setzep

Looking at the two gears I have here (pump gear 36 teeth and e-shaft gear 24theeth) the pump spins at 2/3's the speed the e-shaft spins. When you say 6000rpm do you mean engine speed or pump speed?



I'm trying to figure out why so much oil flow is needed? Do the engine bearings leak that much by and need 20gpm to sustain the pressure need? It would be intresting to see how much oil the rear relief lets by at 20gpm. At 20gpm the velocity is way too high through the pressure lines (32.6feet/second) and the oil pump pickup (14.5feet/second!). There has to be a lot of pressure drop. In the hydraulic field we don't like to see more than 16'/sec in the pressure lines and 4'/sec in the suction. Not to mention the reservoir (oil pan) is WAY too small for that flow.

BUT I suppose the pump has to be sized pretty large to sustain pressure at engine idle speed. It would be neat to see a automotive company make a variable displacement oil pump

I do like the oil pump pickup tube modification. The addition of radius edges are great and the o-ring seal is all the better.

Lynn E. Hanover

The 6,000 RPM is crankshaft speed.



The Swiss operation had calculated the pump output from its displacement. They had needed 16 GPM to get the rated rejection from the coolers they had.



The engine builders use different schemes to increase bearing clearance. Some hone out the bore of the rotor and bolt the bearing in place with little crush fit. Some just bore out the standard bearing down to the copper liner and run on that.

Some just run the factory race bearing that has only .0005" of additional clearance over stock clearance.



The bearing area in the rotors is very large and considerable clearance is run. In addition, a rather large pair of leaks in the nozzles that spray cooling oil into the rotors must be supplied and if racing, about .200MM each. The normal oil pressure would be 100 PSI or as close to that as you can get.



The quick way to improve the whole picture is working on the suction side. The Swiss found the best improvement from removing the bug screen from the bottom of the pickup. They also found good improvement from removing sharp corners from the pressureized lines and galleries.



My pressure lines are dash 12 and my suction line to the scavenge tank is a piece of 1 1/4" steel tubing. What would the speed be through a -12 line for 20 GPM and 16 GPM?



I have adjusted the pressure control on the pump to maintain 100 PSI. at 2,000 idle, oil pressure is 82 PSI.

Lynn E. Hanover

I don't know that, but it should be puting out 16 GPM at 6,000 crankshaft RPM.



It will not do this in the stock installation, but it should have, based on volume testing at low RPM.



We use a Peterson 3 stage dry sump pump.





Lynn E. Hanover

heretic

Hey Mr. Hanover, I've been thinking about that TDC finder setup you have and I think I found a way to refine it further, to make it even more precise and repeatable. The issue I'm addressing is that as you near TDC the amount of motion seen at the lever arm gets smaller and smaller, so the difference between TDC and a couple degree before or after is very small.



Assuming that the oil pan rail is 90 degrees to the rotor housing shape, one could chuck the plate upright onto a known level surface, attach a straight piece of angle iron to the surface of the housing, exactly vertical, in such a way that the e-shaft cannot rotate a full 360 degrees. Turn it until it stops, note the degree wheel numbers, turn it the other way until it stops, note the numbers, then split the difference, re-align the degree wheel to suit, and re-check it. No judgment calls with the dial indicator required, since the side to side motion of the eccentric is much greater per degree of rotation when it's not close to TDC (or BDC for that matter).



Maybe I'm just turd-polishing here. In any event, a jig could easily be produced out of flat stock and angle iron that locates off of the dowel pins, so it takes almost zero set-up time, just set the jig on the housing, pop the pins in, knock the e-shaft back and forth, recenter the degree wheel, recheck it, and done.



I make it sound so easy, I think I'll go and do it tomorrow.

setzep

The formula for velocity in a line is .3208 x V ÷ A = Velocity in feet per second. V= volume in GPM and A= area in square inches.



General rules for oil line velocity:



Suction, 2-4 feet per second (depending on how thick the oil is). If this can't be achieved then try to keep pump inlet vacuum lower than 4"Hg (the radius on the oil pump pickup works great like the swiss proved in the above example)



Pressure, this is a little more open. In a low pressure system like cars oiling system I'd like to keep it lower than 10-15ft/sec because of pressure drop. In high pressure (3000-5000psi), like some hydraulic systems up to 30ft/sec can be used if pressure drop isn't a major concern but space is.



Return, 10-15ft/sec (not really used in a cars oiling system, it's just dumped back via engine bearings).



Soo... if you say by the Swiss calculations the pump should put out 16gpm@6000rpm (engine speed) then that would mean the oil pump would have a CID (cubic inch displacement) of .924

Lynn E. Hanover

My two outstanding traits are that I am cheap, and I am lazy. Your idea will work, I think, but it sounds like more than I want to be involved in.



