friday at the latest to make sure our updates were not an outlier. .

Originally Posted by GSpeed

Its cavitation, for sure. We ran the pump at 200* for 20 minutes on the bench, isolated from the car- and no issues. The pump is rated to 240, we just validated what we saw on track, temp wise.

The pump fails to flow well during cavitation. We correlate this, with the increase in water temp spikes, in the duration leading up to the cavitation shut down.

We too thought about changing the pump- however, that solves the shut down problem, not the root cause. The pump functions fine, when its operating properly. Being that this kit is expensive as is, engineering a proper solution seemed like a better idea than throwing money at it.

We liken it to not filling the oil tank full, and then changing the oil pump, because pressure is fluctuating...

Originally Posted by Edwardz

Thank you for doing all this work on a great car and it sounds like you're going to come up with some interesting solutions to make it better. Maybe I missed it somewhere but I don't understand what is actually shutting the pump down. The cavitation as you introduce air into the pump leads to shut down. Since it last three minutes and then the pump comes back on it sounds like a very controlled event. What is controlling this I am not sure. Do you know at this point the mechanism that is causing this? Also this may not to be a failsafe mechanism to protect the pump. The reason it may be shutting down is to allow air to move away from the pump. Other non-car applications I know they have a strategy where a timer shuts the pump off periodically to allow air to move away from the pump. Allows the fluid to prime the pump. If the pump is just sucking air it may never prime itself again where is if you shut it down fluid may move back into the pump and then be primed and work good after that

Originally Posted by jcthorne

Perhaps I missed it, but in the KISS vein of thought, have you looked at just re-routing the hoses to allow the pump to pull directly from the stock reservoir and have both halves of the heat exchanger down stream of the pump, thus removing the flow restriction upstream of the pump suction? IE keeping all the stock components, just changing the order?

Originally Posted by GSpeed

We've been developing a new cooling system for the temperature-challenged Z06s, and one odd issue we found was repeated temperature spikes in intercooler fluid temperatures when on track. We attributed it to a pump shutdown for some reason, but were unable to pinpoint the problem until now.

This graph from our Cooling Development Thread shows the temperature spikes we were talking about:

Looking over a datasheet for the pump, we saw that it does have the onboard smarts to protect itself if need be. We thought we were getting an overheat condition, so we heated the water up to 180°-200°F and kept it there for 20 minutes. The pump didn't care, and continued happily pumping water the whole time.

We played with starving the pump, and injecting air into the system, and finally figured it out.

Excess air in the system will cause the impeller to cavitate, and enter a three minute shutdown period to recover. That matches our observations on track exactly.

Here's a video of our test. We inject air into the line with a blow gun, and a few seconds later, the pump shuts off. Exactly three minutes later, the pump restarts.


This is HUGE, and gives much more importance to the bleeding process and tank design. Previously, everyone thought that air in the system just hurt the heat transfer and pumping ability, when the reality is much more dire. The system will shut down entirely for three minutes, which is more than enough time for the engine to overheat and go into limp mode.

What also worth noting is that the ECU has nothing to do with this. Throughout our entire test shown in the video, the pump was hooked up to a battery. The ECU has no idea the pump is shutting down until temperatures reach unsafe levels.

Lesson here: Make SURE your intercooler system is bled properly.

Originally Posted by GSpeed

We've been developing a new cooling system for the temperature-challenged Z06s, and one odd issue we found was repeated temperature spikes in intercooler fluid temperatures when on track. We attributed it to a pump shutdown for some reason, but were unable to pinpoint the problem until now.

This graph from our Cooling Development Thread shows the temperature spikes we were talking about:



We set up a "test loop" to simulate on-track conditions in the shop. The intercooler pump is laying on the ground under the wheel, hardwired to a battery (with an inline switch). It's drawing from the big steel pot, and running through the intercooler radiators, intercooler, and back into the pot. Just like in the car, so pressure load on the pump is very similar.



Looking over a datasheet for the pump, we saw that it does have the onboard smarts to protect itself if need be. We thought we were getting an overheat condition, so we heated the water up to 180°-200°F and kept it there for 20 minutes. The pump didn't care, and continued happily pumping water the whole time.

We played with starving the pump, and injecting air into the system, and finally figured it out.

Excess air in the system will cause the impeller to cavitate, and enter a three minute shutdown period to recover. That matches our observations on track exactly.

Here's a video of our test. We inject air into the line with a blow gun, and a few seconds later, the pump shuts off. Exactly three minutes later, the pump restarts.



This is HUGE, and gives much more importance to the bleeding process and tank design. Previously, everyone thought that air in the system just hurt the heat transfer and pumping ability, when the reality is much more dire. The system will shut down entirely for three minutes, which is more than enough time for the engine to overheat and go into limp mode.

What also worth noting is that the ECU has nothing to do with this. Throughout our entire test shown in the video, the pump was hooked up to a battery. The ECU has no idea the pump is shutting down until temperatures reach unsafe levels.

Lesson here: Make SURE your intercooler system is bled properly.

Originally Posted by CSIXX1

I understand getting the air out of the system but don't you want to have a little space in the stock intercooler tank to allow for fluid expansion? There is no over flow bottle in this system to collect the expanding coolant like on the engine cooling system.

Originally Posted by Mikec7z

Have you guys tried running 2 water pumps, at 2 different locations on the loop?

The advantage would be, a person could create a relay that basically keeps them both on 98% of the time, and the 2% or whatever time increment is needed, would be used to shut down one pump for 5 seconds... then it comes back on, and then 2 min later, the other pump shuts off for 5 sec, and they alternate back and forth, but 98% of the time, they are both on together.

What this would do is... use the FLOW from the other pump, to greatly accelerate the time needed to get the air away from the pump that is currently shut off. It might prevent cavitation all together as neither pump is having to work as hard.

2 props on the back of a boat will cavitate less at take off than 1 prop on the back of a boat. You are AT LEAST halving the cavitation's chances from occurring, but more realistically, probably even more... and then you are solving the cavitation in just a matter of seconds, instead of waiting for 3 minutes.

If you've already done that, my apologies, I read this thread a long time ago, know the gist of it.. but forgot if that was part of the solution or not.