PDA

View Full Version : Element wire



darryl
01-17-2016, 05:17 PM
Have gotten to the point where today I want to fit the heating elements into my heat treat oven. It turns out that the original elements were exactly twice as long as these now need to be. I can cut them in half and unwind a bit of each coiled end to use for the connection, and they will fit.

I'm used to dealing with nichrome, but this wire is quite brittle and has an oxide film on it. Everything points towards this being kanthal wire, which doesn't increase in resistance much as it's heated- less than 5% depending on the exact composition from what I've been able to find. My half pieces measure 8.2 ohms, which translates to drawing about 15 amps from 120 vac, or say 14 amps once the wire is hot. That in turn represents about 1600 watts or so per section.

What's interesting is that the original kiln was rated at 6000 watts and 240 vac. It had 8 elements in it, all wired in parallel, so each would have been rated for 750 watts. A half section of each element would then be rated for 375 watts at 120vac- quite a difference from 1600 watts.

I guess my next test would be to apply 120 to this half section and see whether the current drops to the point where the draw would be more according to the calculations. Perhaps I'll also apply 120 to a full length section and see what happens there. Maybe I got fooled by thinking that all the elements were in parallel where they might in fact have been in series/parallel- I don't think I made that mistake but it's possible.

At any rate, I have 5 grooves in which to install element coils- not a good number for a wiring combination. I'll have to see what I can do about that.

darryl
01-17-2016, 09:09 PM
Did a bit more checking. The original element coils are 16.4 ohms, so will draw about 14 amps or so from a 240vac supply. All wired in parallel that's a draw of 112 amps and 26800 watts. Doesn't make sense. Wired in series/parallel the draw would be 30 amps and 7200 watts. That's 7.5 amps per string of 2, x 4 strings. So that makes sense. I hooked up the half element to 120 and got 12 amps draw, where the calcs show 15, so at least I'm in the ballpark with the resistance readings.

It seems that in order to get the wiring combination to work out, each element must be carrying no more than 7.5 amps, which turns out to be 6.5 amps using my test set-up. Again, in the ballpark and probably correct. The wire never gets visibly red at that level, but then again I'm used to nichrome and not kanthal or the other alloy that's apparently sometimes used. I can see that when in the enclosure and operating for some time, the temperature build-up inside would allow the wire to get hotter and thus probably show color. I've looked into kilns in the past and the wire was glowing-

So now that I know what's right here, I can see that I'll have to wire it up as 4 half-elements in series. That works out to 32.8 ohms, 7.3 amps, and 1750 watts, in calculation anyway. In reality, the current is going to be 6.5 amps, as my meter shows in testing. The power then becomes 1560 watts, probably 1500 by the time the wire is running at temperature. It seems like not much power, but then again if the heat can't escape the temperature will rise.

I thought I'd have the coils in by now, but I'm still cleaning in the shop and taking stuff to the bin. More work on it tonite and tomorrow.

boslab
01-18-2016, 12:21 AM
Would think they are parallel, however I've seen a few in series before, there must be a reason but it baffles me.
Elements are easy to wind out of kath on the lathe, only needs a length of round bar as a former, it's bendy and soft when you buy it on a roll, once it gets heated up it goes hard and brittle, you can do a rescue straightening but only a red heat, aka oxy acetylene on the tail, difficult as the damn thing usually breaks at a cold bit.
Be careful if you do wind your own, would not use a lathe without a kick stop button! Or get a helper
Your not going to snap the stuff
http://www.kanthal.com/Global/Downloads/Furnace%20products%20and%20heating%20systems/Heating%20elements/MoSi2%20heating%20elements/S-KA060-B-ENG-2012-01.pdf
Mark

J Tiers
01-18-2016, 12:51 AM
...

I'm used to dealing with nichrome, but this wire is quite brittle and has an oxide film on it. Everything points towards this being kanthal wire, which doesn't increase in resistance much as it's heated- less than 5% depending on the exact composition from what I've been able to find.
...

In the linked blurb, it mentions on page 2 that the stuff changes resistance by 11 times from 20C to 1500. So you may need to consider your wiring differently. Of course, that seems to be for "Kanthal super", which may or may not be what you have. A little session with a variac may show you what's really going on. That might account for your current difference from expected, also.

darryl
01-18-2016, 01:46 AM
I definitely don't have kanthal super. The largest change in resistance I get is less than 10%, more like 7% or so.

Someone mentioned several days ago that after an element has been used it would be brittle. This is definitely true with this stuff, and there's a very significant oxide layer on it. To separate one of these elements into two, I attempted to pull apart a coil or two- it just snapped. Then I figured that in order to pull a short section into a straight wire I'd have to heat it. That worked fine. The oxide layer is pretty durable- I had to work at it to get to bare wire.

My main concern at this point is do I have enough power. I don't know how fast heat will dissipate through the bricks, so it's basically down to a test now. I need to mount the probe first so I can get a reading. I have very nearly 1000 cubic inches inside, roughly 10 x 8 x 12, or about half a cubic foot.

awemawson
01-18-2016, 04:23 AM
The original kiln wasn't 3 phase by any chance? In which case it could be be star or delta giving more combinations to get your 6kW

My kiln was three phase and I've re-wired it single phase as I didn't have three phase at the last place - oh boy it draws amps :) Now I have 3p I must get round to re-wiring it.

boslab
01-18-2016, 04:52 AM
On its first heating it goes through a phase change, it's no longer ductile it's completely brittle,
Just moving a furnace is enough to break them, watch they don't sag out of their grooves too, if you have to push them back, do it quickly while they are red, with the power off!
Although a peice of wood works but I found elements adjusted this way burn out at the spot the wood burnt, must be the carbon or somthing
I used to kill the power and use a steel rod
Mark

Jaakko Fagerlund
01-18-2016, 09:02 AM
My own oven has 75 mm thick bricks around it and 25 mm fire proof wool around it, the opening is about 220 x 150 mm and 450 mm deep. The power is 2.2 kW and it gets up to 1000 C in just less than 5 hours. I made a graph of how long it heats up to certain temps and this graph looks like logarithmic graph. I'm pretty sure the maximum out of my oven is 1100 C, though I haven't tested it.

The ovens volume doesn't play much role in the power vs. temperature, it is more to do with the inside area of the oven as that is what dissipates the heat away. I remember the recommendation was at least 1.5 W/cm² of inside area, otherwise it won't heat up high. Best to run the numbers to see how much power you need.

With Kanthal wire it is easy to do the necessary calcs from their own catalogs, the most critical is the power per area of wire (W/cm² I think it was). They gave 3-9 and I aimed at 3 to get long life out of the wire.

darryl
01-18-2016, 04:04 PM
Hmm- at 1.5w/cm2, I'd need 23 watts/cu inch. I have almost 1000 cu/in- that's 23000 watts. Not gonna happen here. I must have figured wrongly-

But I'll look that up and see if I can go by some numbers to get a handle on this. There must be a reason I'm hesitant to put these elements in- I don't want to go to that trouble only to find that it isn't enough power.

In the original kiln (single phase 240 by the way) the wire is being run conservatively in my estimation. When I tested a section at 11 amps of draw, it still did not get red in open air- and in the kiln it was running at about 7 amps or so.

A bit more homework to do on my part before carrying on-

darryl
01-18-2016, 05:03 PM
Ok, well it seems I need more power to heat my cubic inches. A bit of a re-design is in order. I'm going to line the inside of the chamber with 1 inch thick brick slabs, and make grooves in them so I can use 4 complete element lengths without modification. Then I'll have 3000 watts or so and less cubic inches, plus extra thickness of insulation. I guess I kind of got ahead of myself with this build :( See my sig line :)

Jaakko Fagerlund
01-19-2016, 11:28 AM
No, you don't use cubic inches (volume), you use square inches (area). You need to calculate your oven inside area (all six sides) to get the value you need.

I just checked, my oven data says 0.42 W/cm² and I remember it was a bit too little (found out after building it) versus recommended, so I think it was 1 or 1.5 W/cm² that was the recommended power for an oven.

Your oven is roughly 25 x 20 x 30 cm, so you have 2x25x20 + 2x25x30 + 2x20x30 = 3700 cm² of area inside the oven. With 1...1.5 W/cm² this would mean 3.7...5.6 kW of power needed for proper operation.

darryl
01-19-2016, 04:01 PM
Thanks, Jaako- that explains it. I'll do a recalc and see where I come to. I'll either use four of my existing elements, which calculates to 3500 watts, or buy some wire and wind my own to suit for a little more power. Might as well have it fully capable if I've gone this far with it-

Jaakko Fagerlund
01-19-2016, 11:57 PM
Thanks, Jaako- that explains it. I'll do a recalc and see where I come to. I'll either use four of my existing elements, which calculates to 3500 watts, or buy some wire and wind my own to suit for a little more power. Might as well have it fully capable if I've gone this far with it-
Yep. My oven has that 0.42 W/cm² and it will get to 1000 Celsius, though it takes 5 hours. With 1 W/cm² you shouldn't have no issues getting that high faster and having more control over it.

J Tiers
01-20-2016, 01:54 AM
That's DEFINITELY going to depend on the heat LOSS per cm^2.

What is the assumption you have for heat loss to arrive at that number?

The oven will heat up until the losses equal the input, then it will stop heating and maintain. The losses go up with the temp difference across the insulation. At some point there is a balance.

Losses will involve loss thru the insulation, leaks, and also may depend at least temporarily on the material. If it has a point where it absorbs a good deal of energy in a phase change, or a change in structure, it may even stall the temp rise, either temporarily, or effectively permanently, if that point is near the balance point/natural max temp and the absorption is too large.

macona
01-20-2016, 04:01 AM
Cold resistance is lower than hot so that will throw your numbers off.

Jaakko Fagerlund
01-20-2016, 09:37 AM
That's DEFINITELY going to depend on the heat LOSS per cm^2.

What is the assumption you have for heat loss to arrive at that number?

The oven will heat up until the losses equal the input, then it will stop heating and maintain. The losses go up with the temp difference across the insulation. At some point there is a balance.

Losses will involve loss thru the insulation, leaks, and also may depend at least temporarily on the material. If it has a point where it absorbs a good deal of energy in a phase change, or a change in structure, it may even stall the temp rise, either temporarily, or effectively permanently, if that point is near the balance point/natural max temp and the absorption is too large.
I don't recall where the numbers came, all I remember is that is was from an oven builder in business. It was based on having pretty much standard 75 mm thick insulating brick (K-23) walls, 1000 Celsius temperature that can be maintained (read: not running full blast to keep it barely at that and not needing half a day to warm it up) properly.

I know that an oven will maintain its temperature when the losses equal input, but that's irrelevant to the fact how fast an oven heats up. The numbers I've given give a reasonable estimate to the power needed to run an oven of some size, so you don't over power the thing or run it with too little juice to begin with. They are not be-all-end-all numbers, they are estimates. Of course if you add more insulation and make it air tight you can have less power to get the same results, but there is this limit called "practical".

Of course if you have more accurate way of calculating things like this, please do share. I'm just relying on the information gathered by actual builders of the ovens of what works and what doesn't, myself included.

darryl
01-20-2016, 06:27 PM
Well, I've calculated a new surface area inside of 3400 sq.cm, so I'm pretty much at 1W/sq cm. Of course there will still be a controller, and I'm thinking of tying it in with the thermocouple so it has something to use for a feedback signal. In the original kiln there was a pot, which you would adjust manually to control the average power level being fed to the heating coils. Someone had put a mark on the dial so you knew where to set it- presumably for a correct firing temperature which you could read on the pyrometer. Using it this way, the average power would be supplied to the coils all the time, which means a longer heat-up time than if you could power the coils full tilt until the temperature came up. Because the coils are being operated at a lower current already, I see no need to do this. I'd rather see them get full power to begin with, then throttle back once the desired temperature has been reached.

Throttling back in this case means reducing the duty cycle rather than going into an off, then on again type of control which would be harder on the elements. I'd like to work up a circuit that would start with full-on, then begin to reduce the duty cycle, working through a zero crossing triac so there is no sudden application of voltage to stress the elements (or make electronic noise or mechanical noise through the elements). Pretty sure it uses a triac right now and not an scr, but I better check on that.