Water pump cavitation solution

General engine tech -- Drag Racing to Circle Track

Moderator: Team

User avatar
Wolf_Tm250
Pro
Pro
Posts: 315
Joined: Sat May 13, 2006 7:36 pm
Location: Parma - Italy
Contact:

Water pump cavitation solution

Post by Wolf_Tm250 »

Hi,
have to rebuild an old ST165 3SGTE Toyota engine and make it a track engine.
I heard someone got water pump cavitation over 6000 rpms even with uprated rad cap, so here is the question:
if I'd make some holes on the outer ring (see pic) of the thermostat valve to decrease a bit pump's suck work, could this decrease cavitation predisposition?
Image
Thanks.
Wolf_Tm
TM enduro 250cc 2stroke
Toyota Celica Gt-Four ST205 Snowy White
http://www.youtube.com/WolfTm250
EFI University Advanced tuner
JoePorting
Guru
Guru
Posts: 2997
Joined: Mon Jun 14, 2010 3:16 pm
Location: Lake Elizabeth, CA

Re: Water pump cavitation solution

Post by JoePorting »

Drilling holes in the thermostat isn't going to do much. It'll just make it longer for the motor to warm up and only add alittle bit more flow after the motor is warm. Does it overheat when the rpm's are up??? If not, I wouldn't worry about it. If it does, I'd try and get a better water pump or bigger radiator, or maybe experiment with different sized pulleys to either speed up the water pump or slow it down. It might actually be that the water pump isn't spinning fast enough, although more likely that it's spinning too much.
Joe Facciano
User avatar
dfree383
Guru
Guru
Posts: 1990
Joined: Fri Feb 01, 2008 7:01 pm
Location: The Sand Box

Re: Water pump cavitation solution

Post by dfree383 »

Electric water Pump?
pheyden
Pro
Pro
Posts: 347
Joined: Thu Aug 25, 2011 10:59 am
Location:

Re: Water pump cavitation solution

Post by pheyden »

The problem with water pumps is something that I see every day with the motors that I work on. These things rev to 8500 RPM and many times the pulley on the water pump is 20% smaller that the one on the crankshaft. So at 8000 Engine RPM the poor pump is turning 9600.

When you turn a centrifugal pump that fast you can in fact lower the "pressure" at the inlet to the pump to ZERO. Yes, you can run an 18 lb. cap, and a thermostat, but if the pump is turning so fast that it creates a low pressure zone at the inlet to the pump, then water will instantly boil and turn to steam. With coolant it will be less severe, but it can still happen.

Now centrifugal pump do not move steam at all well. In fact they do not move it at all. So all water movement in the system ceases. This starts a chain reaction. The area where there is stead will not conduct heat. The causes the area adjacent to the steam pocket to overheat, etc etc etc.

I would always run a pulley on the waterpump that is at least 20% and perhaps as much as 40% smaller than the crankshaft pulley. In this way you will eliminate the possibility of cavitation.

Now you might say well the engine will now overheat because the water is moving more slowly. This is old wives tale. The cooling system, including the radiator, has a temperature delta within which it works. So if you move coolant through the systems at a flow rate of 5 GPH, instead of 10 GPH, what will happen. If we assume that the water is moving half as fast as before, it will spend twice the time in the motor absorbing heat, and twice the time in the radiator transferring heat. Whereas before at higher flow rates it simply made TWO round trips instead of one. The total time, both absorbing heat and dissapating heat, is the same in both cases. The engine will reach an equilibrium which is the same in both cases.

If you want to substantively change the "effectiveness" of the cooling system, then you have to provide a greater opportunity for heat dissapation. This would require either a larger radiator or more effective channeling of air through the radiator element. Remember that radiators shed heat two ways, namely radiation and convection. Radiation is limited in terms of its effectiveness, and in automotive applications this means using convection. This means air has to pass THROUGH the radiator core. Air will take the course of least resistance and this means the shrouding and ducting are of paramount importance. If the shrouding is not correct, you could end up with higher pressure behind the radiator than in front of it. There must be a pressure differential between the front and back of the radiator core. You can easily test for this with a differential pressure gauge with two pitot tubes. One is place in from the the radiator at 90 degrees to the air flow, and the other is placed behind the radiator at 90 degrees to the air flow. The differential will display on the guage.

