WTF...Water Pumps!
06-02-2009, 01:51 AM
#166
Former Sponsor
Thread Starter
Mark Anderson called me, today. Interestingly enough, he says that the plastic impeller is indeed still listed as a replacement part and is still listed in the price sheet. I'll check and see if they have any...just for fun.
06-02-2009, 02:01 AM
#167
Addict
Rennlist Member
Rennlist Member
Join Date: Oct 2003
Location: Gone. On the Open Road
Posts: 16,463
Received 1,621 Likes
on
1,059 Posts
No way to know if my theory is correct.
PCNA and 928 Specialists' rebuilt pumps come with a plastic impeller.
Last edited by worf928; 06-02-2009 at 02:06 AM. Reason: No speakada engrish
06-02-2009, 02:02 AM
#168
928 Collector
Rennlist Member
Rennlist Member
I agree.
I'm not even remotely surprised. My pet theory - since all PCNA rebuilds come with a plastic impeller - is that PAG knew that the pump casting couldn't "take" more than 1 or 2 rebuilds and therefore specified the plastic impeller to avoid the consequent block damage of a metal impeller rebuilt pump shredding itself.
No way to know if my theory is correct.
PCNA and 928 Specialists' rebuilds come with a plastic impeller pump.
No way to know if my theory is correct.
PCNA and 928 Specialists' rebuilds come with a plastic impeller pump.
06-02-2009, 03:10 AM
#171
Not the sharpest tool in the shed
Rennlist Member
Rennlist Member
Recent suggestions, in my mind, only increase the complexity of something that should be relatively simple. I'm all for making things better but it sounds like the potential of compounding the propensity for failure. And in regards to putomov's name "double fist", should this new iteration fail would it be the same as a "double fisting"? Sounds painful in any context.
06-02-2009, 03:10 AM
#172
Rennlist Member
Tuomo, I don't want to rain on your parade but it would be very hard to get your third bearing, with its own bracket on the other side of the pulley, to line up with the other two. If it is not aligned within the radial clearance of the other two bearings, then the bearings all get stressed and will fail. This is why blocks get line-bored for main bearings for example.
I think the primary problem is quality, not design. If the design and the bearings were inadequate then every pump would be failing. But neither can they run forever, bearings have a finite life.
This current discussion centers on one failed pump, and unless I missed something Greg B. doesn't yet know how it failed. Absent facts, rampant speculation is a lot of fun, but I think the answer is to find a manufacturer (or rebuilder) who is capable of working to a high standard of quality, and motivated to do so. It's not rocket science, but it does require more care with machining, inspection and assembly than most other WP's.
Just my $0.02.
I think the primary problem is quality, not design. If the design and the bearings were inadequate then every pump would be failing. But neither can they run forever, bearings have a finite life.
This current discussion centers on one failed pump, and unless I missed something Greg B. doesn't yet know how it failed. Absent facts, rampant speculation is a lot of fun, but I think the answer is to find a manufacturer (or rebuilder) who is capable of working to a high standard of quality, and motivated to do so. It's not rocket science, but it does require more care with machining, inspection and assembly than most other WP's.
Just my $0.02.
06-02-2009, 04:15 AM
#173
Rennlist Member
The intake is centered over the impeller. The lower pressure in this region will tend to pull the impeller rearward, acting over a relatively large surface area. Now, the pump is producing pressure, but it is slinging the coolant to the side, not back in the direction of suction. So the higher pressure is mostly pressing inward toward the center from the sides. At the shallow angle of the impeller, it's not trying to push the impeller forward much. Any spillover behind the impeller will likely be at a higher pressure than the front of the impeller, so that's more force tending to push the impeller rearward.
The next thing I want to point out is that the bearing is apparently assembled with the shaft entering the outer race from the rear. The pic I posted above is a bit fuzzy; Leslie posted it here so maybe she or Ed can chime in -- but it looks as if the outer race is composed of two pieces -- the metal appears darker near the ball race. Perhaps this was pressed together, or some other mystical bearing assembly mojo was used. If it was two pieces, this may be a weak point. I'm using weak as a relative term; on the drawing board, two parts pressed together can be VERY strong, but not as strong as a single piece.
