When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
This pump is the most ingenious electro-mechanical device I've ever come across. Some years back I acquired a used one and studied it in detail. I commend you Brettus for undertaking this project but despite correcting some of the initial misconceptions, there still appear to be some misunderstandings. In view of the current interest in pump output, let me provide some input on how the oil is actually pumped, since your calculations appear to be out by a factor of two.
The pump has two "compound' pistons, a big one and a small one. Each piston has two pumping faces. One face is at the small diameter end and the other is the shoulder between the small diameter and the large diameter. The following condition must apply such that the shoulder pumps a volume per stroke equal to that of the small diameter face:
(DL x DL) - (DS x DS) = DS x DS.........................or DL = 1.41 x DS.
The diameters I measured are:
Small piston, small diameter.....1.82 mm
Small piston, larger diameter.....2.58 mm
Big piston, small diameter.....3.84 mm
Big piston large diameter.....5.43 mm
As you can see, the ratio of large to small diameter for both pistons is 1.41. I interpret the language in your description of how you did your calculations to mean you only used the large diameter, which would give a result that is too high by a factor of two. I have not checked the other parts of your calculation.
I thought I knew everything about this pump, but it is clear now from the discussions you have been having with kevink0000 that my understanding of the big piston valving is incomplete. I'll be following those discussions with interest.
Thanks for your input Delmeister but I'm still pretty sure the output is correct.
Yes I was aware that there were effectively two pumps per piston . The smaller one is exactly 1/2 the x-sectional area of the larger one so that each has the same output . That is because each one pumps to an individual output pipe on one upstroke and to another on the upstroke at 180 degrees . Effectively 4 separate pumping operations per rev for each pump.
Yes , I used the large diameter to calculate output because effective x-sectional area is : large diameter area -small diameter area + small diameter area = large diameter area.
I agree that the pump is ingenious , so much so that it took me most of a day and several nurofens to figure out how it actually worked.
I'm still not 100% convinced that there isn't reduced output from the Sohn adapter. Mazda put a lot of design into the flow path for the oil that supplies the pump ,I posted this pic on the other thread.
This showed they provided a means to get air out of the oil and also reduce the pressure of the supply oil. The pressure is still positive however ...someone in the other thread mentioned 5psi but I'm not sure where that came from.
The pump relies on two things to get oil into the pump , the pressure at the inlet plus the pressure from the spring at the end of the piston to push the pump down onto the cam . That spring isn't very strong so I believe the main force on the piston pushing it down comes from the pressure in the oil.
In the case of the Sohn , that pressure is less than what the stock pump setup sees so that puts more burden on the weak spring . At low output and rpm that is likely not an issue . But when the pump has to pump almost 2 litres per hour it's likely that the cam follower doesn't seat back down onto the cam all the way and output of the pump is reduced .
Difficult to test accurately . I think the best test would be two cars running the same number of laps on a track day and comparing oil consumption . If the Sohn was less ..perhaps a larger spring would be worthwhile ?
I think the cam is what actually produces the pressure, and the spring simply keeps the pump following the cam on the return stroke. High oil vis and high cam RPM might cause it to "float" but I doubt if that is a practical circumstance in this engine ever. That would have to be some high engine speed and very high viscosity oil. If the Ren. front cover system is similar to the old mechanical pump models, ( I have no idea if it is or isn't) it will likely be OK as a gravity system. The old mechanical models used Mikuni injector pumps which were identical in the critical design elements we are discussing to units used on various 2 stroke powersport engines at the time. Those were strictly gravity feed units at the time, and the ones in use now, even being and more advanced electronic, are still gravity fed only. This is a guess, but I would think if the OMP feed system was subject to positive pressure, it would also need a high quality, very precise regulator circuit as well, because of the deliberately small quantities of oil the factory metered. Any pressure difference at the input would change the output. Plus, I think the pump itself is not sealed well internally at all, as the presence of vacuum at the metering nozzles is well known to cause the whole line to drain, with oil getting sucked right through the pump itself, hence all the check valves on the output side lines, and the vacuum equalization system attached to the metering nozzles themselves. My thought is that if the output side does not respond accurately to vacuum, then it follows that the input side would likely respond similarly in the presence of positive pressure at the input. Also, I will get my test bench results tomorrow, time has not allowed it yet.
I think the cam is what actually produces the pressure, and the spring simply keeps the pump following the cam on the return stroke. High oil vis and high cam RPM might cause it to "float" but I doubt if that is a practical circumstance in this engine ever. t.
Agreed . Just throwing it out there ..... It may or may not be an issue , but those springs are very light so I can't see the pump relying on spring pressure alone on the downstroke.