Sorry to get carried away with this, but this is one of the areas that is sometimes misunderstood. The water pump is blamed for all manner of ills, but this is not always correct.

Paul
Scuderiatopolino.com
User avatar
BrazilianZ28Camaro
Guru
Guru
Posts: 3939
Joined: Mon Feb 09, 2009 10:52 pm
Location:

Re: Water pump cavitation solution

Post by BrazilianZ28Camaro »

pheyden wrote:
I would always run a pulley on the waterpump that is at least 20% and perhaps as much as 40% smaller than the crankshaft pulley. In this way you will eliminate the possibility of cavitation.

Paul
Scuderiatopolino.com

You mean the water pump pulley should be bigger than the crank pulley to underdrive the pump right?


I like that too, free up some horsepower too! :wink:
'71 Z28 street strip car
Pump gas All motor SBC 427
3308 lbs-29x10.5 Hoosiers
NEW BEST ET
1.38 60' / 4.05 330' / 6.32@111.25mph

https://www.youtube.com/watch?v=99p13UK ... ture=share
User avatar
panic
Guru
Guru
Posts: 2295
Joined: Sat Dec 11, 2004 12:04 pm
Location: Ecbatana
Contact:

Re: Water pump cavitation solution

Post by panic »

Not to argue against your conclusion, but "Whereas before at higher flow rates it simply made TWO round trips instead of one. The total time, both absorbing heat and dissapating heat, is the same in both cases." suggests that the product of flow rate × time in seconds is a constant - which isn't true.
The time is obviously a fixed value, but the heat transfer varies significantly with the Delta T of the 2 surfaces (water and block or core), with the higher differential transferring more.
Unless the radiator is as hot as the block's water jacket (presuming the core is at thermostat temperature, and the block is hotter at least at the exhaust valve), the product of both values won't match except by coincidence.
Since I agree with your conclusion (it doesn't matter) we assume the comparison to be valid in a practical sense for the parameters involved.
pheyden
Pro
Pro
Posts: 347
Joined: Thu Aug 25, 2011 10:59 am
Location:

Re: Water pump cavitation solution

Post by pheyden »

Hello Brazilian- Yes, of course I stand corrected. The crankshaft pulley should be smaller than the waterpump pulley in order to underdrive the waterpump.

Paul
Scuderiatopolino.com
mkodama
New Member
New Member
Posts: 17
Joined: Mon Nov 10, 2008 12:55 am
Location: Berkeley, CA

Re: Water pump cavitation solution

Post by mkodama »

The things I've read to reduce cavitation include:
-slowing down water pump impeller speed
-increasing system pressure
-decreasing system temperature
-I've read impeller design and coolant path design can make a significant difference, but usually those items are out of control for the engine builder.
-I guess reducing coolant flow resistance like what you mentioned would help, but I don't really see that making a huge difference since it's not even known whether that is the point of greatest resistance.

Also, I second the comments about coolant flow not being proportional to cooling system heat dissipation. I think it was Davies Craig Electric Water Pumps that has a video with some tests showing how insignificant coolant flow was in relation heat dissipation, but it still made some difference.
-Miles
User avatar
Wolf_Tm250
Pro
Pro
Posts: 315
Joined: Sat May 13, 2006 7:36 pm
Location: Parma - Italy
Contact:

Re: Water pump cavitation solution

Post by Wolf_Tm250 »

dfree383 wrote:Electric water Pump?

I already thought about them, but I've been told they are not so reliable in longer (24h) track events when used alone with no mech pump...
Wolf_Tm
TM enduro 250cc 2stroke
Toyota Celica Gt-Four ST205 Snowy White
http://www.youtube.com/WolfTm250
EFI University Advanced tuner
User avatar
Wolf_Tm250
Pro
Pro
Posts: 315
Joined: Sat May 13, 2006 7:36 pm
Location: Parma - Italy
Contact:

Re: Water pump cavitation solution

Post by Wolf_Tm250 »

pheyden wrote:The problem with water pumps is something that I see every day with the motors that I work on. These things rev to 8500 RPM and many times the pulley on the water pump is 20% smaller that the one on the crankshaft. So at 8000 Engine RPM the poor pump is turning 9600.
IIRC 3SGTE's pulleys are 1:1



I would always run a pulley on the waterpump that is at least 20% and perhaps as much as 40% smaller than the crankshaft pulley. In this way you will eliminate the possibility of cavitation.
Yes, sure... but unfortunately I don't think there are available larger water pump pulleys...

Now you might say well the engine will now overheat because the water is moving more slowly. This is old wives tale.
I read this too everywhere... but fortunately I already pretty understood that the slower flow will almost have an impact just on the Delta T between block's inlet and outlet, and radiator's inlet and outlet... :wink:


If you want to substantively change the "effectiveness" of the cooling system, then you have to provide a greater opportunity for heat dissapation. This would require either a larger radiator or more effective channeling of air through the radiator element. Remember that radiators shed heat two ways, namely radiation and convection. Radiation is limited in terms of its effectiveness, and in automotive applications this means using convection. This means air has to pass THROUGH the radiator core. Air will take the course of least resistance and this means the shrouding and ducting are of paramount importance. If the shrouding is not correct, you could end up with higher pressure behind the radiator than in front of it. There must be a pressure differential between the front and back of the radiator core. You can easily test for this with a differential pressure gauge with two pitot tubes. One is place in from the the radiator at 90 degrees to the air flow, and the other is placed behind the radiator at 90 degrees to the air flow. The differential will display on the guage.

Sorry to get carried away with this, but this is one of the areas that is sometimes misunderstood. The water pump is blamed for all manner of ills, but this is not always correct.

Paul
Scuderiatopolino.com

Thanks Paul.

Wolf
Wolf_Tm
TM enduro 250cc 2stroke
Toyota Celica Gt-Four ST205 Snowy White
http://www.youtube.com/WolfTm250
EFI University Advanced tuner
User avatar
Wolf_Tm250
Pro
Pro
Posts: 315
Joined: Sat May 13, 2006 7:36 pm
Location: Parma - Italy
Contact:

Re: Water pump cavitation solution

Post by Wolf_Tm250 »

mkodama wrote:The things I've read to reduce cavitation include:
-slowing down water pump impeller speed
Not possible, no underdrive pulley available for water pump
-increasing system pressure
Already done with rad cap
-decreasing system temperature
Already partially done with lower opening thermostat
-I've read impeller design and coolant path design can make a significant difference, but usually those items are out of control for the engine builder.
Yes, not possible
-I guess reducing coolant flow resistance like what you mentioned would help, but I don't really see that making a huge difference since it's not even known whether that is the point of greatest resistance.
Yes, you're probably right

Thanks
Wolf_Tm
TM enduro 250cc 2stroke
Toyota Celica Gt-Four ST205 Snowy White
http://www.youtube.com/WolfTm250
EFI University Advanced tuner
RW TECH
Guru
Guru
Posts: 2398
Joined: Thu Nov 24, 2005 9:08 pm
Location: DETROIT, MI

Re: Water pump cavitation solution

Post by RW TECH »

The best thing you can do is ensure a LOT of water is available at the inlet side of the water pump. Sometimes this means auxiliary plumbing to the rear area of the pump housing from radiator or water inlet pipe.....Whatever you can do to help this is going to pay you back.

