Fitted a K type thermocouple to my furnace to help fuel air mix tuning

Discussion in 'Furnaces and their construction' started by Mark's castings, Jul 20, 2022.

  1. Fitting a large K type thermocouple on the inside of my furnace bore has really made a difference to tuning the furnace fuel/air ratio for the highest temperatures. I bought a 10 inch long K type probe with the ceramic beads over it from Ali Express a while back and last week I drilled through the refractory from the outside and angled the thermocouple into the bore about halfway up the side of the crucible. The thermocouple junction is naked but sits inside a small ceramic cup which I packed with kaolin clay to give some protection from the combustion gases. So while the furnace is running you can set a given fuel flow and vary the airflow and see what the temperature does in realtime: this really does take all the guesswork out of running an oil fired furnace in a big way. I learnt that I'd been running too lean all this time and maximum heat occurs with large yellow rich looking flames spitting 12" out the furnace lid. Also it became apparent the furnace needs constant adjustment as it heats up and the fuel system also warms up and changes characteristics. I can highly recommend a thermocouple as a tuning aid for your furnace, this unit has fairly thick wires so it's reasonably rugged and still works after reaching 1305 degrees C/ 2381 deg F.

    Gettin up there!!

    K type furnace probe.jpg

    Running the furnace at the hottest setting:
    furnace mixture setting.jpg

    Probe tip in the bore, hole sealed with kaolin clay.
    thermocouple install 2.jpg

    Some brass screw terminals to thermocouple extender wire.
    thermocouple install 1.jpg
    Last edited: Jul 21, 2022
    Tobho Mott and Tops like this.
  2. Petee716

    Petee716 Gold Banner Member

    It sure would be helpful to be able to monitor temp that way. I ruined my last iron melt by trying to blindly maximize my furnace temp and ended up too cold. I believe you’re well into the upper limits of a k-type. I wonder if going over it’s limit 15-20% will ruin it or just affect its accuracy at those temps?

  3. They say a K type's life is reduced from oxidation at the upper end of it's range (1300 deg C according to Wikipedia) , this one has fairly heavy gauge wires so it may have a decent life. K types eventually die from metal diffusion at the junction reducing output, the solution back in the day was the Australian development of the N type thermocouple but they are kind of rare. It was surprising how small a change would cause the temps to plummet and the ideal settings changed as the furnace got hot and no doubt as the fuel pump began to warm the fuel up recirculating through the pressure regulator. I said in the past that tuning it was as bad as tuning an Italian sportscar without realizing how true that was.

    K-type Thermocouple Sensor Ceramic Furnace Wire Thermal Detector Probe 2372°f 1300°c .png

    I've just been watching the video footage of the initial startup of the furnace where I shut off the fuel tap completely and let it run off the excess fuel sprayed onto the cold refractory walls and I noticed I had some blue flames as the mix went lean. I had initially thought it was due to the copper scrap I was melting but then that would cause green flames and not blue. While it was running like this there was an excess of air with the butterfly valve more open and the thermocouple temp was plummeting in this mode.

    lean burn 2.jpg
    Last edited: Jul 21, 2022
  4. Petee716

    Petee716 Gold Banner Member

    I think you're on to something very useful here!
  5. That's why I thought I'd post it here, all of a sudden I'm getting reliable feedback on temperatures as the furnace runs. I'll need to find another 100 degrees C to get temps to 1400 deg C if I'm going to melt grey iron (1260 deg C) reliably.
  6. Just an update with these cheap thermocouple: as 1400 degrees C gets close, which is the melting point for one of the K type thermocouple alloys the white insulator beads promptly melt and flow. I had some 6mm borosilicate glass tube lying around and after re-tigging the junctions for the third time I thought I'd use the glass tube in an effort to get a bit more high temp life out of them: Borosilicate is thermal shock resistant and softens at 1650 degrees C and should keep most of the oxygen away from the wires. The only downside I can see is the fragility which I can accept. Who knows, with such a snug fit to the tube, the wires might hold together when soft and survive for another run. It does save a lot of second guessing, but eventually I'll have to wean myself off K type sensors and learn to visually tune the furnace. I used a propane torch to soften the glass and pushed both tubes along the wire until they bent and smooshed together, after that, a bit of squeezing with the pliers had the glass in thermal contact with the junction.

    Borosilicate - K.jpg

    Borosilicate - K2.jpg
    BattyZ and Tops like this.
  7. Melterskelter

    Melterskelter Gold Banner Member

    You really need to be getting the furnace to something like 1650 C as pouring temp should be on the order of 1400C. There is not TC that I know of that can withstand 1650 for very long.

