Prototyping a spinning cup oil burner

ditto to what Myford said...

 

I was finally able to do a test burn on the new design, initially I did a couple of water tests and discovered the blower air would try to blow the cup shaft off the fuel tube causing the drive belt to come off. I did reverse the belt earlier for a few runs on the bench but it kept coming off the pulleys, probably due to the shallow rim on the shaft and not enough crowning on the driven pulley. Water testing also revealed the cup is a bit short for the job as water can enter the cup in a stream and do a lazy spiral off the edge with poor atomisation. Various half arsed attempts to diffuse the flow so it hit the cup didn't work and I bit the bullet and machined up a stainless steel disc with a tubular section on one side that is a neat fit for the bore of the fuel tube. There are four holes to allow fuel to enter the cup and the gap between the rim of the disc and the bore of the cup is close. The water now tends to get evenly spread by the cup rotation and has a chance to get centrifuged. I ran it for about 45 minutes with a break around the 30 minute mark of 10-15 minutes when I ran out of fuel and then ran it for 15 minutes after that. I barely got about 10 lbs of scrap brass melted, definitely not fluid enough to pour a casting. So far the unit runs really well and can be turned up quite high but the flame flow in the furnace is an issue as the flame only does one turn around the inside before exiting, I'll raise the crucible height with a second half brick under it and see if that helps things.

The only other issue was oil inside the stainless air tube, the ball valve appears to leak even when turned fully off so when the cup unit is withdrawn on shutdown, there's oil leaking out the cup in the air tube. Other improvements would be a sheet metal hatch in the cover in case the belt came off, right now the entire cover has to be removed to do this.

 

Hmmmm So what's the plan?

 

I'll open up the cover....yet again ... and see what I can find out regarding the oil leaking, it was dribbling water too during testing, too much volume for what would be in the line after the ball valve.
The flames also seem to be above the crucible mostly so I'm planning on raising the height of the crucible on it's firebrick plinth to see if that helps. This unit doesn't have the Mk2's surging when hot so it solved the surging of the U tube when hot. I think it would be usable once there's some fine tuning done on the combustion in the furnace chamber. J Vibert's concept of a design with a spinning shaft and external bearings is looking to be a better bet: it's entire length spins the fuel before it hits the cup, the bearings can be away from the heat and it can be solidly built. If I were to build another unit it'll be a rear fed, belt driven shaft with external bearings.

This one (Mk3) does everything it's supposed to do (easy ignition, no compressor, no pump, high rate of burn) but it feels a bit complex and finicky.

 

It's taken a while to get organised enough to do another furnace run: I finished the heat shield for the crucible lifter by fitting a 3/8" layer of mineral wool insulation between two stainless discs above the crucible opening. I bought some thick leather boots, even the tongue is leather, but synthetic laces. My overalls have legs long enough to go over the top of the boots and I borrowed a chrome leather apron to add some protection. Fuel is no longer a problem: a local fish and chip shop has 80 litres of used cooking oil a week for free, also a local tyre shop has scrap wheels for AUD $10 each for 8 Kg minimum weight wheels.

So this time there have been some slight changes: the jumping castle blowers turn out to have a motor cooling vent that bleeds air off the blower to cool the motor. I blocked 80% of it with silicone RTV on the secondary blower as lot of air was coming out of it and hopefully the air pressure is increased a bit. In the furnace I noticed most of the flames were above the crucible on a previous run so I added a second fire brick chunk to double the height of the plinth that the crucible rests on. This gives a larger combustion chamber/swirl chamber under the crucible now and hopefully a bit more time for the fuel in the chamber to burn fully. I filtered the oil through some large scouring pad material which got a decent amount of large solids and left some particles like a fine sand size in the oil which should traverse the system fine. The furnace started easy, there was some oil vapour smoke leaking out the cracks which disappeared once warmed up, so the oil must be burning faster and the cracks in the wall could be closing when hot too. The mystery leak of oil turned out to be caused by the static spreader disc inside the spinning cup: it creates a chamber full of oil that allows oil flow backwards through the cages of the roller bearings and between the 16mm spinning tube and the static 12mm tube before exiting as a spray below the angle grinder motor and filling the chamber with an oil mist..... .

