Lucas Mechanical PI System Model

[petertha]

Peter, my cantilevered pump design was intended to simplify construction (I think many commercial pumps are of similar design). The gears should be supported by the fluid film between then and the cavity. The bearing on the input shaft needs to be able to take the forces from the driving mechanism.

Two piece or three piece is an interesting decision. A two-piece pump design, like mine, makes it easy to keep the cavities and shafts concentric and at the required separation but the face of the cavity can only be as good as you can bore it. A three-piece pump is harder to keep concentric and at the same separation unless the parts are doweled together for the boring operation.

Regarding cantilever design, I went down the path of shaft supported by bushings on both ends more because I was loosely envisioning a pump with 2 banks of different thickness gears in one pump body. One to inject oil, the other to scavenge, both driven from a common input shaft. Obviously I have not gotten that far, and it does not pertain to your requirement anyway. I was concerned about alignment of the sandwich package, kind of collective tolerance syndrome of manual machining. My thought being I would bolt it togetether where everything rotated happy & then drill locating pins through the sandwich asembly so it would re-assemble that way. But this pump has taught me the gear width vs body width tolerance has to be just so & adding another gear cluster on top would be even more fun & probably overly complicated. Two seperate pumps might actually be easier in the end.

Regarding the floating plate (or not) thats why I mentioned the mini RC hydraulic pump. See links below which correspond to my post #69, exploded view last picture. There is no floating plate, just O-ring type perimeter seals. It arguably looks more like ours, but amazingly high pressure (20 Mpa)

https://magomhrc.com/en/brushless-hydraulic-pump-without-tank/486-mini-hydraulic-brushless-pump-m3-with-motor.html
https://magomhrc.com/en/120-brushless-hydraulic-pump-without-tank

Thanks for the link to miniature seals, now I know what I'm looking for. My O-rings were a stop gap measure using materials on hand.

[Roger B]

I made a quick test using the multigrade engine oil that I use in my engines. The pump I used has 24T MOD 0.5 gears and reached around 0.8 bar at 1000rpm pumping kerosene.

<a href="https://www.youtube.com/watch?v=OOZ8jw60siI" target="_blank">http://www.youtube.com/watch?v=OOZ8jw60siI</a>

I initially set the test to pump against the pressure gauge but the gauge was blown off as soon as I started the lathe. I then added a nozzle with a 1.2mm bore (nothing special about the diameter, it’s just what I had). This limited the pressure to around 2 bar, but as the pump inlet fitting is only 2mm more I think it may have been starved of fluid.

As a next step I reduced the orifice by inserting a 0.8 mm wire in the nozzle. In this case the pressure rose rapidly to over 7 bar and the pressure gauge blew off 

<a href="https://www.youtube.com/watch?v=Q79MUtT9SwY" target="_blank">http://www.youtube.com/watch?v=Q79MUtT9SwY</a>

[petertha]

That looks encouraging once you get the downstream fittings 'pressure rated' LOL. Thus far I have been using these M3 fittings from AliExpress. They come in different nominal barb OD/ID sizes. I also bought a collection of different flexible tubing suitable for gasoline/oil. I cant tell you which OD/ID combination I'm using now but no leaks there yet. It requires a good yank to pull the tubing off the barb so sealing reasonably well. Now whether it withstands anything close to your pressure, time will tell. They dont list any specs or pressure ratings, I'm guessing 60-80 psi based on industry analog but who knows what I actually have. Still working on machining some adapters for my gauge. Ulitimately in an engine (lubrication environment) this would probably be copper pipe like I see on steam models but thats for later, I'm still at testing stage.

I was trying to think of some easy, off the shelf way to vary outlet flow just to get me in the ballpark on test bench (vs. various drilled orifice nozzles which requires yet more machining). I wonder if if I could use the needle valve assembly from discarded RC carb? I know I can essentially shut off fuel flow with needle fully seated, but OTOH these carbs are not made for any kind of pressure, might just blow the O-ring. The journey continues.