    If I were trying to measure the furnace internal temp, I would encase a TC in a stainless sheath mayb 1/2" OD with a 1/8" bore (dimensions chosen out of the air) and place the assmbly in ice water---for a standard starting temp or maybe in a bucket some constant known temp. Then I would insert it insert the assembly into the oven a set amount for maybe 20 seconds or 30 seconds and pull it out. I'd then observe the peak temp acheived (which would be slightly affected by the ambient air temp) by the assembly when sitting in the open air. It might take 30 seconds to a minute to peak. By observing the max temp reached you could not determine the actual temp of the furnace, but you could compare the relative temp and that could help you tune the furnace. For a repeat measurement, dunk the assembly into the bucket until the TC reading equaled the bucket indicating the assmbly had reach equilibrium and take a new reading.

  8. Eventually I'll have to ditch the internal thermocouple. I bought a cheap S type thermocouple that is claimed to be rated to 1300 deg C despite S type being capable of 1600 deg C. I also had a lead on some guy selling tungsten rhenium thermocouple wires by the metre which is good to 2340 deg C but the poor bugger is located in Kiev.

    The thermocouple allowed me to try all kinds of burn settings and I discovered only one particular setting works best. At this point it's a tuning aid that gives some confidence: I think I've got the fuel settings right, the extra hour taken to preheat 100Kg of dense castable (1650 deg C maximum) was throwing me off a bit. With the A6 crucible allowing much more volume of combustion, the iron is comes out of the furnace sparking all over the place and is hot enough to flow into thin gaps and cause flashing on the castings so I'm happy with the temps. The A25 takes up so much space the iron is barely melted (half full) after 90 minutes and just over 24 litres for 90 minutes burn time versus 12-14 litres over 70 minutes for the A6 and most of that is preheating the furnace.

    If I want to melt larger amounts of iron than say an A12 crucible, I'll have to build a bigger furnace with say a 40cm / 16" bore
  9. Ironsides

    Ironsides Silver

    A cheap way to see how hot your furnace is running is to use steel wire about 1 mm in dia or slightly larger and about a meter or longer in length, place it into the exhaust vent about 25 mm to 50 mm above your crucible. If your furnace is about 1500C or above that wire should melt quickly. With my R type thermocouple the temp is between 1550C and 1570C which is idealy the best temp for a quick melt and hot iron.

  10. Ironsides

    Ironsides Silver

    I had another look at my video and it does indeed take about a minute to see the maximum temp. This is why I don't use it very often in the time taken to see the temp. After all the years melting iron I can now judge by eye and length of time to know if the iron is hot enough to pour.
    Mark's castings likes this.
  11. Melterskelter

    Melterskelter Gold Banner Member

    I use an immeerson S type thermocouple of a type commonly used in commercial foundries. It comes up to temp in about 6 or 7 seconds when immersed in the molten iron. On some of my simpler castings I feel like I can be close enouogh by eye. But a few of them are unforgiving and will not fully fill if I am undrer by 40 deg F. So, you can be sure I measure those melts. and I end up measuring most melts I have also learned that late in the melt my metal rises about 10 deg celsius per minute when the furnace is running correctly. So, if I am 70C low, i can figure that in 7 mins I'll be at the corerect temp. If that target is known to be critical, I do a cofirmatory reading. If there is a little tolerance on the temp, I'll save the TC as each probe is only good for a handful of readings as long as you don't bunp it. They are very delicate. Reliance MT Duratip 12"

    Mark's castings likes this.
  12. Ironsides: I'm getting thin iron (from splashes) to melt in seconds when in the flame above the crucible, it'd be about 3mm thick or so. Any thinner and it blows back out the lid :eek:.

    Denis: I'll check out those disposable S types in future.

    I ran the furnace today and it performed the same as the other A6 crucible runs: it got to 1350 degrees C on the thermocouple at the 70 minute mark and burnt 20 litres per hour. After skimming the slag stuck under the melt and using perlite I added a small amount of ferrosilicon and closed the lid to give a few minutes more. About then (75 minutes) the fuel pump sucked air and lost it's priming so I pulled the crucible out and rather than pour the thin castings, I poured a 10 cm cylinder in a bucket. The thermocouple was working but the glass sheath had melted and run down the furnace wall like water. It had to be above the 1650 melting point of borosilicate glass. I'd run out of fibro cement for under the crucible to prevent sticking and poured some dry silica sand which had partially melted and firmly stuck to the fire brick. I had to break it off after the pour and it pulled a bit of the brick with it so the temps were close to the melting point of silica at 1713 degrees C at that area. The thermocouple is about level with the top of the crucible and the silica at the base would be about 8 inches/20cm below that.

