5kW electric furnace build

The pins are threaded and are holding both halves together, a bit of a nuisance to bolt together but makes a fairly solid box.

 

That would work well.

 

A few comments:

1) Nice part! Looks great all cleaned up.
2) Your flask could have been significantly reduced in width and length so that you would have had enough sound to better fill the cope. Likely not news to you.
3) If you are having breakout along the edges of the mold, your pattern needs some love. Also not likely news. But, the sand does not lie! ;-) I am amzed at how small a pattern defect can be and still cause sand breaks.
4) Remarkable to me that the petro was strong enough to allow successful parting despite the cope layer being as thin as it was. Now that you have proven to yourself that your system works, it is time to buy or make more petrobond.
5) Another good way to prevent cope float is to simply use long screws to pull the cope and drag together. Obviously, the threaded pins did their job just fine. But some nice 4" long GRK R4 general purpose screws are quick and easy and are good for 20 uses before the drive socket in the screws starts to wear out. I use shorter screws that type often to attach top and bottom boards, cope to drag, etc. An impact driver makes quick work of inserting them.

Denis

 

Good tips, thanks!

Haven't had time to use it much but finally started making aluminium ingots.
After maybe 10 hours of runtime I burnt my coils off in several places from aluminium touching them.
I suspect careless loading of the crucible caused some to melt off outside the crucible and hit the coils.
I tried to reconnect the coils after removing the damaged bits by splicing in new resistance wire, this worked for another few hours but seems to have made weakspots where things fell apart.
Got myself a decent stack of ingots now but with just as much left to make, I figure I need a proper solution...


Started by taking out the old coils.
The channels I've made for them were really difficult to work with, requiring a lot of fiddly bending to fill the different levels.
Furthermore I ordered a "real" graphite crucible and it turned out to be like a centimeter too big in diameter for this furnace...
The combination of bad coil channels and just too small size made me consider grinding the chamber out, working a rig now which lets me sink down and rotate an angle grinder around so it comes out symmetrical and nice...

 

3 phases where each leg is 230V relative to neutral. 2 legs are 380V RMS between eachother at highest point.
If I remember correctly I both measured current with one of those "touchless" current meters and also measured resistance/voltage and calculated through ohms law to come up with 5kW.

Looking back at the sales page, they're specified as 220V 3000W which should equal an resistance of ~16.1 ohms per element. I'm using 2 in series at 380V so I think I should be underpowering them.
Checking an old video it seems the current was 7.8A @ 223V AC which would give a resistance of ~28.6 ohm so slightly blow what 2 in series would be.
I can't find any data on when it ran at 380V but the math says 13.3A and 5049W using the resistance value.

From cold start it takes around 40 minutes for aluminium to start melting but I suspect part of that is from direct heat, radiating from the heating coils onto the crucible.
Once the thermocouple actually reads 700C(ever so slightly recessed into the hotface) the melting is much faster but I haven't taken any more "scientific" measurements on how fast.

Currently working on a solution where a small-ish DC motor with attached diamond bit will rotate down a threaded rod and basically spin-mill the chamber from the inside.
Since I used overly thick bricks there should be plenty of 'meat' for this task.
Once the diameter is correct for my crucible I will use a similar approach to make a spiral groove in the wall and place brand new coils there.
I'm thinking of putting two in series again but having 3 cables exit where one of them taps into the mid of both coils.
This would afterwards let me experiment with both 380V across the entire thing or 230V across each coil into neutral in the middle.

 

Been trying to understand the document, thinking out loud a bit.
I think my wall loading should come out to around 21kW/m² with some simplified geometry, this should be very safe up and above 1000 celsius, I will mostly work around aluminium temps though.
I noticed they also recommend dimming control rather than on/off, might be something to implement in the future if durability is a problem.
Further down in the document they recommend chamotte as wall material for grooves and a max of 35 kW/m² @ 1000C.
My wire diameter is a bit thin compared to their recommendations of 2-6.5mm, I still have some new coils leftover but could 'fix' that if these coils don't last(so far only breaks have been from splashing aluminium on them).

I decided to do a makeover due to numerous issues with the current design of the internals. Took the 'hardface' stones out and made more space for insulating firebricks instead, the 'pearlite concrete' was fairly easy to get through with a chisel.


Unfortunately this is how my firebricks showed up, hopefully I can salvage enough of it..


Started cutting some channels. Handheld router with shopvac for dust extraction worked really well.
First a conical bit to get the angled walls and then a round bit further down so I don't have extreme stress risers and also get more depth in the channels.
The perimeter will be built from 10 mitered stones where the 10th will have it's channels angled up(blue marker) so the path overall spirals upwards.
Doing some quick heat conduction calculations and this feels like the right way to go with conduction brought down to 2.6kW/m² from almost 29kW/m²...

 

Thanks for invaluable tips as always!

Tried mortaring everything together yesterday, we'll see how it holds up.
I dipped the bricks in water before applying the mortar but it still seemed to dry up extremely fast and want to come off, a lot like trying to paint an oily surface..
Added some extra around the top and joints just to give it a fighting chance of holding together initially.
To help out a bit I printed a piece to align the bricks against.


This let me bungee-cord the entire thing together whilst it hardens so nothing moves around.


Gonna give it a few days to harden up before I even touch this thing again.

 

I will make some thin steel bands to go around this then try to lower it down into the 'old' furnace and fill the voids back up with perlite, least amount of work and should give a very good overall insulation I reckon.
This time around I think I will loosely fill with perlite and then only at the top mix it with cement for the lid interface. Should further increase insulative properties below the top since the perlite seems to get partly crushed as I mix it with cement.
Should also save a lot of work from not needing to mix so many batches.
Will take a look around the shelves this afternoon to make sure they're still clear incase anything more has dripped down!

 

Checked through all shelves, there's some splatter but removed anything that could interfere with placing my heating coils. Don't want to prod around too much and accidentally break something.

Whilst this is probably unnecessary it felt like a good idea to add more structural integrity.
No idea what it's called in English but basically hole-punched steel strips ment for nailing things together.
Bent it into shape and screwed it together to keep everything clamped.


Had quite a bit of pre-mixed mortar that would otherwise have gone to waste so coated the outside and behind/around the strips for a better connection.


Quite a bit of space around the new 'insert'.


Lightly packed with loose perlite.


And the top finished off with perlite-cement.
I 3D-printed a collar so I had something to smooth the cement against and keep it from dripping down, once the cement sets I can tear this out.
With all the cement packed in place I covered it in plastic wrap and remounted the lid, this let me swing it into place and push it down a bit into the cement so I will have a matched surface for a better seal.

 

Thanks! Haven't gotten that far yet, I've predrilled holes through the insulating fire bricks and rounded over the edges so I don't have a sharp bend in the kanthal and made some extra holes in the steel casing.
I will try to guide the kanthal through a high-temp silicone/fiber sleeve where it goes through the "loose" perlite and steel casing, might also add heatshrink just as an extra layer, seems to have worked well for the first version.
I also got some high-temp caulk I will use to plug the holes from the inside, last version I had hot air escaping through the wire and thermo holes which I'd rather avoid for safetys sake.

 

I hear ya and completely agree about the thermal insulation bit, we can consider this an experiment to see how it will last.
Finally put the coils in, 2x 3kW that enters at top and bottom then comes out as a common middle. The idea is to connect them Phase1 - Neutral - Phase2 to get 230V across each.
SO MUCH EASIER installing coils now... Even if getting the exact length etc was a chore, this was still a 5 minute job instead of 30+ min.
The 'exits' are tripled(top/bottom) and quadrupled(mid) to lower resistance and heat, then sealed with some kind of fire sealant. The cartridge says it can handle 1250+ celsius but it's from amazon so we'll see.


Exits are as I previously explained, a few layers of heatshrink and then the silicone/fiberglass sleeve.
The sleeve says it can handle 300C, the heatshrink is rated much lower but seemed to work last time, we'll see if this works or starts melting.
Since it's all grounded the RCD will trip before anything dangerous can happen.
Since a bit of a cavity had formed behind the sheet(downside of loose perlite) I filled it in with the fire sealant to make sure the leak stopped.


Around the top I now have a high-temp rope, primarily to make a softer surface for the lid to sit on but it might also seal this up even better and reduce leaks.


Finally I gave the fire sealant a try on the lid which had a lot of cracks and small pieces missing. The description says it can be used to fix cracks and small gaps so figured it was worth a test.

 

Well, that entire coating on the lid came off in chunks after 6 hours of firing, the other fire mortar worked much better.
Furthermore Al203 really predicted the failure point
Coils broke right where I sealed the holes to the outside, most likely overheating right there.


In better news, the furnace has worked WAY better in this configuration, heats up to 700+C in 40 minutes and then stays there very reliably.
Managed to melt and pour ~10kg of aluminum before the coils gave up due to my stupid design.
Will overhaul this and not thermally insulate the wires this time.

Mostly made crude ingots but also made these two sand castings.
Unfortunately my 3D-printed patterns didn't hold their shape when I packed the sand but still happy with my second casting.


Some sanding and polish later, looks way shinier in real life than I managed to catch on photo.
A neighbour saw my crude ingots and said it would be really cool to have one for himself, so I made him a fancy one instead.
Second one is for my dad since I figure he might appreciate a weird gift like that.