Can oil pump be bench tested?

Status
Not open for further replies.
To close the loop on my question, here is the follow up.

My issue was that I had removed the oil pump from my 1971 Roadster to see if I could reduce wet sumping. I have 3 other spare pumps, another from a 71 engine and 2 from 73 engines. I wanted to find a way to verify functionality and to compare the performance of these pumps. Then I would decide which one to put in the engine I'm working on, either with or without refurbishing.

I built 2 test rigs, neither of which is scientific in nature. These tests are not designed to measure any real world performance parameters, only go/no go function and a relative performance comparison between pumps.

The first I'll call the static test rig and my intention was to see which pump would be most resistant to allowing oil from the tank to flow to the sump. It consists of a funnel reservoir which gravity feeds a pump thru a spare timing side crankcase which has the pump under test mounted on it. My intent was to measure the amount of time it took for the 24 oz of oil in the reservoir to be emptied. I had pans to collect any oil that dripped out of the pump's feed output and out of the channel to the sump on the return side. The length of time it took to empty the reservoir lasted so long that it would empty in the middle of the night, so instead I switched to measuring the drip period, which was quite regular.
Results:
-My original pump had the longest initial drip period of around 12 seconds. The others were clustered around 7 seconds.
-No oil made it thru to the return path collection point. All the flow was a result of oil passing thru the output of the feed side of the pump, which when assembled, I presume would make it to the sump thru the crankshaft etc.

Here is a photo of the static test set up:
Can oil pump be bench tested?
 
The 2nd test looks at the dynamic function of the pumps. The rig is based on the French reference in post #12 of this thread:


Pompe à huile petit banc de test​



Instead of the "hockey puck" , I machined a rectangular block of aluminum as the fixture that simulates the timing side crankcase, with mounting for the pump and junction block assembly. I used a spare oil tank as the reservoir and connected to the simulated crankcase as on the bike. The pump output goes to a flow restrictor the output of which empties into a container that would effectively be the engine sump. A pressure gauge measures the oil pressure in the line from the pump output to the flow restrictor. The return side of the pump is connected to the container which supplies the pump with the oil that it should return to the tank.
As in the French reference, the pump is driven by a variable speed drill which connects to the nut that holds the worm gear. I used a hose clamp to hold the trigger in various positions, adjusted by turning the screw on the clamp. A hand held optical tachometer was used to measure pump rpm.

Can oil pump be bench tested?Can oil pump be bench tested?

This rig can test both feed and return function of a pump.
For the feed side, I would take measurements of output oil pressure at various pump rpm from around 300rpm to around 2500rpm. I believe that engine rpm is roughly 2 1/3 times the pump's rpm, so that gives a decent range of normal engine operating speeds.
For the return side, I connect the hose from the feed side that went to the sump container to the breather input to the tank. With this connection, the feed side just circulates from & to the tank. Then, I measure the amount of time it takes the return side to remove a specific volume of oil from the container. From this, I get a flow rate for the return side, at a given pump rpm.

I'll post the results of the testing tomorrow.
 
I ran the tests on all 4 pumps.
Test Conditions:Castrol SAE 20W-50 used engine oil (300 miles), 70* F; 0.100" Flow Restrictor.
For the feed side testing, all 4 pumps exhibited a pretty much linear relationship between pump rpm and pressure over the range of rpm used. The figure of merit I used for comparison of the feed side performance of the pumps was the pressure at 5000rpm equivalent engine speed. Again my original pump performed best, at just over 70psi. The next one was about 68psi and the other 2 between 63 & 64psi.

For the return side testing, I ran the pumps at the equivalent of 2600rpm engine speed. The pump from the '73 750 engine performed best with flow rate measured at 0.536 liters/minute. My original pump came next at 0.484 liters/minute. Next was the other '71 pump at 0.366 liters/minute. The pump from the '73 850 engine had no return function.. It didn't move any oil out of the container at any pump rpm.

So what do I conclude from all this?
I feel comfortable just putting the original pump back in the engine and leave it at that. It has the overall best comparative performance. There is only a very small amount of end play (0.0015") on its shaft, so I'm not convinced there would be any benefit to disassembling it and lapping the surfaces.

Sounds like all this test rig building & testing was a waste of time... but no. I'm really glad I did this testing instead of just replacing the original pump with the now known to be non-functional spare (which came from the 850 parts) as I had seriously considered doing (see my original post).
 
If anyone wants a dimensional drawing of the machined fixture, the detail test results or my suggestions on how to improve the test setup, send me a PM.
 
NickZ said:
If anyone wants a dimensional drawing of the machined fixture, the detail test results or my suggestions on how to improve the test setup, send me a PM.
Click to expand...
If I understand it properly, you said one pump, "The pump from the '73 850 engine had no return function.. It didn't move any oil out of the container at any pump rpm." If I did not move oil, I would be curious to see what it looks like inside if you take it apart and hear why, What was going on in it to make this happen
 
Carl H said:
If I understand it properly, you said one pump, "The pump from the '73 850 engine had no return function.. It didn't move any oil out of the container at any pump rpm." If I did not move oil, I would be curious to see what it looks like inside if you take it apart and hear why, What was going on in it to make this happen
Click to expand...
Carl the norton pump can wear to the point that 1. gear end clearance is big, 2. the tips of the gear to the body clearance is big, and the gear to gear profile is a poor mesh with a big clearance. With a drain down of the oil back to the sump, it is apparent there is not enough vacuum to lift the oil from the sump up to the scavenge set of gears.
Real cases make the pump vacuum lift the oil a little over 2" before it pumps liquid.
I should note that my test rig uses a real set of norton cases and not an artificial test fixture that makes test results suspect.
One pump from a NENO friend in Connecticut was tested and found to be poor and would not scavenge adequately . As the oil got hot and viscosity decreased, it would get worse.
I do applaud Nick's efforts.

Is the pump missing the slave gear on the scavenge side? LOL
 
NickZ said:
To close the loop on my question, here is the follow up.

My issue was that I had removed the oil pump from my 1971 Roadster to see if I could reduce wet sumping. I have 3 other spare pumps, another from a 71 engine and 2 from 73 engines. I wanted to find a way to verify functionality and to compare the performance of these pumps. Then I would decide which one to put in the engine I'm working on, either with or without refurbishing.

I built 2 test rigs, neither of which is scientific in nature. These tests are not designed to measure any real world performance parameters, only go/no go function and a relative performance comparison between pumps.

The first I'll call the static test rig and my intention was to
Results:
-My original pump had the longest initial drip period of around 12 seconds. The others were clustered around 7 seconds.
-No oil made it thru to tcrankshaft etc.

Here is a photo of
Click to expand...
dynodave said:
Is the pump missing the slave gear on the scavenge side? LOL
Click to expand...
This was my thought too.
 
Whist I applaud all your effort, there is , to me some aspects you appear to have missed with this set up, namely hot engine oil.
This will drip, when stationary, pass the gears far more readily than cold oil. Also, don't forget that the oil pump and crankcase will also be hot, finding gaps that don't appear when cold.
But intresting all the same.
 
Bernhard said:
you appear to have missed
Click to expand...
I assume you mean Nick.
My dynamic rig uses ATF because I knew of the hot/cold viscosity difference. In addition, I run a thermocouple readout in my oil tank reservoir. Even the pumping action heats the oil...
My static rig is strictly to measure resistance to static pump ATF flow cooresponding to wet sumping.
 
Carl H said:
If I understand it properly, you said one pump, "The pump from the '73 850 engine had no return function.. It didn't move any oil out of the container at any pump rpm." If I did not move oil, I would be curious to see what it looks like inside if you take it apart and hear why, What was going on in it to make this happen
Click to expand...
I plan to take that pump apart to see what is going on there. Probably get to that in a week or so.
 
dynodave said:
I assume you mean Nick.
My dynamic rig uses ATF because I knew of the hot/cold viscosity difference. In addition, I run a thermocouple readout in my oil tank reservoir. Even the pumping action heats the oil...
My static rig is strictly to measure resistance to static pump ATF flow cooresponding to wet sumping.
Click to expand...
Good point. I could repeat the tests with ATF but I don't expect the comparison between pumps would change.
I also noticed that the oil did heat up a bit while running the tests, but obviously nowhere near normal operating temps.
 
NickZ said:
I plan to take that pump apart to see what is going on there. Probably get to that in a week or so.
Click to expand...
I finally had time to open up the oil pump I tested that had no sump return function. The reason was that the path inside the pump body that connects the return side gear chamber to the output that leads back to the oil tank was clogged with something that resembled dirt. You can see some of it in the center hole here:
Can oil pump be bench tested?

The rest of the pump doesn't look too bad.
Can oil pump be bench tested?
 
Thanks, It explains the unexplainable, The little pumps do the job and last a few life times, I wish I was so lucky!!
 
NickZ said:
I finally had time to open up the oil pump I tested that had no sump return function. The reason was that the path inside the pump body that connects the return side gear chamber to the output that leads back to the oil tank was clogged with something that resembled dirt. You can see some of it in the center hole here:
View attachment 79130

The rest of the pump doesn't look too
Click to expand...
 
I'm beginning to suspect that the bike that pump came from might have been a flood victim. The return tube inside the oil tank from this bike (which I used as the reservoir for my test rig) also was clogged with that substance, along with some of the galleys in the case.
 
NickZ said:
I'm beginning to suspect that the bike that pump came from might have been a flood victim. The return tube inside the oil tank from this bike (which I used as the reservoir for my test rig) also was clogged with that substance, along with some of the galleys in the case.
Click to expand...
Down in the southern USA , they have insects that make nests with mud in holes, like gas tank vent hoses. I had a problem with some insect that got into some NOS carb bodies and left crap in the idle fuel circuit. And I live up north in Massachusetts.
 
Don't know the bike's history but I did buy its carcass in the south.
 
Status
Not open for further replies.
Back
Top