Yes , I used the large diameter to calculate output because effective x-sectional area is : large diameter area -small diameter area + small diameter area = large diameter area..
OK. I thought you had used that diameter for each pump
Well, I spent a lot more time than I expected playing with this thing yesterday. It was kind of addictive, honestly. I ran so many combinations, and didn't get my HoneyDo list done at all. Found out something interesting too, that could be a big flaw in this design.I was less impressed with unit when I finished than when I began.
Conclusions:
Mod#2 does not need the pump cylinder drilled out to get even flow to both sides. That pump hits the cam 2x per revolution of the pump,(not the main shaft) so its either a full oil load to each set of holes every revolution, or a half load 2x per revolution to each set. The quantity is the same. The feed holes to the rotor injectors are top and bottom directly across from each other. I ran it with the original small pump, the big pump undrilled, and the big pump drilled. The big pump obviously flows more, but the ratio of big to small was unchanged regardless of drilling or not. I think Mikuni ran it with two holes because of the small output designed into that side of the pump, and wanted a more constant flow due to the low oil quantity pumped by the small pump at low throttle. (really a minuscule amount) But, I found I didn't especially like that idea, so I left the large pump in the small pump chamber undrilled, for reasons below:
I am not sure if this occurred because of the leaky nature of my test stand, the repeated disassembly and introduction of air into the system, the low rpm of my drive motor( a 900rpm rated drill, which is probably not even that fast), or if this OMP unit is faulty in some way, but the output from the bottom pump lines was consistently greater than the upper outputs. The flow began faster from the two bottom outputs, and was consistently more over the time period I was checking it. The only time they were all even was at more than 40% throttle, and the oil was flowing fast. This disparity was consistent regardless of all the pump configs I tried. It may be because I was opening the pump and starting with air in the lines, (which I would purge for 5 min or so each time, so I didn't see any air, and I would purge air as much as I could before closing the chamber cover each time. ) or because my system of oil feed leaks a bit, or maybe because of internal leakage at the low rpm test speed combined with the low throttle openings.( I feel this is most likely) This is why I kept the one-hole large pump for mod#2, I wanted as much separation as possible between the pump outputs, as early as possible in the flow chain. I was concerned that gravity would bias the flow even more to the bottom outputs if there was an output pass-through in the pump itself especially at low throttle settings.
Also, the pump outputs that I saw the larger outputs both serve only the front rotor. They have the shortest run to the injectors, and are the lowest on the pump itself. I think both are advantages. Maybe this is another reason we see more wear and lower compression on the rear rotor. At high throttle I don't see any difference, but at the typical throttle ranges a street motor sees, there may be a bias. Again, I can't be 100% sure how large the difference is, because my test RPM was so low. Equal to 1800 rpm ES speed, or less. I think internal leakage plays a larger part at low throttle,low rpm settings on the OMP.
Feel free to give any input.
Last edited by kevink0000; 10-16-2018 at 01:35 PM.
Reason: comma
Wow ...nice work Kevin . The way the aluminium body is drilled vs the sleeves would support my earlier theory for sure but the fact that in practice you don't get a different output of drilled vs undrilled suggests to me that there is leakage around the brass sleeve.
Yes, I think there is a lot of leakage in the pump itself internally. Kinda like the Wankel itself!
I am not sure the sleeves leak enough to make the kind of output I am seeing, I still think the holes lining up 2x per rev, but with half a load vs 1x per rev with a full load, is why it works out OK.
Plus, it seems Mikuni felt the larger pump was adequate with one set of side ports for the big pump turning faster speed in the large pump chamber. I think they may have found poor accuracy or repeatability, or recovery time, or something else, with one set of ports on the small pump turning slower. (That is a very tiny amount of oil that the small pump handles).
Even now with your work, Delmeister's, and my own redneck apparatus, unfortunately I am still only guessing with this thing.
Also, how did you remove the main worm drive shaft assembly? Did you have to remove the blind plug on the side? I worked on it a bit yesterday couldn't give it enough time.
I am not sure the sleeves leak enough to make the kind of output I am seeing, I still think the holes lining up 2x per rev, but with half a load vs 1x per rev with a full load, is why it works out OK..
As oil is not compressable or expandable ..... the pump chamber has to be filling with oil as it rotates past the blind holes in the side of the aluminium body. You can't pump oil twice via the cam with oil that is already in the chamber. It needs to be refilled.
Originally Posted by kevink0000
Plus, it seems Mikuni felt the larger pump was adequate with one set of side ports for the big pump turning faster speed in the large pump chamber. ..
The large pump sleeve only has one set of ports because both sides of the aluminium body are drilled . So it picks up oil in 2 locations per rev.
