Gear pump

[Roger B]

The gears are commercial Delrin 12T MOD1 with the bosses turned off to give a width of 9mm. The flow is more than adequate for water cooling a couple of my engines.

I didn’t initially think much about the clearances but the first one seized due to the expansion of the plastic. The bores of the pump are around 0.05mm over size and the gears were a bit under the nominal so I would guess there is a clearance of around 0.1mm when cold which will reduce to almost nothing at 100°C. The initial version had almost zero face clearance so I increased this to 0.1mm which seems to work.

This is the first trial some years ago with my 500 rpm battery drill:

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

I am currently building a three gear oil pump for my Junkers CLM based on the same concept with commercial steel gears, an aluminium body and a brass thrust/port plate. As it hasn’t been run yet I can’t comment on the success or not

[jcge]

I am sure that this has already been said but gear pumps have to leak, slightly. If they don't the oil trapped the teeth, as gbritnell says, will produce very high pressure. This in turn fatigues the teeth. A running clearance, say 0.010", between the gear face and end plate will surfice.

AVTUR

I wanted to revisit this because I ended up adding spare PTFE gasket sheets between the middle cavity lock & end plates to alleviate what I thought was slight friction somewhere in the system. I still don't have a definitive answer.

I'm not sure that you'll need to address the tooth/housing pitting issue at such a small scale due to the expected greater leakage, and limited running time. However for visualisation purposes, here's an example of a small spot face relief on the side plate in the tooth mesh zone of a cam housing scavenge pump of the Porsche 917 (typ 912 engine).

Regards
John

[petertha]

Thanks John. I think that feature is what GeorgeB was getting at with his V8 gear pump if I interpret drawing correctly. I'm unclear about sizing though. Is the intent to encompass a certain proportion of intersecting teeth area from perspective of looking at the gear face on? Or is recess depth dimension important if it relates to gear width? Thicker gear = more pump volume so more 'reservoir' required?

[steamer]

I am sure that this has already been said but gear pumps have to leak, slightly. If they don't the oil trapped the teeth, as gbritnell says, will produce very high pressure. This in turn fatigues the teeth. A running clearance, say 0.010", between the gear face and end plate will surfice.

AVTUR

I wanted to revisit this because I ended up adding spare PTFE gasket sheets between the middle cavity lock & end plates to alleviate what I thought was slight friction somewhere in the system. I still don't have a definitive answer.

I'm not sure that you'll need to address the tooth/housing pitting issue at such a small scale due to the expected greater leakage, and limited running time. However for visualisation purposes, here's an example of a small spot face relief on the side plate in the tooth mesh zone of a cam housing scavenge pump of the Porsche 917 (typ 912 engine).

Regards
John

thanks for the shot John!

[jcge]

Not sure I can give any meaningful guidance with regard to sizing the relief in the meshing zone...it obviously must not extend beyond the gear pair center line, but it seems that backlash between the gear teeth will also provide for relief of the pressure spike at point of meshing at the center line.

https://www.modelenginemaker.com/index.php/topic,11357.msg265561.html#msg265561

Refer to Section 2.3,  Fig 26 & Fig 27 relating to pressure spikes.

At such small scale in these model engines, and given the limited run time I really wouldn't be too concerned about it.

[petertha]

I added a 0.004" shim to the pump body sandwich so effectively another 0.002" clearance per side of gear face assuming they self center while running. For one thing it runs very smooth now when I tighten the bolts hard. So its obviously is a bit tight axially somewhere in the assembly. I can tell no difference in pump performance with extra clearance, mind you this is just visual observation until I rig something up to better quantify things & ISO-32 oil. It readily self primes from a height of 6" (the length of my hoses) & drains a 30 ml cup in about 10 secs of 'slow Milwaukee drill' speed whatever rpm that is. So I think it would be satisfactory for an engine sump pump as-is. Maybe elevated pressure (lubrication pump) will be more challenging, that will be next. Another little observation when I was re-assembling. When I pushed the gear axle into the blind bearing hole it promptly ejected out of the hole under hydraulic pressure. These pumps make great fidget toys. One day when I find time I'll figure out a way to post videos on YouTube or similar.

[jcge]

See also section 4.2.1 Relief Groove of the previous reference
Regards
John

[petertha]

(VW Beetle oil pump, YouTube video). What do you suppose the ring groove with a single feeder is about?

[jcge]

The VW pump you've shown is sealed with a flat plate cover and paper gasket, and not an oring as the peripheral groove might suggest.

I believe the short radial groove connects the peripheral groove to the suction side (low pressure side) of the pump and is to provide for the return of leakage from the periphery of high pressure oil against the cover plate.

Note also the pressure relief spot face similar to that posted previously

Regards
John

[Roger B]

Hi Peter,

Have you got any further with the gear pump? I had the same fun as you with cheap internet brass gears 

[Mike R]

I will need a gear pump if I ever get around to making a multi cylinder diesel so I dropped a few dollars on one of these VW transmission gear pumps that seem popular with the hydraulic RC model crowd.

The first thing to note is that its already pretty small, not quite tiny and model engine sized but getting close.
It would fit in a 45mm cube (just under 2"), its actually only 37mm deep not including the boss for the shaft or the screw heads.
And yet it could probably be shrunk even more for model engines.

made up of 17 individual pieces total including fasteners, washers, gaskets and pins.
 
The example I received has 16mm diameter gears and they are 5.80mm wide, and if I counted right have 19 teeth.
They ride on 8mm diameter shafts.
The gear centers appear to be set at 14.62mm
The floating bearing blocks are 8.5mm wide. The main housing is 22.9mm wide, the gears and bearing block stack is 22.83mm
The next important observation is that the fits and clearances on this pump are amazingly close for something that was $9 plus shipping.
The gears and shafts are steel, the main housing is an extrusion, the floating bearing supports and end plates are cast aluminum.

A few more details that seem like they are important for this gear pump design.

The gear end faces and shafts appear to be ground to a high finish.
The floating bearing blocks have a channel that runs front to back on each of the bearings surfaces, it looks like it was cast in place and almost completely removed by machining. See picture.   Its placement makes me believe its an oiling groove by design.
The blocks also have a small channel that runs from near the gear mesh point to the shafts - its on the suction side where the gears are moving apart.
The blocks fit in the housing like a piston to cylinder fit or better, sliding fit but no noticeable clearance.
It is evident that the gears have cut their own clearance in the housing.
There is no seal on the input shaft or the inlet port, evidently this particular unit lives in a bath of oil and leakage out is not a concern.  The outlet port has a single o-ring seal.
I think what this examination is telling me is that absolute minimum clearances are a must, even to the point of interference to allow running in.

[Admiral_dk]

I presume that the Pump has been Run-In where it was produced and in a way where it only receives clean oil that also flushes out the Aluminium Chips from the Housing => a very tiny clearance and clean Pump to be delivered to the Transmission assembly line.

Just for reference - I had a Suzuki GSX1100EZ and from the Original Service Manual : see attached file.
Just to be clear - ALL the Suzuki MC's I have owned use this kind of pump and they all have a minimum Crank RPM of 1050-1150 + a gear reduction between Crank and Pump 2:1-3:1 (model dependent). That minimum is 50 RPM above, where the pressure drops off a cliff.

Per     

[petertha]

Hi Peter,

Have you got any further with the gear pump? I had the same fun as you with cheap internet brass gears 

Hi Roger, sorry for lapse. I've been on vacation. Yes made some progress, both in terms of cutting gears & lapping the widths relative to body. I thought I posted some pics but maybe it was partial and/or other forum. I'll gather my wits & update.

[Roger B]

That commercial pump is interesting, but as you say a little large for model engines. I wonder what the design pressure is, I guess it uses some form of automatic transmission fluid?

Tight tolerances are obviously a requirement however self-machining must require some quite clever production engineering and tolerancing.

I came across this site for the model hydraulic fraternity:

https://magomhrc.com/en/39-hydraulic-pumps

Some are based on, quite expensive, commercial pump units and others are standard closely machined gear sets. I expect the maximum pressures depend on the fluids being used.

My Triumph cars had the same design of Hobourn Eaton pump that Per shows. If I remember correctly the manual suggested lapping the end clearances with some fine abrasive paper on a piece of plate glass.

[john mills]

when i worked for company that reconditioned VW engines  they just ground the assembly of body and gears
flat across the end  clearance which the cover plate also ground flate would only have been the thin paper gasket
they did quite a number every day so it must have worked