So, having mentally poked and prodded this, here is my final speculation (for the moment
)When you rev the motor in a low gear or neutral, especially one of your monsters, all of the pieces accelerate very quickly. That's kind of the point. The water in the block does not move instantly, it has its own inertia. It cannot be compressed, and we hope that there is no cavitation, so it is going to put extra strain on something until it gets moving. I theorize that the rearward axial forces acting on the impeller are greater in a stroker because of the higher rate of acceleration. The first thing that's going to give is the impeller -- it's going to want to walk rearward right off the shaft. It will be interesting to see if, when you pull the pump out, the impeller/shaft relationship is different than when you inspected the parts during assembly. This would explain the accelerated failures in high-power engines.
Add to that, if in fact there is some weak link in the way the bearing is assembled such that we are exceeding the designed rearward axial load, it could be overstressing the seam where it is assembled, causing it to run loose and accelerating wear. A constant in-out motion of the shaft as play in the ball bearing increases would compromise the seal and it's all downhill from there.
A test could be devised where a scrap block with good water jackets(weld a cracked bore if needed) could be set up with a radiator to flow through and a decent-size electric motor rigged to spin the WP, say a 1 or 2hp motor. This would accelerate the WP very fast and with a speed control you could keep the acceleration/max speed below the cavitation point. I bet such a test rig could suck the impeller right off of any pump that didn't have the tightest possible shaft/impeller fit.
Speculating further, if I assume that the above is the problem, I'd say there are two things that could be done to address it.
One, weld the impeller to the shaft. This could be done by putting a suitably sized washer on the end and welding the center to the shaft and the edges to the impeller. The shaft gets tossed at rebuild time anyway, the washer can be ground off the impeller and the impeller re-used since this approach to welding shouldn't damage the bore.
Two, spec a variation on the existing bearing/seal cartridge that is better suited for a rearward axial load, maybe with a thrust bearing or ball bearing designed for high axial loads at the front. This would probably require a pulley redesign, but really the only important dimension for the pulley is the width and OD; other features are irrelevant so long as it is oriented correctly in the belt path. I think many people would be OK with part of the pulley protruding through the timing cover if it solved the problem.
06-02-2009, 12:33 PM
#175
Rennlist Member
Dave,
What you are suggesting may well be part of the failure mechanism, but Greg's motor only had a few hours on it. So either that particular motor is way different from anything done before, or that particular water pump was different (in a bad way). And that pump seized, something not seen before (AFAIK) for an almost-new pump.
The concern I have about quality is that water pumps in most applications are non-critical parts, and a 1 or 2% failure rate is no big deal. So the guys who build (and re-build) water-pumps have that mind-set, and respond to requests for a less-expensive part by using less-expensive castings, spending less time machining and checking, and including fewer (if any) inspection steps.
I don't think we can ever know for sure, but I doubt if the shaft and bores are checked or measured before they get pressed together, and I doubt if the guy running the press watches the pressure-gauge. Yet if you don't do those things then you have no idea how tight that fit is, and for a press-fit the amount of interference is everything.
Part dimensions vary, that's why engineers put tolerances on drawings. Critical parts generally have tighter tolerances and get machined with better tooling in good condition that gets carefully monitored, and the parts get carefully checked. Ordinary parts get machined with ordinary tooling and only spot-checks to save money. And rebuilding introduces an added dimension of unknown wear.
So, without careful machining and checking, an impeller (or pulley) with a slightly-too-large bore can easily get pressed onto a slightly-too-small shaft. And as long as it doesn't fall apart on the way out the door, I don't think anyone is going to catch it.
And yet we are part of the problem. When was the last time that anyone suggested paying more for a water pump? Other than Greg B. that is, in his opening posts here. The problem is that spending $1003 (going up next month) on an OE factory part is still no guarantee that the extra bucks are going into quality rather than corporate overhead.