    The K type protruded about an inch from the wall and has slumped down and the borosilicate glass has run like water (1650 deg C melting point)
    dead K type 1.jpg

    It's a bit hard to see but there's now a solid 6mm/0.25" white layer of sintered but not melted sand on top of the cream-off white coloured firebrick.
    sintered sand on brick 1.jpg

    The bottom of the crucible has a similar solid layer with the fused fibre cement patch centre-left.

    sintered sand on crucible 1.jpg
    Last edited: Jun 26, 2023
  13. I took some close photos of the sand on the fire brick as well as samples of unheated sand as a kind of sanity check. The heated sand is actually very glossy and shiny from certain angles and is definitely sintered, even with some kind of impurity fluxing it the temperatures would have been North of 1600 degrees C under the crucible to have this effect if silica begins to melt at 1713 degrees C. The unheated sand is entirely different in colour compared to the white of the heated sand. It took a few solid hits on the dirt to separate the hot crucible from the brick after the pour.

    Original plain old washed beach sand:
    unheated sand.jpg

    Heated sand under the crucible is now snow white with a few small basalt flakes in it:
    hot silica 2.jpg

    Same sand showing the partially fused shiny gloss of the sand.
    hot silica 1.jpg
    Last edited: Jun 27, 2023
  14. Ironsides

    Ironsides Silver

    If you can get above 1650C in your furnace have a go at melting steel in your A6 crucible because 1650C is the pouring temp for most steels. If you use ball bearing races to melt they are a high carbon chromium steel which will have a much lower melting point than low carbon mild steel. I had a look at at a iron carbon phase diagram and it is well under 1500C.
    Mark's castings likes this.
  15. I'd been thinking of putting some 304 stainless in the crucible and seeing if it slumped (like a primitive but cheap pottery cone) which would be the 1400 degree C range. I'll be happy if I can an A6 of iron hot enough to cast these 1/8" thick stove parts.
    Melting steel under a layer of perlite would be cool but doesn't it ignite without air protection like glass or perlite?.
  16. metallab

    metallab Silver

    For measuring temperatures of 1400 C and above, type S is the way to go. I have an alumina sheathed type S which works flawlessly. It is located just below the level of the top of the crucible, opposite the burner entry, via an 8mm hole in the furnace wall. When I melt cast iron, I let it go to 1500-1550 C and when melting copper alloys 1350-1400C.
    Alumina is much better heat resistant (mp 2030 C) and thermal shock resistant. When I am going to pour, I pull it out of the furnace when still white hot and let it cool.
    Unlike borosilicate glass, it does not soften and does not get sticky.
    Tops likes this.
  17. Ironsides

    Ironsides Silver

    Years ago I watched stainless steel being melted with an induction furnace and no flux was used. They were pouring fire grates for model steam locos. It was at a tafe college hobby foundry course and it was closed down 21 years ago so now there is nowhere in Victoria to learn foundry skills.
  18. I have a pink ceramic sheathed S type from Aliexpress which might be alumina, what's interesting is they claim a 1300 degree C service temp. I was thinking of connecting it to a digital multimeter and using a chart to work out temps and hope it survives.

    You'd get away with it using an induction furnace but may manage to oxidize if not burn it in a fuel fired furnace. There's a few videos on youtube, one of a USA foundry induction melting stainless and another of a Pakistani foundry melting stainless in an oil fired furnace. I think I'll stick with iron and non ferrous for for now.
  19. There's an Ebay seller in Ukraine with a lot of obviously Soviet era scientific and engineering supplies for sale: They have tungsten-rhenium thermocouple wire for sale, 95%-5% and 80%-20% available which they say will give 24.59 millivolts per degree C and go up to 1800 degrees C. According to Wikipedia this has to be in non oxidizing environments ideally, I'll try it out and report back on the results. I expect I'll need some sort of high temp ceramic sleeves like alumina to hold the wires and keep the worst of the flames off, Wikipedia's thermocouple article mentions the challenges with this type of material reacting with common sleeve ceramic and also with water vapour at high temperatures. I can't find the exact type of thermocouple junction these ratios form but it's similar to a type C or D thermocouple

    From Wikipedia:

    Tungsten/rhenium-alloy thermocouples
    These thermocouples are well-suited for measuring extremely high temperatures. Typical uses are hydrogen and inert atmospheres, as well as vacuum furnaces. They are not used in oxidizing environments at high temperatures because of embrittlement.[19] A typical range is 0 to 2315 °C, which can be extended to 2760 °C in inert atmosphere and to 3000 °C for brief measurements.[20]

    Pure tungsten at high temperatures undergoes recrystallization and becomes brittle. Therefore, types C and D are preferred over type G in some applications.

    In presence of water vapor at high temperature, tungsten reacts to form tungsten(VI) oxide, which volatilizes away, and hydrogen. Hydrogen then reacts with tungsten oxide, after which water is formed again. Such a "water cycle" can lead to erosion of the thermocouple and eventual failure. In high temperature vacuum applications, it is therefore desirable to avoid the presence of traces of water.[21]

Share This Page