Jump to content
  • Announcements

    • Alan Longmire

      IMPORTANT Registration rules   02/12/2017

      Use your real name or you will NOT get in.  No aliases or nicknames, no numerals in your name. Do not use the words knives, blades, swords, forge, smith (unless that is your name of course) etc. We are all bladesmiths and knifemakers here.  If you feel you need an exception or are having difficulty registering, send a personal email to the forum registrar here.  


  • Content count

  • Joined

  • Last visited

Community Reputation

1 Neutral

Recent Profile Visitors

147 profile views
  1. First smelt failure post-mortem analysis

    We do! That's actually how we noticed the furnace was dying off, we initially thought the temperatures droping were a sign of a thermocouple failure (which was kind of expected as we used K type thermocouples up to 1200 °C). We did see the furnace going darker and darker. Nice Facebook group, I'm joining Thank you for your kind words.
  2. First smelt failure post-mortem analysis

    Hi there! Thank you for your kind words. As it turns out, I have news about the project: basically, we're idiots. Using the experience we gathered during our previous attempt, we set up a new bloomery. We're still using a combination of precast blocks to build the furnace, except this time we're going with clay chimney liners (rated to a very high temperature) instead of cinder blocks. We're still lining the entire thing with clay for a nice, thick clay wall. The furnace was way better than the first one, and this time we were able to monitor the temperatures at 3 different heights very closely (more to follow). We used the same fan as we did the first time, but we now have some doubts it is still able to maintain a high enough air flow when the counter pressure builds up in the furnace. What are you guys using? Anyway, we made a very simple mistake, but still gathered some interesting results. Last time, we used a bowl to measure our dose of ore (one small saucepan being approximatively 500 g of ore). So without thinking too much about it, we re-used the same pan. As it turns out, since I have been so much more selective about the quality of the ore (aiming for ore that had about 90 % magnetic content, as suggested by Lee earlier), the bulk density of my ore has dramatically changed. Our saucepan was now holding 2.5 times the mass of ore it did on our first attempt! This is great news for the ore quality, since it does mean that we're charging much more iron oxides into the furnace. However, it also means that we charged the ore too fast by a factor of 2.5 in the furnace! We were aiming for a 4:1 fuel/ore ratio to begin with, then increasing up to a 1:1 ratio. Because of our mistake, we actually began with a 1.6:1 ratio (which works well enough), and went as high as a 0.4:1 ratio (which is terrible). So basically, after a brilliant start, the furnace died out being saturated with ore. This is quite evident on the temperatures that we recorded: Please note that T1 denotes the temperature closest to the bottom of the furnace, T3 denotes the temperature closest to the top and T2 is somewhere in between. We're more interested in the variation of temperatures than in the actual values, which may vary based on the depth of the thermocouple. The first thing we can see is that based on the evolution of the T3 temperature, it took about 1.5-2 hours to pre-heat the furnace (which is not much since we used pre-cast blocks). We waited for this temperature to stabilize to begin the reduction. Then, we see that there is a ~40 minutes delay between a variation of the ore charging rate and the temperature. This is due to the ore "hang time" in the furnace, and it correlates almost exactly with the fuel consumption rate we observed (by measuring the speed at which the top surface of the fuel/ore mix goes down in the furnace). So about 40 minutes after we introduced ore (at an actual ore/fuel ratio of 1:1.6), we observed a gradual increase of the bottom temperature, which was expected (yay!). This increase was even more noticeable when we increased the ratio up to almost 1:1. But then we went way to far and everything went off, as you can see on the graph. So what next? Well, we do believe that we have some appropriate ore and thankfully we still have more than enough to try again. This time, obviously, we'll be more cautious about the ore weight. I expect us to build a new furnace next month, so stay tuned!
  3. First smelt failure post-mortem analysis

    Alright - we're getting ready for our next try. Before that, I wanted to know a little more about the ore we'll be using. My ore mostly comes from the cliff side I showed in the first post. I was able to classify it in 3 major categories: Banded sandstone I focused on the dark bands that I thought were more likely to have a higher-than-average iron content. Surprisingly, this dark coloration only exists on the surface of the ore up to a depth of about 5 mm. This appears very well after roasting (which renders it quite magnetic): Beneath that, the ore is mainly made of reddish sandstone, much like the layers above and beyond the ore in the cliffside. Dark homogenous sandstone This type of rock is very dark and appears to have a much larger grain size. The color is uniform throughout the rock. It does not change colors much after roasting and does not become magnetic: I took it after I saw the same kind of rocks in a quarry nearby, which is supposed to yield iron-bearnig rocks: Iron crust At the bottom of the cliff, I found large crusts of layered deposits. It really looks like thin layers of rust! Some of the crusts have very interesting-looking inclusions: This one becomes somewhat magnetic after roasting: Test protocol I decided to try to use magnetic roasting on the ore. My reasoning is that the ore is mostly made of (hydrated) iron oxides and mostly silica. Roasting the ore *should* transform Fe2O3 into Fe3O4 which is magnetic, allowing me to remove it from the silica. I roasted small samples (8-10 g) of each of the ore types using a gaz burner. I then powderized the ore and used a magnet to remove everything in the powder that is magnetic. I repeated the experiment on 3 samples for each ore type to verify the results. Using a precision scale, I weighted the ore before and after roasting (to get an idea of how much it loses), I weighted the powder and the magnetic and amagnetic contents separately. I also checked that magnetic + amagnetic = original weight, just to check that I didn't lose anything in the process. Results The results are presented below: Banded sandstone : 30-40 % magnetic content Dark homogenous sandstone : below 1 % magnetic content Iron crust : 80-92 % magnetic content The error (difference between the original weight and the sum of magnetic and amagnetic contents) was less that 0.5 % of the original weight each time. I was quite amazed at the amount of magnetic content in the iron crust. I don't know how it translates in terms of iron content, but I believe it is quite promising. The dark homogenous sandstone was quite a surprise as it appears to yield very little iron at all. This was confirmed by a test on the similar stuff I found in the nearby quarry, which yielded almost the same results. What now? Well, according to everything I read (which was confirmed by Mr. Sauder here), I shouldn't use anything below 50 % iron. I also know that too high of an iron content will inhibit slag formation and prevent the reduction reaction from happening at all. Let's say that the magnetic content translates well in terms of iron content. I'm thinking about mixing the right proportions of iron crust and banded sandstone to aim at something like 67 % magnetic content. What do you think?
  4. First smelt failure post-mortem analysis

    Hi Lee, I was not sure you would read this on the forum. Thank you for taking some time to answer me, much appreciated. As Joshua stated, this is great reading material. Hats off to you sir! That's a great and simple idea. I'm going to try it on several samples to get an idea of the iron content. I'm thinking about using a paper filter to get rid of the iron chloride solution at the end (and maybe monitoring the pH to know for sure when I don't need to add any more acid). I'm also going to try to get a rough estimate by comparing the density of iron-bearing sandstone with regular sandstone (that I can find right over or under the ore layer). The use of this specific ore (that is, on this precise location) is well documented in the middle ages, so I'm rather confident about the iron content. I didn't include that in my original post for the sake of brevity, but the cinder blocks were indeed lined with kaolin clay (as was the tuyère). We probably underestimated the thickness that was required though, we're thinking about doing it with a much thicker clay wall next time. The cinder blocks idea originated from this video, where they seem to have wielded "good enough" results. Indeed it was, the type lies into the unit, we measured around 5-6 m^3/h, not l/min. This equates to around 100 l/min. Still not good enough! We're working on improving this. Follow-up question: do you think the appropriate air flow correlates well with the furnace dimensions? Duly noted. I'm planning on mining ore on a larger scale next week. We'll try again in a few weeks. It really does. Thank you again.
  5. First smelt failure post-mortem analysis

    Yes, my region has a rich (no pun intended) history of iron extraction during the late Middle Ages. I found the ore at the exact spot that is documented in archives. I had no luck finding analysis results, though. I'm trying to get in touch with a geological lab which could help me on this particular matter. However, the fact that the ore is both quite magnetic and able to get a very nice red color after roasting leads me to think that iron is indeed present. What do you think?
  6. First smelt failure post-mortem analysis