Anyway the furnace was run with the fuel valve at about 45 degrees open or 20% of maximum flow rate on the throttle and tuned for a semi transparent orange, leaner flame rather than a bright yellow flame. The furnace was run for about ten minutes to preheat and the crucible was full of dull red molten aluminium in 20 minutes or 30 minutes total burn time. The furnace used 12 litres over 45 minutes or 16 litres per hour consumption at the 20% of max flow setting. I used my crucible gripper to get the pot out of the furnace, this time the heat shield was above the rim of the furnace and there was no hurry or heat at all when lifting the crucible out of the furnace although I expect bronze will be a different story altogether.

So I'd made a styrofoam safety guard model for a 10" grinding wheel and packed it in sand, being a large surface area and forgetting to place weights on top of the sand it failed badly. But that's ok I was able to get the furnace running hot enough to melt the perlite/silica sand/concrete/fireclay refractory. I'm pleased with how well the burner runs and will use it a bit before going to a Mk4 design.

 

So after a long delay, I'm making some improvements to the spinning cup burner in the hopes of getting decent performance. During the last run there was poor atomisation which was caused by the belt drive slipping on the flat pulleys. The longer it ran the worse the problem became: afterwards the pulleys looked chrome plated from the polish they got off the slipping belt. So it now has 5mm pitch timing belts running on decent size steel pulleys and the inner fuel tube ends several inches before the cup so the oil can come in contact with the spinning outer tube before the cup, hopefully that will spin up the oil better than the existing short cup with diffuser and reduce the oil backflow via the roller bearings. It's a compromise to avoid making a longer cup.

I have a 2000W single stage vacuum cleaner stuck in a 6 inch to 4 inch PVC downpipe adapter which has less volume of air than the jumping castle blowers but massively higher air pressure. The blower feeds a butterfly valve and then into the spinning cup unit. It's much more compact, but I may run a longer length of PVC to get the blower with it's sparking brushes away from the furnace and potentially sucking flammable clouds of oil vapour into it. I had a short trial run with water using the old wide ring nozzle and had a 90 degree spray pattern, but to take advantage of the higher vacuum's pressure I've sealed any leaks in the housing and machined a 60 degree cone steel air nozzle that reduces down to about 1/8" gap around the spinning cup. It's probably not practical but I want to see how far I can bend the oil spray pattern forwards.

So the unit is a bit better built than the Mk3 unit and should have better performance.....in theory.

 

I gave it a quick test run this morning with water: the cup was slightly off center to the right of the photo and it shows in the spray pattern with less deflection on the right hand side. Pressure is high enough that the vacuum blower unit partially popped out so it'll need some physical restraint. The interesting thing was that the deflection pattern stayed pretty much the same over a wide range of airflow settings, possibly the nozzle airgap is too small. I'm going to machine a decent, longer oil cup and see how it goes with a wider nozzle.

This version with fixed inner tube and spinning outer tube is too finicky and needs better accuracy of the carrier sliding in the tube. A Mk4 unit with a spinning tube in ball bearings connected to an oil tube via a teflon shaft seal is going to be easier to make, more reliable and less reliant on accurate construction. This current version is a bit of a design dead end but will serve as a cup/nozzle design test bed for now.



You can see in the photo, there's about 60 degree deflection on the left: a new record over 45 degrees and about 30 degrees on the right with it's smaller air gap. The cup was centred in the air nozzle on the bench but the build quality lets the cup move a bit in the tube.

 

I love the blower design... Great idea.

 

So far it's been an easy to build unit, after building a slide valve and now a butterfly valve, I'll go for butterfly every time. It's easier to build by far but needs a lathe to machine the butterfly. I didn't quite get the shaft on centre so I had to file the butterfly screw holes oval and mount it while in the closed position to get a tight closed seal but that was the only issue. It would ideally have a butterfly secondary vent for the fully closed position so the motor has cooling air and doesn't burn out if it's shut for long periods. The foam seal round the rim of the pump unit will seal it fine but it will pop out, so I now have some retaining blocks made from PVC to stop that in future.