It is similar the piston stop method of finding TDC in a piston engine.



The rig I built has the fulcrum at one end, the crank throw in the center and the dial indicater at the other end.



The dial indicater moves twice as far as the crank throw. And you can use a version of the positive stop method with a dial indicator.



You set the indicator by rocking back and forth over the position that you suspect to be TDC and set the indicator at zero. Now set the degree wheel to the zero position and lock it down.



Move the crank in one direction until you get substantial movement on the indicator. For example .030". Mark the degree wheel next to the pointer with a sharpie. Then move back across TDC to the .030" position again and mark the degree wheel with a sharpie.



True TDC is between the two marks on the degree wheel. If the zero on the degree wheel is still the point between those two marks, your TDC indication is dead on, and you lock down the degree wheel on the crank and carry on.



If the zero on the degree wheel is not the point between the two marks, put that point in front of the pointer, and without moving the crank, loosen the degree wheel, and move the wheel only until the zero on the wheel is in front of the pointer.



Lock down the degree wheel onto the crank and repeat the procedure until the point half way between the marks is the zero on the degree wheel. Time and time again I find no difference between the raw selection of TDC based only on the first dial indicator setting, and a check of the TDC position based on the above.



Even the error based on the obvious parallax generated by this setup is inside of one degree. (moving the crank lobe toward the fulcrum generates slightly more motion than moving it away from the fulcrum).



I can resolve inside of one degree with just the indicator. I have seen no port timing data reported in less than full degree callouts. So, accuracy inside of one degree is close enough.



This sounds dreadfully complex, until you walk through it one time. Totally repeatable. Cheap and easy.



Lynn E. Hanover

Lynn E. Hanover

I'm sorry.



I skipped right over part of your question. The displacement I did not know.



THe 18 GPM at 100 PSI, actual from a flow meter on the dyno for 9,600 RPM.



The Calculated 16 GPM at 85 PSI from the engineers for 6,000 RPM are the figures I have seen on paper.



The 18 GPM is available from an aftermarket external pump.



The bearings are very large for the amount of stress involved and that tends to cover up the often poor performance of the stock pump. When trying to fathom what path the development my have taken, especially when the Japanese have been at it, one can only surmise that they have different divisions working on each

piece of the pie. No one is ever told that their idea just flat sucks, lest they "loose face" so they work for years on a piece that could have been completed here in an hour. Honda sends people here to work, so American supervisors "round eyes" can tell them when they are full of ****, and make them do the proper research and modeling before submitting proposals. Worse, the ones who have family connections cannot be told anything in Japan, and often arrive with one hell of a chip on their shoulders. They leave here as much better engineers, and are able to think on their own, rather than as a part of a big committee.



But once again I digress.



So if you take off the screen from the stock (86 style 13B pump at 40MM diameter)

the pump starts working pretty well. If you add a thick washer as in the drawing, so as to eliminate the vena contracta, it works even better. If you change the body, so that both sections of the pump have their own oil supply, it puts out 26 GPM at 6,000 RPM at 85 PSI.



Damn fine pump, and it was in the engine all along.



So then comes the 93-95 pump, and it is bigger (50MM) and has suction oil supplied to both ends of the pump, and it works great. But admit that the bug screen was a problem? Never. Admit that the shaded rotors were a problem? Never. Admit that pulling suction oil through the rear section and then forcing pressurized oil back through the same section was a problem? Never.



Why spend millions to change the pump if the old one could put out 26 GPM with a few common sense changes?



So none of the engineers that worked on it loose face.



This is a new pump, with all new design people.



If we loose 7,000 engines to foamed oil, it was us who were racing around operating outside the limits of the design, not the fault of the engineers in Japan.



I have yet to see an RX-8 pump. But the design probably sucks.



There you are. More than you wanted to know about oil pumps.





Lynn E. Hanover



The picture is the dash 12 pressure line into the engine. Note the oil pressure gage line in dash 4 and the oil temp bulb. The bellhousing is Chevy S-10 pickup truck.

Drago86

Lyn, would you recoomend removing the bug screen for a daily driver? I dont really see what good it is unless you drop a tiny nut into your oil pan or something..

setzep

Hey Lynn, I like the adapter you made for your rear plate. It looks a lot like the one I made

I thank you for all the info you've shared and I'll try to store it in my head for future. BUT, I have one question, the oil pump is a positive displacement pump. How can it put out more flow at a set rpm with one inlet and another amount of flow with a different inlet? I'm guessing tolerances between the gerotor ring and housing?

CGeek2k

I think it has to do with the inefficiencies of trying to supply both pump rotors off of one inlet vs. each rotor off of its own inlet. Im sure lynn will correct me if Im wrong though.