    Hi, Thank you for your answer. How do you suggest I should do this? I've looked into Lee Sauder's "iron dumpling" method, but even with the forge at full blast, all I can do is heat toe inside contents of the dumpling to a nice red hot. No trace of any iron bead inside. The contents went from a 50/50 redish/black mix (roasted ore/charcoal) to an even black mix. The charcoal is still there and I guess the iron ore transformed into magnetite hence the darker color. I wonder what kind of temperature I should aim at to get the iron bead...
  7. Hi all! I recently got interested in iron smelting and decided to give it a try with a couple of friends. We just got back from our bloomery. Although it was not a complete failure, we were unable to obtain a workable iron bloom. There are a couple of things that we do not understand and we would like to get your advice on them. Furnace construction and protocol Ore Our ore was iron-bearing sandstone that I found near my home: We roasted it until bright red. It became quite magnetic after that so I am confident it carried at least some amount of iron oxides. We broke it in small pieces and dust and used all of it. The overall quantity was quite small because of the preliminary nature of the experiment (about 10 lbs). Furnace Our furnace was inspired by this example (Coated Tyle furnace by Skip Williams and Lee Sauder), i.e. made from 10" cinder blocks stacked about 35" high. We dug out underneath the cinder blocks and lined the bottom of the furnace with clay bricks. The total height from bottom to top was about 40". We used a steel pipe about 1" in cross-section to build our tuyère and planted it about 6" from the bottom at a 20 degree downward angle. We measured the air flow at about 5 liters/min, which we thought was at the low end. We added 3 thermocouple ports at various heights to monitor the temperature (but only ended up using one because of cable melting issues). The whole thing was thermally insulated with foam concrete. Charcoal We decided to calibrate our charcoal so that we only used 1" pieces. Otherwise it was classical barbecue charcoal. Results Charcoal burn rate and ore charge We were surprised by the high charcoal burn rate as we naively expected it to be way lower. In the end we had to charge about 1 lb every 10 minutes in order to keep the level steady. We were focused on using a 1:1 ore-to-charcoal so we burned through our small ore quantity in about an hour and a half. We then continued adding charcoal so that ore could travel all the way down the furnace based on the burn rate (about 1 hour to get from the top to the bottom). Temperature We encountered a lot of issues with our thermocouple. We know that the core of the furnace was about 1100-1200 °C (2000-2200 °F) when we added our first ore charge. After that, we tried to play with the air flow to adjust the charcoal burn rate (that we thought was way too high) but since we were a bit anxious we went back to our max flow for the remainder of the experiment. We know that temperatures were about 1200 °C (2200 °F) when ore was present at the bottom of the furnace. Slag formation We had planned for a slag tapping door at the bottom of the furnace. When we opened it (some time after the final ore charge), nothing went out. We observed a lot of dark slag forming through our viewport, and periodically ran a steel rod through the tuyère to remove some of it the best we could. Final product The final product is depicted here: It was a dome-shaped structure located in the center of the furnace, underneath the tuyère (as was expected). However, when putting it on the anvil, we discovered it was mostly moss-like slag. It certainly contained iron oxides as it was observed to be magnetic when cool. Some places had traces of the original redish ore. We found no indication that a reduction reaction of any kind had taken place. We theorize that we only got slag with dissolved iron oxides. A lot of material from the foundation bricks was vitrified and had mixed with the slag. We think that the bricks were not rated to temperatures that high and might have polluted the furnace, preventing the bloom from forming, but we have no proof of this. Questions Do you have any indications about what went wrong inside the furnace? We believe that either: - The bricks were not adapted for building the bottom of the furnace. We are planning on using Mr. Sauder's furnace clay recipe (that we used for the furnace door here, which went well) next time. Our bricks might have polluted the furnace. - Our ore charge rate was very bad. We should have begun with 1:4 ore-to-charcoal ratio and increased progressively to 1:1. Do these hypotheses make sense? Do you have any other recommendation? More generally, we would be very interested in your opinion about this experiment. Many thanks!