Heinrich mentioned risk/loss, that goes straight to the point. If my theory is correct-- that the design is fine and the problem is shoddy quality-- then why not offer an insurance deal: We'll pay $100 more for a water pump that comes with a secondary-damages clause: If it fails and damages our engine, then we get $5000 for engine repairs. If we want more coverage, then we can pay more. If the failure rate is really 1% or better then that is a great deal for the supplier. And if they find a way to improve quality then they get to make more money. The beauty is that it puts the customers and the suppliers on the same side of the table.
Cheers, Jim
What you are suggesting may well be part of the failure mechanism, but Greg's motor only had a few hours on it. So either that particular motor is way different from anything done before, or that particular water pump was different (in a bad way). And that pump seized, something not seen before (AFAIK) for an almost-new pump.
The concern I have about quality is that water pumps in most applications are non-critical parts, and a 1 or 2% failure rate is no big deal. So the guys who build (and re-build) water-pumps have that mind-set, and respond to requests for a less-expensive part by using less-expensive castings, spending less time machining and checking, and including fewer (if any) inspection steps.
I don't think we can ever know for sure, but I doubt if the shaft and bores are checked or measured before they get pressed together, and I doubt if the guy running the press watches the pressure-gauge. Yet if you don't do those things then you have no idea how tight that fit is, and for a press-fit the amount of interference is everything.
Part dimensions vary, that's why engineers put tolerances on drawings. Critical parts generally have tighter tolerances and get machined with better tooling in good condition that gets carefully monitored, and the parts get carefully checked. Ordinary parts get machined with ordinary tooling and only spot-checks to save money. And rebuilding introduces an added dimension of unknown wear.
So, without careful machining and checking, an impeller (or pulley) with a slightly-too-large bore can easily get pressed onto a slightly-too-small shaft. And as long as it doesn't fall apart on the way out the door, I don't think anyone is going to catch it.
And yet we are part of the problem. When was the last time that anyone suggested paying more for a water pump? Other than Greg B. that is, in his opening posts here. The problem is that spending $1003 (going up next month) on an OE factory part is still no guarantee that the extra bucks are going into quality rather than corporate overhead.
Heinrich mentioned risk/loss, that goes straight to the point. If my theory is correct-- that the design is fine and the problem is shoddy quality-- then why not offer an insurance deal: We'll pay $100 more for a water pump that comes with a secondary-damages clause: If it fails and damages our engine, then we get $5000 for engine repairs. If we want more coverage, then we can pay more. If the failure rate is really 1% or better then that is a great deal for the supplier. And if they find a way to improve quality then they get to make more money. The beauty is that it puts the customers and the suppliers on the same side of the table.
Cheers, Jim
06-02-2009, 01:45 PM
#176
Rennlist Member
Dave,
What you are suggesting may well be part of the failure mechanism, but Greg's motor only had a few hours on it. So either that particular motor is way different from anything done before, or that particular water pump was different (in a bad way). And that pump seized, something not seen before (AFAIK) for an almost-new pump.
The concern I have about quality is that water pumps in most applications are non-critical parts, and a 1 or 2% failure rate is no big deal. So the guys who build (and re-build) water-pumps have that mind-set, and respond to requests for a less-expensive part by using less-expensive castings, spending less time machining and checking, and including fewer (if any) inspection steps.
I don't think we can ever know for sure, but I doubt if the shaft and bores are checked or measured before they get pressed together, and I doubt if the guy running the press watches the pressure-gauge. Yet if you don't do those things then you have no idea how tight that fit is, and for a press-fit the amount of interference is everything.
Part dimensions vary, that's why engineers put tolerances on drawings. Critical parts generally have tighter tolerances and get machined with better tooling in good condition that gets carefully monitored, and the parts get carefully checked. Ordinary parts get machined with ordinary tooling and only spot-checks to save money. And rebuilding introduces an added dimension of unknown wear.
So, without careful machining and checking, an impeller (or pulley) with a slightly-too-large bore can easily get pressed onto a slightly-too-small shaft. And as long as it doesn't fall apart on the way out the door, I don't think anyone is going to catch it.
And yet we are part of the problem. When was the last time that anyone suggested paying more for a water pump? Other than Greg B. that is, in his opening posts here. The problem is that spending $1003 (going up next month) on an OE factory part is still no guarantee that the extra bucks are going into quality rather than corporate overhead.