The new test cup has been made from some 1045 steel I had lying around, the taper is nearly twice as long as the older unit. You can see the internal ends of the tapers marked with the red and blue marks for comparison. It will still need an oil diffuser for very high oil flows but for all normal flow settings it should be ok (hopefully ). It's actually much lighter than the stubby original and the cone is less than a millimetre thick, it has a slight curl outwards at the rim to encourage droplet formation. The shape is difficult to machine as it rings like a tuning fork, so I left the inside textured with very fine chatter marks like knurling and sanded the outside smooth. Afterwards the shape looked awfully familiar....it hit me after a while that it looks like the flash suppressor on a Bren light machine gun (see attached photo for comparison).

 

What burner?????

 

The Mk 3.5 had it's test run yesterday and results were a little mixed. A few minutes in I heard a pop inside the furnace which turned out to be the steel ring nozzle coming out due to the air pressure from the vacuum cleaner. After that I had to run the air valve wide open and the oil valve at 50% open to get a higher throttle setting than usual but running acceptably. The oil spray would have been a flat plane in the furnace and some was obviously pooling in the bottom and leaking out, wasting it completely.

The furnace melted the scrap lump of brass easily but due to a hard crust of oxides on top I didn't notice the exact melting time but it was on the hotter side of temperatures with the brass very fluid to pour and burning zinc visible. I threw a tablespoon sized lump of zinc into the melt to compensate for the amount lost and poured into a heated ingot mould to get 5 lbs of brass. I have no doubt it would have easily melted more brass for the amount of fuel burnt.

Fuel consumption was around 15 litres over 1 hour and 20 minutes run time with a lot of that oil wasted until the furnace warmed up: I may need an entry into the furnace a bit higher up the wall or some fancy compound air nozzle to bend the oil spray forwards more.

It's time for a decent furnace body to be made, this one has too many oil vapour leaks and the 2" wall thickness loses a lot of heat I suspect. The burner concept has issues with the wide, hollow cone spray that need a breakthrough but I love the ease of use: just throw a lit rag in the furnace, turn on the power and turn on the oil flow. The earlier oil leak issue between the inner and outer tube was solved by packing the air gap between with axle grease.


Here it is running at half oil setting and full air flow from the multi stage vacuum cleaner pump. The air ring nozzle has come out at this point.





Over 5 lbs of scrap brass (all the scrap I had ) and the 4" ring nozzle that came loose.

 

I'm almost done with spinning cup burners as there are few technical issues I can't yet solve. The last experiment is a combination of a spinning disc burner with a scientific paper that found a spinning cup with fine textured points along the rim gave excellent atomization, and had more effect than cup diameter or RPM did on droplet size. I dug out the "Fire" brand motor tool jig from the early experiments and machined an adapter to mount a 2"/50mm brass plated wire wheel on the shaft. It happily goes to 30 ooo RPM and the balance is vibration free (got lucky again). There is no need for goose neck bends or hollow spinning fuel shafts: a 1/4" copper tube almost touches the back of the wire wheel boss and the fuel gets into the wire bristles to be flung off.

Atomization is excellent: it's a fine mist with no droplets coming out the front like the spinning cups had at high flow rates. The full air flow of the blower blew most of the droplets out the front but a significant proportion still hits the rim of the tube and dribbles in a steady stream out the front: I'm tempted to make a small drain tube and sump to capture this for recycling but it would be a fire hazard I suspect.

 

Don't be too concerned if you have to shelve the project for a spell. With the groundwork you laid here, there shouldn't be a cold person left on earth if one had half a brain to read this thread. You've accomplished fine atomization without the use of compressed air to do the dirty work. Hurry up and drop some kind of patent protection on it before the &hinese steal it and dump it in HF for $29.95.