I've been o-ringing the nose of all my oil pumps for years, a single ring between the ports and the inner end of the body. That's the most important one, for preventing air being sucked in from inside the crankcase. Since my wbx's all use hydraulic lifters non-aerated oil is especially important.
So since it's one of those topics that's like the weather, that is everyone talks about it, but no one does anything about it, here's a little treatise for you all on how I o-ring oil pumps.
The easiest way to do that most important one is on a lathe. But you quickly find how you can't mount the awkwardly-shaped pump body in a normal 3-jaw chuck so you can machine the nose. You can center it up in a 4-jaw for a one-off, sure, but I'm doing several at a time, often, and that gets pretty old pretty quick.
So, obviously, make a fixture. Take a junk pump body, punch out the steel idler gear pin, and saw or turn off the protruding bosses on the back of the round part of the body. In a 4-jaw chuck you can then turn off the back of the cut off body so it's dead flat. Now you can chuck the round part of that body quickly in a 3-jaw, and using either two M8 nut-bolt pairs at opposite corners, or a longer M10 bolt and nut thru the driven gear hole, mount the workpiece pump face-to-face with your fixture pump, and zero it in with a dial indicator (within .002" radially is good enough for the o-ring). That's pretty fast and easy to do, even over and over. The groove is then easily cut with a parting tool.
The o-ring sizes that work are a #37 or a #145. Buna-N is what I use, 70 durometer. The only material I would consider superior is fluorosilicone but it's much more expensive and the advantages it offers are marginal for the cost. Buna-N works well with about .008-.010" projection above the groove.
The #37 o-ring is nominally 1/16" (.0625") in section, so for ~.010" projection it needs a groove ~.052" deep and .070-.080" wide. A common 5/64" parting tool is .078" wide so that can be used in a single pass.
The #145 is nominally 3/32 (.094") in section so for ~.010" projection it needs a groove ~.084" deep and ~.105-110" wide. That can be done with that same 5/64" tool, moving it aside the extra .025" or so, and doing a nice sweep across the bottom of the groove for the last couple thou of depth.
Depending on the pump, the distance from the larger suction-side port to the nose of the body varies, so you may favor the thinner #37 o-ring if it is a large-suction-hole pump. The ones I use have the larger hole and I have been using the #145 o-ring in spite of that, but will be switching to the #37 for reasons I will explain below.
The amount of o-ring projection can be as little as ~.005-.006" and do the trick. How much you want to have depends a bit on the material selection, and whether you would like to be able to install the pump in an assembled case. No matter what, the outcome is always more certain if you install the pump when mating the case halves (using bolts rather than studs), but I have pulled and reinstalled my pumps with the #145 o-ring projecting ~.008" with no damage to the rubber, and that is without taking the precaution of chamfering the port holes in the case. Grooving deeper for a smaller projection of course will improve the odds, as will chamfering the ports. O-rings should always be oiled as well as the surfaces they will ride against, which must be smooth and clean. Don't gob everything up with sealant, that obviates the whole reason you're fitting o-rings, which is to have a flexible seal that responds to dimensional changes under expansion and contraction in a way few sealants can.
Now, as long as I have had the pumps centered up on the lathe, it takes only a minute more to cut a second groove up under the flange, as a belt-and-suspenders approach. You still want a gasket/and/or sealant under the flange, and your bolt threads need to be treated with a thread sealant, so oil doesn't seep out the threads from inside the crankcase (thats the oil the plastic sealing inserts in the OEM-spec pump cover nuts are there to contain, but better to stop that seepage at the source by sealing the threads). Don't cut the groove butt up against the flange, move it away a bit more than the thickness of the gasket so the gasket will center on the pump body like it's supposed to and not interfere with the o-ring.
This pump has both body grooves and they accept the #145 o-ring:
OK, that's how I used to do the pumps. Nowadays I'm going yet a step further and placing an o-ring in the pump face to give a flexible seal under the cover as well. That's 3 o-rings per pump! Now I know someone much smarter than I am is going to say this is too much trouble and not needed, and y'know it probably is too much trouble for a hobbyist, but I'll tell you what: when you are building engines constantly, and once you work out your processes, the time spent on machining is made up two-fold in time saved during assembly, and I have the assurance of a leak-free pump for the life of the engine. My new process lets me do all the operations on the mill.
Below are pics of the fixture I use on my mill's rotary table. The first pic shows how I faced off the backside of a pump body, leaving a boss that fits very snugly into the center hole of the rotary table. I then milled out slots on either side that allow a 3/8" x 1" flat steel bar to slide thru.
The bar has holes on either end so it can be bolted down to the t-slots in the rotary table top, holding the fixture down with the boss keeping it on-center. The bar has an M10 threaded hole that ends up aligned with the driven gear shaft hole in the workpiece pump. A little relief had to be milled out of the bar to make room for the driven gear shaft boss on the back of the workpiece pump.
Most critically, the face of the fixture was milled out on the rotary table until the round part of a new pump body would nestle snugly into the round depression. This way, a pump rests perfectly flat and on-center every time it is dropped onto this fixture.
It gets bolted down to the fixture thru the driven gear shaft hole, with a wide flat washer and nylon washer to hold it down snug without marring the inside of the workpiece.
On the pump face, a 1/8" endmill cuts a nice square-bottom groove that supersedes the circular oil return groove, allowing oil that does reach that groove to still be drawn back to the suction side. That groove accepts a #145 o-ring.
Then I cut my two grooves around the pump body with a slitting saw. My fixture allows pump bodies to be swapped in and out very fast and they settle right on center, so I can change out my workpieces faster than I can change the mill's tooling and positions.
I tried these body grooves with a nominally 1/8" slitting saw, but the resulting groove was about .130" wide and left very little meat between the large suction port and the groove walls. But leaving more meat there meant the o-ring was almost being squeezed out to the inside of the crankcase. That's why I'm going back to a #37 o-ring, so I can use my 1/16" saw and not have the grooves so close to the port while being sure the o-ring is fully contained and compressed within the pump bore.
The saw blade needs to be at least 2.5" diameter to be able to reach in under the pump flange ears.
And finally, I face off the covers with a flycutter. An .005" depth of cut takes out all wear and corrosion and leaves a nice dead-flat surface on one pass.
When I install the cover, I use a very light coat of Reinzosil on the pump face but strictly outside the face o-ring groove. That's just extra insurance against any oil that may seep up the bolt threads finding its way out. Same goes for the gasket under the pump flange. All the pressurised oil is contained by the o-ring seals and the nose o-ring of course also prevents air being drawn in on the suction side.
As for setting the all-important gear end-clearance, most pump bodies aren't dead flat across the face, the flange ears almost always turn up, so while each body is in the mill fixture I can graze them with an endmill to bring the face dead flat. In doing this I may calculate the cut depth so I end up with my desired .003" gear endplay. Or, if the endplay out of the box is too tight, it's very easy to chuck each gear in the lathe and take off exactly enough to set the endplay. It just depends on the pumps, they vary in where the errors are as all new parts do, which is why we blueprint things.