Heinrich mentioned risk/loss, that goes straight to the point. If my theory is correct-- that the design is fine and the problem is shoddy quality-- then why not offer an insurance deal: We'll pay $100 more for a water pump that comes with a secondary-damages clause: If it fails and damages our engine, then we get $5000 for engine repairs. If we want more coverage, then we can pay more. If the failure rate is really 1% or better then that is a great deal for the supplier. And if they find a way to improve quality then they get to make more money. The beauty is that it puts the customers and the suppliers on the same side of the table.
Cheers, Jim
What you are suggesting may well be part of the failure mechanism, but Greg's motor only had a few hours on it. So either that particular motor is way different from anything done before, or that particular water pump was different (in a bad way). And that pump seized, something not seen before (AFAIK) for an almost-new pump.
The concern I have about quality is that water pumps in most applications are non-critical parts, and a 1 or 2% failure rate is no big deal. So the guys who build (and re-build) water-pumps have that mind-set, and respond to requests for a less-expensive part by using less-expensive castings, spending less time machining and checking, and including fewer (if any) inspection steps.
I don't think we can ever know for sure, but I doubt if the shaft and bores are checked or measured before they get pressed together, and I doubt if the guy running the press watches the pressure-gauge. Yet if you don't do those things then you have no idea how tight that fit is, and for a press-fit the amount of interference is everything.
Part dimensions vary, that's why engineers put tolerances on drawings. Critical parts generally have tighter tolerances and get machined with better tooling in good condition that gets carefully monitored, and the parts get carefully checked. Ordinary parts get machined with ordinary tooling and only spot-checks to save money. And rebuilding introduces an added dimension of unknown wear.
So, without careful machining and checking, an impeller (or pulley) with a slightly-too-large bore can easily get pressed onto a slightly-too-small shaft. And as long as it doesn't fall apart on the way out the door, I don't think anyone is going to catch it.
And yet we are part of the problem. When was the last time that anyone suggested paying more for a water pump? Other than Greg B. that is, in his opening posts here. The problem is that spending $1003 (going up next month) on an OE factory part is still no guarantee that the extra bucks are going into quality rather than corporate overhead.
Heinrich mentioned risk/loss, that goes straight to the point. If my theory is correct-- that the design is fine and the problem is shoddy quality-- then why not offer an insurance deal: We'll pay $100 more for a water pump that comes with a secondary-damages clause: If it fails and damages our engine, then we get $5000 for engine repairs. If we want more coverage, then we can pay more. If the failure rate is really 1% or better then that is a great deal for the supplier. And if they find a way to improve quality then they get to make more money. The beauty is that it puts the customers and the suppliers on the same side of the table.
Cheers, Jim
I was just speculating as to failure modes, and came up with a theory that I think may explain the higher failure rate in high-HP engines.
06-02-2009, 02:30 PM
#177
Former Sponsor
Thread Starter
Tuomo, I don't want to rain on your parade but it would be very hard to get your third bearing, with its own bracket on the other side of the pulley, to line up with the other two. If it is not aligned within the radial clearance of the other two bearings, then the bearings all get stressed and will fail. This is why blocks get line-bored for main bearings for example.
I think that slightly oversize holes in the mounting plate would allow perfect alignment. Slide on the bearing, captive in the new plate and tighten it down.
I think the primary problem is quality, not design. If the design and the bearings were inadequate then every pump would be failing. But neither can they run forever, bearings have a finite life.
Agreed.
This current discussion centers on one failed pump, and unless I missed something Greg B. doesn't yet know how it failed. Absent facts, rampant speculation is a lot of fun, but I think the answer is to find a manufacturer (or rebuilder) who is capable of working to a high standard of quality, and motivated to do so. It's not rocket science, but it does require more care with machining, inspection and assembly than most other WP's.
Agreed.
Just my $0.02.
I think that slightly oversize holes in the mounting plate would allow perfect alignment. Slide on the bearing, captive in the new plate and tighten it down.
I think the primary problem is quality, not design. If the design and the bearings were inadequate then every pump would be failing. But neither can they run forever, bearings have a finite life.
