leitmotif
Member
A couple minutes resulting in burn thru can result in weeks of repair. I remember rebricking furnaces and the melt furnace lid taking matters of days - and all I did was jackhammer out the old brick - that was only one shift but it was enough for me.
I think I would go into this a little slowly starting with taking Peters advice and shooting an arrow and seeing what you get and fine tune from there.
1. Monitor where you last burned thru. This may have been a matter of one rod a little longer or a real large chunk of scrap. Use one probe.
2. Use data logger with at least 4 probes - set other 3 equidistant
3. Start furnace from ******t temp
4. Record temperatures at 30 sec interval. Time is arbitrarily chosen mostly for ease of computation and comprehension at this stage.
5. Transfer time stamps and temps to computer
6. Now calc max min ave temp values for all 4 channels.
7. Now compute for dT/dt ie t@30 - t@0 and divide that value by 30 (second) do this for all 4 values recorded.
8. Now you have a baseline for how fast a normal heat up is. I predict your graph will NOT have four distinct lines. They will probably have a wide variation and just may look like a haystack. If this is the case ON THE GRAPH ONLY skip 10 or 20 values that may clean it a little. This graph will be one analysis method to determine how fast you want take your readings ie 30 second minute two minute etc. With this application the more data points the better - your overall data sets will not overload the computer.
9. If you repeat then your baselines will vary but you will get better data with repeats.
Look over all the data and determine if the 4 or 5 baselines are fairly consistant. If so now you are ready to drop some bucks and get a big data logger. Fluke is the first name that comes to mind mostly because I have used their 1982 version extensively and have their Hydra Data Bucket (circa 1995). I am sure they have advanced.
You could also go with a computer. If so then I would NOT connect it to anything else PERIOD. I would have very minimal software in it. If they want the data in another unit then use some memory medium to copy data and manually transfer. Minimize chefs in this stew - I would have only two or three people allowed to get into it and do any changes.
RAW DATA is SACROSANCT copy to a working file for any computations. NEVER compute in the raw data. COPY all raw data files for backup.
It is also worthwhile to spend a little time to think about what you are really going to do with this data and how you are going to apply it.
Peter you are more qualified to discuss computer option than I including software.
Dan Bentler
I think I would go into this a little slowly starting with taking Peters advice and shooting an arrow and seeing what you get and fine tune from there.
1. Monitor where you last burned thru. This may have been a matter of one rod a little longer or a real large chunk of scrap. Use one probe.
2. Use data logger with at least 4 probes - set other 3 equidistant
3. Start furnace from ******t temp
4. Record temperatures at 30 sec interval. Time is arbitrarily chosen mostly for ease of computation and comprehension at this stage.
5. Transfer time stamps and temps to computer
6. Now calc max min ave temp values for all 4 channels.
7. Now compute for dT/dt ie t@30 - t@0 and divide that value by 30 (second) do this for all 4 values recorded.
8. Now you have a baseline for how fast a normal heat up is. I predict your graph will NOT have four distinct lines. They will probably have a wide variation and just may look like a haystack. If this is the case ON THE GRAPH ONLY skip 10 or 20 values that may clean it a little. This graph will be one analysis method to determine how fast you want take your readings ie 30 second minute two minute etc. With this application the more data points the better - your overall data sets will not overload the computer.
9. If you repeat then your baselines will vary but you will get better data with repeats.
Look over all the data and determine if the 4 or 5 baselines are fairly consistant. If so now you are ready to drop some bucks and get a big data logger. Fluke is the first name that comes to mind mostly because I have used their 1982 version extensively and have their Hydra Data Bucket (circa 1995). I am sure they have advanced.
You could also go with a computer. If so then I would NOT connect it to anything else PERIOD. I would have very minimal software in it. If they want the data in another unit then use some memory medium to copy data and manually transfer. Minimize chefs in this stew - I would have only two or three people allowed to get into it and do any changes.
RAW DATA is SACROSANCT copy to a working file for any computations. NEVER compute in the raw data. COPY all raw data files for backup.
It is also worthwhile to spend a little time to think about what you are really going to do with this data and how you are going to apply it.
Peter you are more qualified to discuss computer option than I including software.
Dan Bentler
Last edited:
