how to check oil pump health

[yelcab1]

Considering that you bought a used oil pump (3.2L), how would you go about checking its health before you install it?

[KTL]

Take it apart and inspect the housing for any scoring.

I know from experience that an oil pump can ingest bearing material, spit it right out and still feel nice and smooth to your "scientific" test of spinning it by hand. Only way to really know is open it up. Be sure to keep the gerotors/impellers marked so you can reassemble exactly as they came apart. Use a paint pen to put a dot on meshing teeth and don't mix up the gerotor stack.

Here's a picture inside of my now wall art 964 pump that gobbled up a bunch of bearing debris. Spun just fine by hand. Not even the smallest hint of dragging or roughness.

Engine Bearing Failure- Disassemble Heads Too?

[Turbo_pro]

Originally Posted by yelcab1

Considering that you bought a used oil pump (3.2L), how would you go about checking its health before you install it?

Send it to Supertec. They can evaluate the condition and modify the flow to a reported 15% increase.

PS: if you have your old pump, they buy core and oil pump parts.
Good luck

[yelcab1]

Originally Posted by KTL

Take it apart and inspect the housing for any scoring.

I know from experience that an oil pump can ingest bearing material, spit it right out and still feel nice and smooth to your "scientific" test of spinning it by hand. Only way to really know is open it up. Be sure to keep the gerotors/impellers marked so you can reassemble exactly as they came apart. Use a paint pen to put a dot on meshing teeth and don't mix up the gerotor stack.

Here's a picture inside of my now wall art 964 pump that gobbled up a bunch of bearing debris. Spun just fine by hand. Not even the smallest hint of dragging or roughness.

Engine Bearing Failure- Disassemble Heads Too?

Kevin,

Even with the scoring that you show on your old pump, would that result in significantly lower oil pressure? Or zero oil pressure? or ... ? I mean the depth of the scoring is not very much and the oil is much thicker than that.

[KTL]

The clearance between the pump housing and the gerotors is very precise/tight. The scoring in the housing opens up that clearance and pressure is reduced.

The pump is quite easy to open. Just have to use a little bit of patience tapping the inner case off of the dowel pins.

For comparison, here's a nice SC pump that I sold before I bought the to-be-doomed 964 pump.




Obviously a clean pump is not going to have any scoring. My point of posting the SC pump picture is to show that the clearance is so tight that a good pump can show a little bit of contact against the housing, as can be seen at the end of the pump housing

[AlfonsoR]

I am not familiar with any repair procedures for automotive gear pumps, but in industrial applications, two key checks are a tooth contact check and back lash. There is some judgment involved in the contact check, but you would like to see 80% or better and the backlash is to be within some specified range.

To do this on a pcar oil pump, it seems like you would have to build a fixture to allow you to locate the rotors and allow you to perform the checks. Basically the housing would have to be open so you could do the checks.

You might also have a reference set of rotors to do a "go/ no-go" check on the housing.

Lastly, you should also do a flow test.

I don't know of anyone that does all this, but that's not to say that no one does. Maybe try googling "oil pump repair" or "gear pump repair"?

[Green993]

Originally Posted by turbo_pro

send it to supertec. They can evaluate the condition and modify the flow to a reported 15% increase.

Ps: If you have your old pump, they buy core and oil pump parts.
Good luck

+1

[Walt Fricke]

Alfonso - I've never heard of anyone making a pump test jig. A simple one wouldn't be hard to make - mill recesses in a pair of plates for the in and out O rings, drill/tap/add fittings, use clamps across the pump to hold plates in place. Maybe make provision for the scavange side also - could just place the intake in your oil container - scavange side most likely to have crap go through it. And use what to turn it - a drill? Not strong enough, maybe need a mill? to turn it. Hoses in buckets, stop watch, etc.

You'd think you could measure output of a specific oil viscosity/temperature at some set RPM without too much trouble. And compare with a new pump, or a bunch of pumps which seemed to be doing the trick in their motors when removed.

How does pressure enter in, though? If a pump will put out 100 liters per minute into an open container, can that be translated into 100 times 0.X liters per minute into a restriction set at 80 psi? (thinking of this like Ohms law, but for hydraulics not electricity). Or might a worn pump perform worse against a resistance than a new pump of the same base output?

I think you might be able to measure backlash on an assembled pump through the ports in its side.

[AlfonsoR]

Walt, in the industry I'm in, jigs are not too uncommon. For example, we use mandrels for checking tilt-pad bearing clearances, coupling tapers, etc. Yes, I would think a jig for testing oil pump rotors should be easy enough to make.

Well, just thinking out loud, but it seems that a flow test rig would not be too difficult to design either. The pump driver would need to be one where speed (rpm) is indicated and maintained fairly accurately.

On the scavenge side, the flow resistance for the filter would have to be accounted for, measured on a live engine would be best.

You would make the rig to mimic a typical engine set-up, possibly including oil tank and filter, then return the oil to pressure side.

On the pressure side, yes, a calculation for the sum resistance of the various passages would have to be developed to come up with a single flow orifice. It might again, be easier to figure out a way to get measure pressure data on a live engine. Calculations are great, but nothing beats a measurement.

Once you build your rig, which would include a temperature control system to maintain normal running temp (warmer for simulating track use possibly), then you would need to benchmark at least three, maybe 5, new pumps. I know there is data out there that states certain pumps should flow so much. If you were doing this for a living, you would want to validate that as best possible, in my opinion. You would want to build a flow vs pressure vs RPM curve. After analyzing the data, you would build a master test curve.

Use the master to compare to flow data for customer's used pump. Supply customer with a dyno sheet, so to speak, of their pump compared to master curve. You might possibly make a judgment on whether the pump is acceptable or not, etc.

Well, OK, that sounds like there might be a possible business opportunity there. But maybe Supertec already beat us to the punch. That Henry is a crafty guy. I believe that's why some call him Yoda.

[Walt Fricke]

I think that, when it comes to fixing pumps, about all you can do is shorten the housing a smidge to make up for internal end wear, which you could clean up with a mill. And you could add bushings for shafts if their holes (unbushed on earlier pumps) have gotten enlarged. But if the gears are dinged, or the insides of the housings scored, not much you can do, I think. At least not cost effectively. Which would make being able to test a pump all the more useful - the scores and dings might, or might not, reduce flow/pressure enough so that you'd not want to use the pump again.

Henry has some magic to improve pump flow - I imagine it is akin to porting heads, but that's not the same as repairing a pump which has been damaged and isn't able to do what is needed as a result.

I would think that by testing some new pumps, and some used but apparently good pumps, against about whatever reasonable approximations of what a pump has to do, and thus coming up with a baseline, you could then compare about any used pump against that data. You'd think a really badly worn pump would stand right out doing that. No real need to simulate flow resistance as long as you had some and kept it consistent.

[KTL]

Just FYI on pump flow improvement. Aaron (BURN-BROS) mentioned to me that the later pumps like 964, 993-up are already flow improved. So Henry's improvements are apparently on the 3.2 and older pumps?

Point being is that someone could visually compare the 964-up pump to the older pumps and see how things differ inside. Going by memory, the gerotors are basically the same. I suspect it's improving the port contours? There's not much to an oil pump. Precise animals they may be, but they have very few parts within them to modify!

[Walt Fricke]

I will soon, alas, have my GT3 pump out on the bench again. If I remember, I'll compare what I can see of the in and out ports to the couple of really early 911 pumps (one I think has a sand cast case) I have and should try to sell.

[AlfonsoR]

Here is a video by Melling on comparison of pumps. Granted they are for a SBC, but i would infer that Porsches oil pumps all have the same gear shape and that the only difference is the length of the gears. Melling Oil Pumps - YouTube


In Bruce Andersons book, he talks about lengthening of pumps as the pumps went through various upgrades in capacity (my understanding). He also mentions for example that the 964 oil pump on the scavange side was 1.83x greater in capacity than the pressure side. Which I take to mean that the length of the scavange side is 1.83x that of the pressure side.

I don't know whether it is worth it to repair oil pumps, butI think a lot of people tend to throw away perfectly good pumps becase they see some FOD on the gears or at least that's my impression, that people are throwing them away. When if you had a decent method of flow testing, i think it would take a lot of the unknown out. Maybe because the housing has a couple of scores in it, it will now only pump 60 liters per minute instead of 65. Would 60 be too low? I don't know, but having the data, lets you make an informed decision.

[Walt Fricke]

I'm with Alfonso - nothing like having good information to make good decisions.