Pump inlets can easily run at negative pressure in the closed cooling system and you don't have to be turning a gazillion RPM for that to occur. Drown the pump inlet before you change much of anything else.
pheyden
Pro
Pro
Posts: 347
Joined: Thu Aug 25, 2011 10:59 am
Location:

Re: Water pump cavitation solution

Post by pheyden »

Here is an idea that someone may wish to take further. On by small motor (1000cc 13.5:1 compression) I decided that I wanted to reverse the flow of the water. That is take the water from the low Delta T from the radiator and introduce it into the head FIRST and then channel it down through the head gasket to the water jackets around the cylinders. This would remove heat from the area around the exhaust valves more efficiently and also heat up the block more evenly.

Removing heat from the heads would assist in staving off detonation, allowing you to run slightly higher compression. Having the block heat up more evenly and consistently (as opposed to having the cooler water enter the front of the block and then migrate to the rear of the block, as would be with the standard water pump location) would aid in better cylinder roundness and ring sealing.

My main problem was that keeping the water pump on its block mounted location, and finding a way to rechannel to flow via an external line, introduced more problems (complexity, space etc). So I decided to use a Davies Craig electric centrifugal water pump. Now it could have been mounted anywhere, but I came up with the following idea.
Header tank complete.jpg
Waterpump Mount.jpg
This is the expansion tank (header tank) that I fabricated. As you can see the Davies Craig pump is attached to a flange on the very bottom of the tank. Water level in the tank is maintained just above the 90 degree -12 connector, about halfway up the column. The area above the water level is there for system expansion when hot and under increased pressure. By placing the pump in this situation it always has a head of water sitting against the impeller.

BTW - The two smaller AN fittings, further up on the tank are for "vent lines" coming from the head and from the front mounted radiator (this is a rear engine car). The head vent line is placed as high as possible in the head and has a very small pill orifice in the line. A small amount of fluid is always passing through this line, and should a steam bubble occur, then this would be flushed though the line to the upper portion of the expansion tank. As it passes through the pill orifice the steam would condense on the walls of the expansion tank (they being less that the temperature of steam) and become liquid again. The vent line coming from the front mounted radiator is more of a convenience. It comes from the highest point of the radiator and again passes a small amount of liquid to the same upper portion of the expansion tank. This makes the radiator "self-purging" with respect to any trapped air that may be present.

The idea of reverse flow water flow is not new. Pontiac used it in the 50s and the latest Corvette uses a system that incorporates many of the same principles. However in implementing it I was confronted with how to make the water pump work most efficiently.

Hope this gives you some ideas.

Paul
Scuderiatopolino.com
You do not have the required permissions to view the files attached to this post.
tonyg64
Member
Member
Posts: 178
Joined: Fri Sep 04, 2009 4:19 pm
Location: uk

Re: Water pump cavitation solution

Post by tonyg64 »

Ive seen water/head gasket problems on 3SGTE's plenty of times. When running lots of boost they can lift the head slightly and induce air into the system.
Simple fix ? Run EVANS COOLANT. Not cheap but worth it imo. Nucleate boiling is common around no3 cylinder and every detted piston ive seen on these engines is No3. The Evans stuff wont nucleate boil around the water jacket surfaces and the prob will dissapear :)
Just to add, the non turbo version of the same engine which uses the same block and water pump set up does Not boil even when thrashed on track for 40 minutes, so in my mind its the boost that adds the issue

Tony
User avatar
MadBill
Guru
Guru
Posts: 15024
Joined: Tue Nov 15, 2005 10:41 am
Location: The Great White North

Re: Water pump cavitation solution

Post by MadBill »

When you say "Already done." to the question about raising system pressure Wolf, raised to what? We run a 29 psi cap on our road racer (as do many circle track and RR cars), which increases the B.P. of straight water to 275° F. It's going to take a ton of suction and temperature to cavitate at that pressure...

Also, sorry, but when you say: "Not possible, no underdrive pulley available for water pump" It does not reflect the correct racer attitude. How not possible? Cut the center out of a stock one and weld it into a larger pulley, carve one out of billet, substitute a smaller crank pulley instead...Faced with this issue, 40 years back when such were not available, I made my own crank pulley from an aluminum billet and am still using it...
Felix, qui potuit rerum cognscere causas.

Happy is he who can discover the cause of things.
Post Reply