Agreed.
This current discussion centers on one failed pump, and unless I missed something Greg B. doesn't yet know how it failed. Absent facts, rampant speculation is a lot of fun, but I think the answer is to find a manufacturer (or rebuilder) who is capable of working to a high standard of quality, and motivated to do so. It's not rocket science, but it does require more care with machining, inspection and assembly than most other WP's.
Agreed.
Just my $0.02.
06-02-2009, 02:34 PM
#178
Former Sponsor
Thread Starter
Saw one, exactly like this, two weeks ago. That has to be the result of a poor press fit on the pulley. That is one of the issues with rebuilt pumps. The pulley always gets reused...and never measured. How many times can it be pressed on and off, before it fails?
06-02-2009, 02:46 PM
#180
Former Sponsor
Thread Starter
Couple things... having studied up on this a bit, it seems that the tendency is for the impeller to be pulled inward after all, at least for most pumps of this type, most of the time. Here's why:
The intake is centered over the impeller. The lower pressure in this region will tend to pull the impeller rearward, acting over a relatively large surface area. Now, the pump is producing pressure, but it is slinging the coolant to the side, not back in the direction of suction. So the higher pressure is mostly pressing inward toward the center from the sides. At the shallow angle of the impeller, it's not trying to push the impeller forward much. Any spillover behind the impeller will likely be at a higher pressure than the front of the impeller, so that's more force tending to push the impeller rearward.
That's possible. I was thinking that the attempted compression of the fluid between the impeller and the block would push the impeller forward. Perhaps I've owned too many boats with props and can only think one way.
The next thing I want to point out is that the bearing is apparently assembled with the shaft entering the outer race from the rear. The pic I posted above is a bit fuzzy; Leslie posted it here so maybe she or Ed can chime in -- but it looks as if the outer race is composed of two pieces -- the metal appears darker near the ball race. Perhaps this was pressed together, or some other mystical bearing assembly mojo was used. If it was two pieces, this may be a weak point. I'm using weak as a relative term; on the drawing board, two parts pressed together can be VERY strong, but not as strong as a single piece.
So, having mentally poked and prodded this, here is my final speculation (for the moment )
When you rev the motor in a low gear or neutral, especially one of your monsters, all of the pieces accelerate very quickly. That's kind of the point. The water in the block does not move instantly, it has its own inertia. It cannot be compressed, and we hope that there is no cavitation, so it is going to put extra strain on something until it gets moving. I theorize that the rearward axial forces acting on the impeller are greater in a stroker because of the higher rate of acceleration. The first thing that's going to give is the impeller -- it's going to want to walk rearward right off the shaft. It will be interesting to see if, when you pull the pump out, the impeller/shaft relationship is different than when you inspected the parts during assembly. This would explain the accelerated failures in high-power engines.
I would agree...if the one I have failed from reving the engine quickly. Mine failed after a steady state dyno load, a shut down period, and the restart at idle. I"m thinking it is more "heat soak" related.
Add to that, if in fact there is some weak link in the way the bearing is assembled such that we are exceeding the designed rearward axial load, it could be overstressing the seam where it is assembled, causing it to run loose and accelerating wear. A constant in-out motion of the shaft as play in the ball bearing increases would compromise the seal and it's all downhill from there.
A test could be devised where a scrap block with good water jackets(weld a cracked bore if needed) could be set up with a radiator to flow through and a decent-size electric motor rigged to spin the WP, say a 1 or 2hp motor. This would accelerate the WP very fast and with a speed control you could keep the acceleration/max speed below the cavitation point. I bet such a test rig could suck the impeller right off of any pump that didn't have the tightest possible shaft/impeller fit.
Speculating further, if I assume that the above is the problem, I'd say there are two things that could be done to address it.
One, weld the impeller to the shaft. This could be done by putting a suitably sized washer on the end and welding the center to the shaft and the edges to the impeller. The shaft gets tossed at rebuild time anyway, the washer can be ground off the impeller and the impeller re-used since this approach to welding shouldn't damage the bore.
Again, tough to weld to a hardened bearing shaft. If this was a good thing to do, I'd think all the high performance makers of aftermarket American water pumps would be doing this...they are not.
Two, spec a variation on the existing bearing/seal cartridge that is better suited for a rearward axial load, maybe with a thrust bearing or ball bearing designed for high axial loads at the front. This would probably require a pulley redesign, but really the only important dimension for the pulley is the width and OD; other features are irrelevant so long as it is oriented correctly in the belt path. I think many people would be OK with part of the pulley protruding through the timing cover if it solved the problem.
The intake is centered over the impeller. The lower pressure in this region will tend to pull the impeller rearward, acting over a relatively large surface area. Now, the pump is producing pressure, but it is slinging the coolant to the side, not back in the direction of suction. So the higher pressure is mostly pressing inward toward the center from the sides. At the shallow angle of the impeller, it's not trying to push the impeller forward much. Any spillover behind the impeller will likely be at a higher pressure than the front of the impeller, so that's more force tending to push the impeller rearward.
That's possible. I was thinking that the attempted compression of the fluid between the impeller and the block would push the impeller forward. Perhaps I've owned too many boats with props and can only think one way.
The next thing I want to point out is that the bearing is apparently assembled with the shaft entering the outer race from the rear. The pic I posted above is a bit fuzzy; Leslie posted it here so maybe she or Ed can chime in -- but it looks as if the outer race is composed of two pieces -- the metal appears darker near the ball race. Perhaps this was pressed together, or some other mystical bearing assembly mojo was used. If it was two pieces, this may be a weak point. I'm using weak as a relative term; on the drawing board, two parts pressed together can be VERY strong, but not as strong as a single piece.
So, having mentally poked and prodded this, here is my final speculation (for the moment
)When you rev the motor in a low gear or neutral, especially one of your monsters, all of the pieces accelerate very quickly. That's kind of the point. The water in the block does not move instantly, it has its own inertia. It cannot be compressed, and we hope that there is no cavitation, so it is going to put extra strain on something until it gets moving. I theorize that the rearward axial forces acting on the impeller are greater in a stroker because of the higher rate of acceleration. The first thing that's going to give is the impeller -- it's going to want to walk rearward right off the shaft. It will be interesting to see if, when you pull the pump out, the impeller/shaft relationship is different than when you inspected the parts during assembly. This would explain the accelerated failures in high-power engines.
I would agree...if the one I have failed from reving the engine quickly. Mine failed after a steady state dyno load, a shut down period, and the restart at idle. I"m thinking it is more "heat soak" related.
Add to that, if in fact there is some weak link in the way the bearing is assembled such that we are exceeding the designed rearward axial load, it could be overstressing the seam where it is assembled, causing it to run loose and accelerating wear. A constant in-out motion of the shaft as play in the ball bearing increases would compromise the seal and it's all downhill from there.
A test could be devised where a scrap block with good water jackets(weld a cracked bore if needed) could be set up with a radiator to flow through and a decent-size electric motor rigged to spin the WP, say a 1 or 2hp motor. This would accelerate the WP very fast and with a speed control you could keep the acceleration/max speed below the cavitation point. I bet such a test rig could suck the impeller right off of any pump that didn't have the tightest possible shaft/impeller fit.
Speculating further, if I assume that the above is the problem, I'd say there are two things that could be done to address it.
One, weld the impeller to the shaft. This could be done by putting a suitably sized washer on the end and welding the center to the shaft and the edges to the impeller. The shaft gets tossed at rebuild time anyway, the washer can be ground off the impeller and the impeller re-used since this approach to welding shouldn't damage the bore.
Again, tough to weld to a hardened bearing shaft. If this was a good thing to do, I'd think all the high performance makers of aftermarket American water pumps would be doing this...they are not.
Two, spec a variation on the existing bearing/seal cartridge that is better suited for a rearward axial load, maybe with a thrust bearing or ball bearing designed for high axial loads at the front. This would probably require a pulley redesign, but really the only important dimension for the pulley is the width and OD; other features are irrelevant so long as it is oriented correctly in the belt path. I think many people would be OK with part of the pulley protruding through the timing cover if it solved the problem.