[Roger B]

I made a new body for the MOD 1 steel gears in the same way as before. The gears seemed to mesh well with a, hopefully, 11.95mm spacing. I had problems fitting the gears into the cavities, measurement said they should be around 0.05 mm below the body but they were definitely above the sealing surface    The problem was that the milling cutter I was using as a boring bar had chipped    and there was not a square corner at the bottom of the cavities. I selected another undamaged cutting edge and re cut the corner. All then fitted together and the gears were around 0.05mm below the sealing face. I decided to make a trial with this clearance:

<a href="https://www.youtube.com/watch?v=TK0LVQSpRrg" target="_blank">http://www.youtube.com/watch?v=TK0LVQSpRrg</a>

I reached 5 Bar with one of my 1mm injectors (made at the start of this thread). The spray pattern was not very good so I decided to make new needles with an angle of 16° instead of 11°.

These were made in the same way as before, making two needles, one on each end of some 1.5mm silver steel.

[Roger B]

I will try the new needles during the next pump trials.

Another item still to be developed is the delivery valves for the shuttle pump. These need to open at a fairly low pressure, ~1 bar, and have to seal into the cavities in the sleeve with a small O ring. To keep things compact the spring is retained by the cone of the outlet union. (I will probably regret this idea when the spring and the ball have shot out a few times    ) The seat for the ball is reamed 2mm and flattened with a modified 3.2mm drill.

[Admiral_dk]

Always interested to read your posts and trials Roger.
Things always get complicated when they are scaled down far enough - no matter the subject !!!

It wasn't clear to me - are you happy with the latest pump or is the next step in Needle design needed to prove Pump progress ?

Best wishes

Per 

[Roger B]

Thank you Per    This is at the moment a series of experiments to see what is possible. A self-made gear pump that will reach 5 bar using lamp oil (probably less with petrol/Benzin/gasoline) is feasible. I now need to look at suitable injector designs to work in this pressure range. My current designs are all made with diesel injection in mind.

The delivery valves are sealed to the body using bonded (Dowty) seals. The ones I have chosen are 1mm thick so I needed to confirm that the clearance with the valve bodies screwed down was less than 1 mm. It was a little too wide so I shortened the thread by 0.2 mm. The tip of the delivery valve was then given a 45° chamfer so it would easily enter the 2.5 x 1 mm O ring that seals it to the sleeve.

The sleeve is only held in place by the two O rings that seal it to the body and the delivery valve O rings. The body O rings are arranged so that they do not have to slide over the ports (which can cause them damage).

[petertha]

Roger, I had not heard of Dowty seals, so I Googled them a bit. Where do you get them? I searched around my usual haunt (AliExpress) they have some but seems like smallest nominal diameter is 6mm. There is another style, what they call 'combined seal washer bonded'. According to the specs, good for 39 MPa. Do you think these would work? I dont have anywhere these pressures to contend with but I do see a bit of leakage using nylon-ish plastic washer under threaded nipple on my pump

[Roger B]

Hello Peter,

This is my supplier for bonded seals. They are available for threads down to 6BA and M3.

https://www.brw.ch/en-CH/screw-seal-usit-ring-nbr-70-shore-a-dichtomatik-860430

Various different materials are available and some have little ‘ears’ in the middle to centre them on the thread.

[Roger B]

A family shot of the gear pumps so far:

24T MOD 0.5 5mm wide steel gears.

12T MOD1 9mm wide Delrin gears.

12 MOD1 8mm wide steel gears.

I set up my commercial pump for some tests as the speed is more easily controllable using the bench power supply rather than using the Proxxon lathe which does not control down to zero rpm.

The first test was the delivery valves for the Lucas shuttle pump. These opened at around 0.8 bar, close to the calculated figure. The ball seatings are as machined, I haven’t lapped them yet (I may not need to    ).
 
https://youtube.com/shorts/dXQoXuAp128?feature=share

<a href="https://www.youtube.com/watch?v=dXQoXuAp128" target="_blank">http://www.youtube.com/watch?v=dXQoXuAp128</a>

The next test was the 16° injector needles. These were not so good, the opening pressure was not consistent and there was no atomisation   

https://youtube.com/shorts/n8yr5YsVdAE?feature=share

<a href="https://www.youtube.com/watch?v=n8yr5YsVdAE" target="_blank">http://www.youtube.com/watch?v=n8yr5YsVdAE</a>

Some further thought is required. The original Lucas injectors had a very wide cone angle, maybe I need to try something like 30°