Yet another keg furnace

Nice!

 

I'll get there in the end, it's becoming an epic quest to get a reliable oil burner that just works .... I was aiming for simple low maintenance equipment but I'll now go for a conventional approach instead. Nobody local supports oil burning furnace hardware anymore so it means cobbling together something from gardening, gas, plumbing, air, hydraulic and airconditioning hardware. Using off the shelf hardware is a priority for me, rather than mystery Ebay hardware that ends up unavailable down the track.

 

NOOOOOOOOOO!!!!

 

The spray nozzle was successfully tested today, I had to upgrade to copper plumbing from 1/4" to 1/2" which gave about 4 times the cross sectional area to the plumbing. With 1/4" tubes the pump was having serious problems with flow and the upgrade to 1/2" copper fixed it nicely. The DIY brass spray nozzle is working well enough to be usable although the spray stops working at lower flows. The pump has an internal regulator set to about 40 PSI and I should be able to deactivate it and fit an external adjustable regulator to get higher pressures and a finer spray.

 

That is a pretty slick little pump arrangement.
Should be interesting to see how it functions with the burner in the furnace.

 

I have high hopes for it, today was a proof of concept and to make sure the pump was working ok: it was, but it can't suck through 1/4" copper and then a few metres of 1/4" plastic tube. Right now there's a slight oil leak that a new gasket and silicone will fix if I allow the silicone RTV time to cure properly. I've ordered a $25 140PSI adjustable fuel regulator to bypass the outlet and I'll jam the internal 40 PSI pressure regulator which is a stiff spring and a piston that blocks a drilled bypass gallery. A friend suggested having multiple small nozzles to get around any lack of flow variation, each with a separate valve.

The fittings are all hardware store 1/2" British standard pipe household water hardware with nylon olives (optional copper olives). The plan is to have the pump a few metres away from the furnace with rigid 1/2" copper tube to the furnace and nozzle.

Edit: A power steering pump would also be a more suitable candidate if you used one of those inline diesel fuel filters with it and would be a gasket free solution, although I'm not sure of the power requirements for a steering pump. This current oil pump is easily driven by a half horsepower 1440 RPM motor.

 

The last few days have been spent working on the oil pump to hopefully fix some leaks: The pump drive shaft gets it's lubrication from a small internal groove from the high pressure chamber to the shaft bore. Halfway along the shaft's bore is a small drain where the oil leaks out after oiling the spinning shaft. The pump is designed so the oil drains back into the engine sump for reuse and needs no seals. This is fine with waste oil as the slow drip can be caught in a sardine can below the hole, once kerosene is used it becomes a squirting stream of fuel beside an electric motor with internal sparking on the start winding centrifugal switch . A tiny power steering pump is looking like a better option than doing this again.

There are similar pumps that use a gland and packing online and after asking around a bit I found a rotating shaft seal in Viton polymer, good for 140 PSI and high temperature to boot!. It turned up yesterday and I had the cast iron pump housing sleeved ready to take the seal. The oil hole now has a small brass elbow and I'll hook up a return line to the fuel tank and if it reduces pressure too much I'll play around with a restrictor of some kind.

 

Today I removed the hose barb and fitted a 1/8" BSP grease nipple and gave the shaft a few squirts of 2% moly grease and ran the pump for a few minutes on diesel: complete drip free success!. Last time I tried this with diesel, there wasn't enough pressure for the nozzle as it all freely squirted out the oil port all over the place. I should be good for a test run on diesel in a few more days.

 

Today was the most successful run of the furnace, The original 1.5mm nozzle was spraying too much fuel so I machined a new brass 1mm orifice which is less than half the area: 1.76mm^2 vs 0.75mm^2.
After a quick test of the new nozzle inside my prototype flame can, I trimmed the can to about 5 cm length or 2" so that it did not intercept any fuel droplet spray. Today I ran it from cold on diesel fuel and the modified oil pump ran beautifully, no leaks this time!. Starting involved some oily rags and it took a bit of fiddling to get it running: the ball valve is too difficult to use as a fine throttle, I'll need a needle valve of some kind next time. The flame can or baffle allowed the nozzle to be drawn back into the tuyere to any distance and bring the flame front with it, it works as intended and I ran it at 4-6" inside the tuyere. I did note the flame in the chamber making two revolutions of the chamber before exiting on some settings. The heat is now down low: the plinth was orange hot where earlier experiments in the old furnace had carbon deposits down low and melted refractory at the top. It's too early to get a feel for how it runs, I'll have to modify the fuel system and also come up with a way to seal the burner tube entry hole/tuyere to prevent leaks. This time the vacuum cleaner could not provide enough air flow, so I'll have to try a jumping castle blower and see how it works. I had about 15 litres/three imperial gallons of diesel in the tank and ran it for about an hour maximum and stopped when I ran out of fuel.


The cable tie on the copper tube (near the ball valve) indicates how far inside the tuyere the fuel nozzle is:

 

I like watching videos of furnaces operating.
Something about the fire and melting metal and all that is cool to watch.
That is a nice clean furnace design.