Originally Posted by kevink0000
Also, how did you remove the main worm drive shaft assembly? Did you have to remove the blind plug on the side? I worked on it a bit yesterday couldn't give it enough time.
Just remove the cam first then pull out the worm drive.
Both sides of the pump chamber are drilled for oil intake on the small side too. The small pump cylinder only has one set of intakes, same as the large one, IIRC, So the big pump or small pump picks up oil 2x per rev. Or am I looking at it incorrectly? I don't have it in front on me now. I need to bring this thing to work so I can post about it 24/7!
Both sides of the pump chamber are drilled for oil intake on the small side too. The small pump cylinder only has one set of intakes, same as the large one, IIRC, So the big pump or small pump picks up oil 2x per rev. Or am I looking at it incorrectly? I don't have it in front on me now. I need to bring this thing to work so I can post about it 24/7!
Thanks again for all the info.
lol .... If you take another look at it you will see there is no way for oil to get to the blind holes on the small pump, except via leakage past the brass sleeve. The holes you can see are just there because that would have been the only way to get the galleries drilled. They are plugged at one end and have no access to the oil flow.
Both opposing holes on the aluminium body have access to oil flow on the large pump however .
I would run the larger pump that you drilled ..... even though your results suggest it isn't necessary , because I can see how it is supposed to work.
Yes I see that blind feed hole now. I ran more tests last night, I will do more in the coming days. My new theory is that the one hole large pump works ok at my test rpm, but may not get enough seepage to fill on the second stroke at higher rpm.I will install my drilled large pump and see if consumption increases over what I had before.
Also, interestingly, the one hole big pump running in the small pump chamber should be biased to the two upper oil outputs, but again in test after test, configuration after configuration, the bottom holes see bias at low throttle settings.
Also, interestingly, the one hole big pump running in the small pump chamber should be biased to the two upper oil outputs, but again in test after test, configuration after configuration, the bottom holes see bias at low throttle settings.
Can you clarify where you are seeing the lower outputs please ? I'm not sure if you are talking about the bottom holes of the pump piston or the bottom holes of the output hoses. If it's the latter ........ can you post a pic of the outputs you are talking about because 'bottom' could mean anything depending on which way up it is.
FYI for the record, I have had mod#2 running on my car now for about 4 months now
So, FWIW I personally don't think its too much oil,
Mazda really had to starve this engine of oil to get it to market, past Ford, or past EPA/EU regs, or whatever. Too bad, really.
I'm really leaning toward your way of thinking .
With a stock OMP :
*Street cars that spend very little time in higher rpm/load range wear out the housings very quickly .... I've seen heaps of housings in engines from 50000 miles up to 100,000 miles . The housings are always worn out.
* Race car engines last quite a long time and wear on the housings is minimal after many races .
Conclusion : not enough oil at low load rpm on the smaller pump but sufficient oil for high load/rpm once large pump kicks in.
So ... for an NA:
fit large pump (drilled) in place of small pump . Adjust oil settings in higher rpm load ranges to around 45 (for same oil flow as stock)
Probably could adjust the lower settings as well flat lining at 1 to around where stock reaches 20.
For turbo :
fit large pump (drilled) in place of small pump . Adjust oil settings at 100% load to around 45 and taper up to 60 at 150% load .
The choice to run the Sohn is neither here nor there IMO but doing the above would definitely remove any need to add premix to the tank with or without the Sohn.
Something like that .......... could help our engines last a LOT longer !
The two lower banjo bolt fittings as the pump is attached to the front cover, this is where I am seeing higher oil output consistently on the test stand. Those are also the shortest lines, as both go to the front rotor. I have the pump in the same orientation as in the car on the stand, minus the injectors.
Agreed on the oil issue for this engine. The stock 2007 unit I took apart at 186k miles had a lot of housing and apex seal wear, but very little bearing, iron and side seal wear. The sides of the apex seals that contact the rotor had almost as much wear as the facing side. It was a fretting kind of wear, like what you get on a dry ball bearing. "dry" being the operative word. The FB and FC cars don't see this kind of housing and apex seal wear, and manage to stay in one piece with good compression for over 250k miles.
Mazda really tightened up the oil output, I think for emission reasons only. Many have said," trying to market a car that uses oil is hard," etc, but I don't believe that is the cause at all. That pump produces good oil output at high throttle settings, and the factory fuel injection flashes that I have seen will go open-loop even on the highway at moderate cruise throttle settings, so I think they were only concerned with the various emission laws and the test loops for same. Correct me if I am wrong, but I believe the RX8 to be the first and only OBDII rotary from Mazda. I don't think the FD ever went to OBDII, since it stopped selling everywhere but Japan right around the time of universal OBDII adoption. I thought Japan had more relaxed emission regs for many years vs the rest of the word, and that was the main reason it got pulled from most markets in the mid-late 90s, but was still sold in Japan almost until the RX8 came out. I think the OBDII and ULEV ratings were the main cause of the oil issue for sure. As soon as the map is out of that possible test range area, it can go very rich on both fuel and oil, even with stock settings. Less so on oil , because I still think Mazda were concerned with cat poisoning, from injecting regular motor oil with its metal anti wear compounds, and poor burn characteristics, again for OBDII compliance. Anyway, my thinking on this engine is oil,oil,oil and more oil. The cooling system sucks, too, but that is another matter. Less radiator area than an FB, is a recipe for the overheat and runaway temps we see all the time. When the only thing truly regulating temp on most of these cars is ambient temp and output, somebody got their sums wrong!
The two lower banjo bolt fittings as the pump is attached to the front cover, this is where I am seeing higher oil output consistently on the test stand. Those are also the shortest lines, as both go to the front rotor. I have the pump in the same orientation as in the car on the stand, minus the injectors.
Agreed on the oil issue for this engine. The stock 2007 unit I took apart at 186k miles had a lot of housing and apex seal wear, but very little bearing, iron and side seal wear. The sides of the apex seals that contact the rotor had almost as much wear as the facing side. It was a fretting kind of wear, like what you get on a dry ball bearing. "dry" being the operative word. The FB and FC cars don't see this kind of housing and apex seal wear, and manage to stay in one piece with good compression for over 250k miles.
Mazda really tightened up the oil output, I think for emission reasons only. Many have said," trying to market a car that uses oil is hard," etc, but I don't believe that is the cause at all. That pump produces good oil output at high throttle settings, and the factory fuel injection flashes that I have seen will go open-loop even on the highway at moderate cruise throttle settings, so I think they were only concerned with the various emission laws and the test loops for same. Correct me if I am wrong, but I believe the RX8 to be the first and only OBDII rotary from Mazda. I don't think the FD ever went to OBDII, since it stopped selling everywhere but Japan right around the time of universal OBDII adoption. I thought Japan had more relaxed emission regs for many years vs the rest of the word, and that was the main reason it got pulled from most markets in the mid-late 90s, but was still sold in Japan almost until the RX8 came out. I think the OBDII and ULEV ratings were the main cause of the oil issue for sure. As soon as the map is out of that possible test range area, it can go very rich on both fuel and oil, even with stock settings. Less so on oil , because I still think Mazda were concerned with cat poisoning, from injecting regular motor oil with its metal anti wear compounds, and poor burn characteristics, again for OBDII compliance. Anyway, my thinking on this engine is oil,oil,oil and more oil. The cooling system sucks, too, but that is another matter. Less radiator area than an FB, is a recipe for the overheat and runaway temps we see all the time. When the only thing truly regulating temp on most of these cars is ambient temp and output, somebody got their sums wrong!
They don't appear to be the holes affected on that chamber. The upper holes should get bias not the lower, in my opinion. (feed hole gets exposed and takes oil, then next event in that chamber is to pump to upper holes, not lower holes) But, it is interesting that the front rotor seems to be getting more lube, and its is usually the rears that show lower compression first, and ultimate failure. The bottom line is, if there was enough oil, it would be enough oil, even if one rotor got more. That is my premise with running extra oil, more is better, and too much is still better than too little.
I agree with you, that the mysteries of this pump are a contributor to the lack of success of this engine. I think the S2 design is how it should have been done from the beginning. True "oil injection" not oil dribble, maybe, sometimes,varying amounts, if the moon is aligned, etc.
Yes, and interestingly, it will pump backward if reversed.
And higher output in those lines with every possible combination. Even 2 small pumps. But, at high throttle, parity is achieved, in all cases. (RPM the same always) That was why I thought internal leakage was a possible explanation.
So .................we still can't explain the output variation !
I really don't think it's the pump design . So long as the pump refills on each stroke , and I can't see why it wouldn't, it should have equal output to all four hoses.
Maybe something to do with gravity feeding it .... but even that doesn't add up .
I think it's kinda sloppy in there, that's my guess. As the pump strokes get larger, the disparity soon becomes nil. I would guess the same for rpm. That sounds like internal leakage to my untrained mind. I can't guess at what else it would be. It doesn't seem designed in, or it wouldn't favor the front rotor with the shortest run. If anything, if there was a built-in bias, deliberate or circumstantial , Mikuni would spec the rear rotors fed from that output, or even a 50/50 split. We may never know...
10-14-2018, 02:32 PM