 

Thanks for the comments guys, at this stage it will need some changes but the basic design seems sound. The ability to withdraw the flame up the tuyere/air pipe seems to work. I'll have to take off the flame can for a comparison run, but this is the first time the plinth area of the furnace was orange hot. The A20 was in the furnace for less than 10 minutes before I ran out of fuel but was glowing a dull orange in that time. The so called flame can could rightfully be called a baffle as it's open ended and no real combustion takes place in it. That said, it does prevent the flame blowing away from the nozzle, so for the first time I can adjust where the combustion occurs in the furnace.

Some modifications:
1. Block off the internal pump regulator and fit an external adjustable pressure regulator and needle valve throttle.
2. Some sort of quick connect fuel plumbing would be less messy.
3. Test a higher volume air blower, I had to run at low fuel settings to avoid smoke/soot.
4. Some kind of gas tight seal at the tuyere is needed, some sort of flange

Soot deposits after approximately one hour of run time.


The outside of the flame can shows the colour change from heat discolouration while the 4" air pipe shows the soot from vapour leaking out the join.

 

The fuel regulator arrived yesterday and I was able to source a brass needle valve from a hydraulic shop. This morning I spliced them into the fuel system, I think there's about 5-6 different plumbing systems now with 19 TPI BSP tapered fittings going into 20 TPI SAE parallel holes with lots of thread tape, Australian metrified versions of British imperial water plumbing T junctions with nylon olives, sweated soft solder joins, brass compressed air fittings with brass olives, gas ball valves running into a brass hydraulic needle valve....all in all, a crime against plumbing to accomplish my goal. After today's run I'm very happy with the current fuel system: once lit, I can set the flow rate to the correct setting and use the ball valve to shut the furnace off when charging the crucible or skimming the melt, then shut the lid and crack open the valve and it auto ignites instantly.

It's apparent the vacuum cleaner pump can't supply enough air for the amount of fuel and I was just able to melt 9 kilograms of brass to pour into the resin sand mold. Of course I didn't clamp the mold and the cast iron and steel weights weren't enough to stop the lid from lifting off. If I had slammed the brass in or if it was a bit hotter I may have succeeded with the casting and my skimmer was pretty useless: I wasted a lot of time cleaning the molten scrap brass. The finish of the brass is quite good, leaving it to cool in the mold while the resin chars to carbon keeps the oxidation form forming on the metal.

 

Yes it's a record player turntable complete with 50 and 60 Hz strobe marks for 33&1/3 and 45 RPM. Binder is no bake polyurethane resin, I can't find any photos but I've cast three in aluminium so far: I'll go take photo now. The pattern is epoxy soaked MDF and sheet metal. It ended up being 9 kilos...so yes 19 lbs, it's a hefty sucker, I can't wait to machine a brass casting. This example was made from recycled runners that weren't cleaned of sand and you can see sand embedded in the lightly skimmed surface: it's deliberately cast upside down to try and float any contaminants away from the top surface. I have a Youtube video showing the resin sand mold being made: it takes about four minutes per mold, two minutes to pack the mold before the resin cures and goes off.

 

If I'd been a bit more prepared, I would have used clamps and steel bars which solves the problem neatly, I had been warned before. I think they'll almost certainly have to be balanced, I can machine them quite accurately but that only gets you so far. An active electronic balancer is on my "To Do" list eventually probably using those cheap silicon gyro accelerometer IC's as sensors. I have lots of brass scrap and it's just an easy alloy to test the furnace performance with. Bronze is an option, but brass does machine better and is a harder alloy than the bronzes I can currently get hold of. While I can passivate brass it would need a clear lacquer to have any chance of lasting more than a few years.

 

The furnace had it's best run to date today, the changes include:

• A smaller 0.8mm orifice running at 58 PSI (4 Bar) instead of the earlier 36 PSI (2.5 Bar).
• A jumping castle air blower for higher airflow than the vacuum cleaner blower.
• Running on Jet A fuel which is very similar to diesel fuel.

The small nozzle was to try and improve atomization at lower throttle settings which was where the furnace ran last time. After about 11 minutes of running, I remembered to withdraw the nozzle further back into the tuyere instead of right at the chamber (for starting)and the flame withdrew back into the chamber. The combustor can neatly fixes the combustion delay that allows unburnt fuel out the lid opening at high rates of burn: I could run it at higher fuel flows if I had an even bigger blower.

At the 36 minute mark all of the brass (9Kg) was liquid in the crucible, so it took 25 minutes in a preheated furnace to melt it. If I want to equal Peter's furnace on propane: 30 Kg in 20 minutes, I'd be looking at even higher rates of fuel and airflow.

The only sand mould I had is for a spindle housing, so I cleaned it up and shook out a cockroach and some loose sand: a cockroach + molten brass would be a spectacular steam explosion in a confined space so I dodged a bullet there. At the 58 minute mark, with a nice hot crucible and some of the zinc burning blue from the heat, I skimmed then left a chunk of hardwood on the metal to provide CO2 to reduce the oxidation and poured into the mould. Fuel consumption was 18 litres per hour, say 4-5 US gallons.




After cutting the runner off and a bit of a fettling on the belt sander:



There are three shrinkage holes: one where each runner was and a third on the upper left foot near the left edge: