Post by thread:High temperature sensor

Truncated match.
PICList Thread
'High temperature sensor'
1998\10\26@172233 by Mark Willis

Anyone ever do temperature sensing in a vat of molten metal? (This'll be a ceramic-lined steel vessel.) I'm thinking the consensus for temperature control and sensing here was a nice thermocouple, IIRC, but what about at as high temperatures as say 2600 F or thereabouts? (Thermal remote sensing of the top surface might well work here, I need to learn more.) I've searched my piclist archives on this machine, no joy, must've lost that set of messages describing the sensors when I lost a bunch of mail in August. I'm open to whatever -works- well & is reliable, exact +- 1 degree F temperature sensing isn't mandatory, but +-25 degrees F or so would be nice (I think.) Lots of thermal mass here <G> This'll be a future PIC project (December or January) in a factory setting, but I want to start getting a clue now in my spare time... Idea is to warm a batch of metal up (ramping the temperature up nicely without wasting fuel etc.) and then hold it in controlled temperature conditions of whatever profile for a set length of time, telling a host PC what's going on (the PIC's the sensor front end, the PC will handle announcing "When it's Soup" and so on <G>) and controlling things from the host PC, probably, for safety. I still "don't know what I need to know" on this all. But I haven't signed an NDA either yet <G> Anyone used super high temperature sensors of any kind in this sort of project? A commercial sensor would be wonderful, this is a "Need it to work right forever" type of application (so I may use redundant sensors & have the PIC tell me when one fails, and plan to replace sensors periodically.) May be talking a water-cooled PIC module, with optoisolated outputs & some severe power de-spiking... I don't think any Dallas chips like the DS1287 (?) will work here! <G> I probably will just do subcontracting on some of this sort of package, too, as this hits (It's one of those NDA contracts, but if I can subcontract everything to get everything done I think they'll be happy, so long as I end up with PCB data, source code for the PICs, etc. so I can maintain things for them. Looks like these folks may actually launch, finally, after years on my back burner getting things together... I won't have time to do it all quickly, myself. I may have to do some program re-working as they may want all the gory details kept to myself, for all I know. They're a little worried by some past problems they've had!) I'm thinking of just doing all their various code on 17C family chips, but maybe I'll try to fit everything on 16F84's (so much to learn...) Thinking about it <G> Mark, spam_OUTmwillisTakeThisOuTnwlink.com

1998\10\26@180622 by Dave VanHorn

Mark Willis wrote: > > Anyone ever do temperature sensing in a vat of molten metal? > > (This'll be a ceramic-lined steel vessel.) > > I'm thinking the consensus for temperature control and sensing here > was a nice thermocouple, IIRC, but what about at as high temperatures as > say 2600 F or thereabouts? (Thermal remote sensing of the top surface > might well work here, I need to learn more.) You might need to move to more exotic metals, but thermocouple would be the most accurate. Sensing off the top surface would mean knowing the temp of the top layer of insulative oxides. They are nice and reliable, and you know if it's working or not, for sure! You could take a specific mass, held relatively cold, of something that won't contaminate the brew, and dunk it for a set time. At any given temp, that would transfer a set amount of heat, so the peak temp would tell you (albeit indirectly) what the temp of the bath was. Takes more smarts that way.

1998\10\26@193514 by Craig Lee

Here's a thought. I have absolutely no experience here, so take it with a grain. I would attach a bunch of high temperature thermocouples to the outside of the vat. The temperature on the outside of the vat should be directly proportional to the inside of the vat. Taking an average of the ADC readings, and knowing the calibration factor of molten metal to vat temperature, you should come up with something reasonable. Of course the outer temperature will lag the molten temperature by some time, but perhaps this is acceptable. Another consideration would be keeping the air temperature outside the vat from influencing the thermocouples. This could be done by insulating the thermocouples with material with the same insulating factor as from the coupler to the molten metal. Just some thoughts.. {Original Message removed}

1998\10\26@215848 by Mark Willis

Only one problem: If they're heating the vat from the outside with gas or such heating, those thermocouples are going to melt, Craig <G> Good thought, but I think I'm stuck with inside the lining of the vat or something. And I *know* I'm stuck with learning a lot on this project! (Gonna love it <G> Been "stuck" for years helping my girlfriend, she's about at the point where I can let someone else take over, it'll be SO nice to focus on technology for a while on a good hard project that pays reasonably, AND know I kept her safe with some good help!) Mark, .....mwillisKILLspam@spam@nwlink.com Craig Lee wrote: {Quote hidden}> > Here's a thought. I have absolutely no experience here, so take it with a > grain. > > I would attach a bunch of high temperature thermocouples to the outside of > the > vat. The temperature on the outside of the vat should be directly > proportional > to the inside of the vat. Taking an average of the ADC readings, and > knowing > the calibration factor of molten metal to vat temperature, you should come > up > with something reasonable. > > Of course the outer temperature will lag the molten temperature by some > time, > but perhaps this is acceptable. > > Another consideration would be keeping the air temperature outside the vat > from influencing the thermocouples. This could be done by insulating the > thermocouples with material with the same insulating factor as from the > coupler > to the molten metal. > > Just some thoughts.. > > {Original Message removed}

1998\10\26@223313 by Gabriel Gonzalez

I knew of a foundry where the vessel was made of ceramic tiles, so they drilled little holes on some tiles around the vessel and inserted thermocouples in the holes. Calvin {Original Message removed}

1998\10\27@112333 by Peter L. Peres

On Mon, 26 Oct 1998, Mark Willis wrote: > Anyone ever do temperature sensing in a vat of molten metal? I had to do with something like this, once. The sensor was a 1 meter long metal tube with no weldings or solderings on it. Inside it was hollow, had ceramic macaroni insulators holding a tungsten wire that made contact with the tube at the tip. It worked as a thermocouple. The last 30 cm were of covered with some other material, probably ceramic. I don't know who makes these but I know that devices like this are used in the alloy industry to check on batches etc. I suppose that asking a sensor supplier for a contact thermocouple for 1500 degrees celsius will yield the right answer. I remember that there are some contamination issues with this, depending on what is being melted and how much. Thermal treatment is way less hot than the actual melting for casting or forging. You need to find out more... Peter

1998\10\27@144155 by Wes A Brzozowski

Greetings, all! As a long time lurker to the list, I guess I should first introduce myself, and then thank the many contributors here for their insights and expertise; you've saved me much time and provided much enlightenment, folks. You should be proud! On Mon, 26 Oct 1998, Mark Willis wrote: > Anyone ever do temperature sensing in a vat of molten metal? I've been interested in doing this for some time, but from a somewhat less precision, hobbyist standpoint. (If you peek into rec.crafts.metalworking, you'll see that ther are more than a few nuts like myself who cast aluminum or even brass in backyard foundries. A few of the more advanced lunatics even cast iron, but we'll not go there...) In any case, I'd be glad to give you some of my thoughts and ideas, and I'd greatly appreciate it if you might be able to report back with any interesting results. I'll be working them through eventually, but you'll easily be able to get to them before I do... > (This'll be a ceramic-lined steel vessel.) > > I'm thinking the consensus for temperature control and sensing here > was a nice thermocouple, IIRC, but what about at as high temperatures as > say 2600 F or thereabouts? (Thermal remote sensing of the top surface > might well work here, I need to learn more.) There are metal-pairs that work at this temperature. A web search for a company called Omega would be very helpful. A phone call to them will score you a pile of catalogs/application notes that's some 6 inches thick, and will tell you more than you ever wanted to know about thermocouples, as well as strain gauges, pressure sensors, platinum RTD sensors, and a host of other products that Omega makes. As you search through the different thermocouple types, be aware that even if you select metals with melting points far above the temperature of the melt, they may be soluble in the melt at the working temperature. For example, while copper won't melt at the temperature normally used for casting aluminum, it will dissolve readily enough in a ladle of molten Al. (In fact, dissolving a little copper makes the melt flow better through the mold and conform better, and we sometimes do it on purpose.) The point of this is that a thermocouple dipped into the melt may not only dissolve on you, but depending on the application, you may contaminate the melt by doing so. (I note your mention of a ceramic-lined vessel, which may possibly be used to avoid iron contamination in the melt.) Omega also sells optical pyrometers, which measure the temperature from the samples's black body radiation, in a totally non-contact measurement. Unless you're planning to make sensing systems by the dozens, Omega may be able to sell you a complete solution. The obvious disadvantage is that you won't get to use a PIC, but if your schedule is agressive enough, that may be OK. {Quote hidden}> I've searched my piclist archives on this machine, no joy, must've > lost that set of messages describing the sensors when I lost a bunch of > mail in August. I'm open to whatever -works- well & is reliable, exact > +- 1 degree F temperature sensing isn't mandatory, but +-25 degrees F or > so would be nice (I think.) Lots of thermal mass here <G> > > This'll be a future PIC project (December or January) in a factory > setting, but I want to start getting a clue now in my spare time... > Idea is to warm a batch of metal up (ramping the temperature up nicely > without wasting fuel etc.) and then hold it in controlled temperature > conditions of whatever profile for a set length of time, telling a host > PC what's going on (the PIC's the sensor front end, the PC will handle > announcing "When it's Soup" and so on <G>) and controlling things from > the host PC, probably, for safety. I still "don't know what I need to > know" on this all. But I haven't signed an NDA either yet <G> > > Anyone used super high temperature sensors of any kind in this sort of > project? A commercial sensor would be wonderful, this is a "Need it to > work right forever" type of application (so I may use redundant sensors > & have the PIC tell me when one fails, and plan to replace sensors > periodically.) May be talking a water-cooled PIC module, with > optoisolated outputs & some severe power de-spiking... I don't think > any Dallas chips like the DS1287 (?) will work here! <G> Thermocouples can be run on long, long, "extension cords", also made of the same thermocouple metals. At least your sensing circuitry should be able to be kept in a cooler location. An engineer I've known who was involved in the design of nuclear power plants told me that this was done commonly to measure temperatures in the "hot" areas. I beleive runs of hundreds of feet were workable. As far as making the sensor last a long time, I've had an idea that may or may not work, but seems worth investigating. I've used many thermocouples in the past, but never at a temperature where the sensor was warm enough to glow by its own (visible) light, as yours would do. I know that even at cooler temperatures, one problem is that the thermocouple junction occasionally comes apart. If your thermocouple might dissolve a little each usage, this problem could become severe. When the sample to be measured is molten, it may be possible to do something tricky. Here's the gist of the idea. The Omega catalog's thermocouple tutorial mentions that if we have junctions of three metals, A, B, and C, all at the same temperature, then a pair of junctions in series, --AB----BC-- will produce the same voltage as a single junction, --AC--. They give this a name like, IIRC, the law of dissimilar metals, or something. This name is very vague in my memory, but the concept is real. In any case, if your thermocouple metals are A and C, and the melt is metal B, then you may be able to dip in two separate bars of metal to form your thermocouple, and let the melt join the two and complete the junctions. The bars could be fairly hefty, and if you do the dip only long enough for the temperature to equilibrate, you might not have to sacrifice too much dissolved thermocouple to each dip. But thick bars might be able to last a long time without being replaced. Of course, this won't work until until the solid-to-liquid phase change has taken place, and the bars can actually be inserted in the metal. If the folks you're doing this for can set you up with a small test furnace that can melt a quart or so of metal, it might not be too hard to get some data. Melt up a cruciblefull, shut off the furnace, dip the (thoroughly, thoroughly, very *very* dried out and free-of-moisture) bars, and watch the voltage as it decreases over time. There should be a lengthy "leveling out" of the voltage during the phase change, then a decrease again as the solid metal drops in temperature. If you can measure a repeatable voltage at the leveling out point,I'd be fairly confident you could measure the other temperatures fairly repeatably, since the melt will be mostly liquid during the early part of the phase change. The solid metal will be on the crucible wall, while the laws of physics hold the liquid part in the center (where your sensor is) at a very, very constant and repeatable temperature for you. The surface is likely to solidify early too, so a ceramic sleeve on the bars that forces the contact area to be in the interior of the melt might be helpful. One small source of error will be the additional thermocouples you create by hooking up the sensor thermocouple to your instrumentation. When using a thermocouple to measure mere hundreds of degrees, this error is significant, and must be accounted for (the app notes mention this under the name "cold junction compensation"). At the temperatures you're working with, it may not be a problem at all, depending on the ultimate precision you need. {Quote hidden}> I probably will just do subcontracting on some of this sort of > package, too, as this hits (It's one of those NDA contracts, but if I > can subcontract everything to get everything done I think they'll be > happy, so long as I end up with PCB data, source code for the PICs, etc. > so I can maintain things for them. Looks like these folks may actually > launch, finally, after years on my back burner getting things > together... I won't have time to do it all quickly, myself. I may have > to do some program re-working as they may want all the gory details kept > to myself, for all I know. They're a little worried by some past > problems they've had!) > > I'm thinking of just doing all their various code on 17C family chips, > but maybe I'll try to fit everything on 16F84's (so much to learn...) > Thinking about it <G> Sounds like a fascinating problem to solve! Let us know how it turns out. Wes B. ***************************************************************************** * wesbKILLspamspectra.net * "It's a magical world, Hobbes ol' buddy... * * * ..Let's go exploring." - Calvin * *****************************************************************************

1998\10\27@220410 by Mark Willis

Wes A Brzozowski wrote: > <Snipped a lot throughout> > company called Omega would be very helpful. A phone call to them will > ... Great! I'll ask <G> > ... they may be soluble in the melt at the working temperature. I know (Part of the problem I'm having is that I know little of what they're doing, what I do know isn't in accordance with what they're saying, and as we all know, adding more metals to a mix usually decreases the melting temperature of all of them (as they dissolve in each other, sorta!) Trying not to bore the whole list with a discussion of how much confusion they have me in, here <G> > The point of this is that a thermocouple dipped into the melt may not > only dissolve on you, but depending on the application, you may > contaminate the melt by doing so. (I note your mention of a ceramic-lined > vessel, which may possibly be used to avoid iron contamination in the > melt.) They mention something about melting the steel vessel, as well, but with external heating, I think it's a "We'll tell you the truth when we get to where you've signed the NDA agreement" thing. I'm focusing on learning facts that'll be useful for the project, once they do tell me what I "need to know" etc... > Unless you're planning to make sensing systems by the dozens, Omega may > be able to sell you a complete solution. The obvious disadvantage is that > you won't get to use a PIC, but if your schedule is agressive enough, that > may be OK. So long as it interfaces to the rest of the system & isn't $3 billion, apiece, I can deal with it <G> Not home use, here! > Thermocouples can be run on long, long, "extension cords", also made of > the same thermocouple metals. ... Yep <G> Those parts I remember (I may series connect a bunch of type S thermocouples myself, to raise the voltage, which is good.) > But thick bars might be able to last a long time without being replaced. > Of course, this won't work until until the solid-to-liquid phase change > has taken place, and the bars can actually be inserted in the metal. I'm pretty sure they want to control the whole batch from initial heating to final pour, but sneaky idea, and that'd sure be rugged, provided you used thick bars (cast them into the ceramic covering the steel vessel? Dunno <G>) > Sounds like a fascinating problem to solve! Let us know how it turns out. > > Wes B. Interesting ideas; I can deal with just about anything that works for them. I'll know much more, and more of what I can & cannot talk about, later <G> (Right now I'm learning what I still remember & what I need to learn more about <G>) Mark, .....mwillisKILLspam.....nwlink.com

1998\10\28@071728 by Tom Handley

>On Mon, 26 Oct 1998, Mark Willis wrote: > > Anyone ever do temperature sensing in a vat of molten metal? Mark, check Omega's "The Temperature Handbook". For more info and to order their (normally free) series of handbooks, contact: http://www.omega.com/ - Tom

1998\10\28@130358 by Wes A Brzozowski

On Tue, 27 Oct 1998, Mark Willis wrote: > I know (Part of the problem I'm having is that I know little of what > they're doing, what I do know isn't in accordance with what they're > saying, and as we all know, adding more metals to a mix usually > decreases the melting temperature of all of them (as they dissolve in > each other, sorta!) Trying not to bore the whole list with a discussion I don't think there would be anywhere nearly enough dissolving to make a noticeable change in melting points; the problem is that one metal in the solid phase can be soluble in another metal at the liquid phase. Melting points are not the issue; I can dissolve sugar in water, without any interest in the melting point of sugar. This situation is similar in many ways. > Yep <G> Those parts I remember (I may series connect a bunch of type > S thermocouples myself, to raise the voltage, which is good.) One little problem there... when you connect thermocouples in series, you generate a whole lot of reverse thermocouples in the process of connecting them. These will mostly counteract the thermocouples you're interested in, and unless you can keep the unwanted thermocouples at a very controlled temperature, they'll add uncertainty to your measurement. Still, it's a neat problem to solve. Be sure to tell us all you can right before you sign that NDA. Best of luck, & take care... Wes B. ***************************************************************************** * EraseMEwesbspam_OUTTakeThisOuTspectra.net * "It's a magical world, Hobbes ol' buddy... * * * ..Let's go exploring." - Calvin * *****************************************************************************

1998\10\28@171605 by Mark Willis

Wes A Brzozowski wrote: > > > Yep <G> Those parts I remember (I may series connect a bunch of type > > S thermocouples myself, to raise the voltage, which is good.) > > One little problem there... when you connect thermocouples in series, you > generate a whole lot of reverse thermocouples in the process of > connecting them. These will mostly counteract the thermocouples you're > interested in, and unless you can keep the unwanted thermocouples at a > very controlled temperature, they'll add uncertainty to your measurement. Other end of the thermocouples are cooled to a known, stable, measured temperature, I thought that went without saying <G> (Chilled water, probably?) That or I use an amplifier to raise the single thermocouples' voltage to near the A/D's max. range, if needed (I need to go stare at spec sheets for type 'S' thermocouples <G>) Doesn't do to have an A/D converter that gives 8 bits 0-5VDC and then feed it a signal that only changes from 0.45VDC to 0.78VDC or something, giving 1 bit ADC resolution with 0 bits accuracy <G> (and most thermocouples will source loads of current, and are quite low impedance IIRC, but do not source much voltage at all.) So an Op Amp insulating the couple(s) from the melt are both probables unless I go a different route (I'll have a couple months to learn here <G>) Mark, mwillisspam_OUTnwlink.com

1998\10\29@142619 by Wes A Brzozowski

On Wed, 28 Oct 1998, Mark Willis wrote: > Wes A Brzozowski wrote: > > > > > Yep <G> Those parts I remember (I may series connect a bunch of type > > > S thermocouples myself, to raise the voltage, which is good.) > > > > One little problem there... when you connect thermocouples in series, you > > generate a whole lot of reverse thermocouples in the process of > > connecting them. These will mostly counteract the thermocouples you're > > interested in, and unless you can keep the unwanted thermocouples at a > > very controlled temperature, they'll add uncertainty to your measurement. > > Other end of the thermocouples are cooled to a known, stable, measured > temperature, I thought that went without saying <G> (Chilled water, > probably?) We each have our own personal style of working, and have found various methods and procedures that have served us well. Myself, I've found that *nothing* goes without saying when talking to someone I don't know, and whose experience in the matter is unknown to me. Also, a peek at my earlier comment, which you've quoted above, contains an allusion to the controlled temperature of the unwanted thermocouples, and hints at this as a possible disadvantage, rather than a design solution. Still, you may want to investigate this first hand. More on that in a moment. It's possible that you may find otherwise, but when you consider the cost of the plumbing, the lower reliability and cost (not to mention the inconvenience!) of the numerous lengthy "thermocouple extension cords" you'll need for the distance to keep your water from boiling, let alone keep it chilled, you may start thinking instead in terms of the cost and convenience of one hot thermocouple, possibly a single remote cold junction. and a decent instrumentation amp. Count also the number of commercial temperature measurement solutions that use series thermocouples. This may prove helpful. Now, the fact that no one else uses a particular solution doesn't make it a bad thing. If your solution is novel, it may be a very good thing. But, getting back to that "personal style" thing again, I've found that it's been only a small percentage of the time that my "novel" solutions were truly novel. More often, I've found that the reason no one uses a particular solution is not because no one's tried it, but rather because they already have. Your track record may well be better than mine in this area, and that would be great, but it's still prudent to view a novel idea with suspicion and skepticism, at least until getting a little data. That's not to say that I follow my own advice in this matter nearly enough. The excitement of a new idea often has an amazing amount of momentum, and I continue to learn "the hard way"... If the data ends up looking good, then it's full speed ahead! > That or I use an amplifier to raise the single > thermocouples' voltage to near the A/D's max. range, if needed (I need Yep. Start with just a few "back of the envelope" preliminary designs, just comparing the costs. It's fairly straightforeward to estimate the amount of radiant heat you'll be dealing with at that temperature, for different distances from a certain sized melt; consider what you'll need to do to minimize the error you'll get from heat coupling into those unwanted junctions. Again, you may have an interesting solution, or not. > to go stare at spec sheets for type 'S' thermocouples <G>) Doesn't do > to have an A/D converter that gives 8 bits 0-5VDC and then feed it a > signal that only changes from 0.45VDC to 0.78VDC or something, giving 1 > bit ADC resolution with 0 bits accuracy <G> (and most thermocouples will > source loads of current, and are quite low impedance IIRC, but do not > source much voltage at all.) So an Op Amp insulating the couple(s) from Think "low offset instrumentation amp". > the melt are both probables unless I go a different route (I'll have a > couple months to learn here <G>) Taking the time to learn is an excellent idea. I mentioned taking some data. If you're a fairly "hands on" kinda guy, you might want to peek through rec.crafts.metalworking, under a series of threads with titles like "coffee can foundry". You'll find you can melt small amounts of aluminum for almost no cost, and you could use such a setup to get a vague feel for the kind of environment you'll need to deal with. Don't try to actually pick up a melt using this apparatus, no matter what they say, though. While fascinating, I do think it's quite crazy. Just produce a melt, perhaps try a measurement or two, (MAKE SURE THAT YOUR SENSOR IS **ABSOLUTELY** DRY!!! BAKE THOROUGHLY BEFOREHAND!) and let it cool, undisturbed. This will of course be a very lightweight version of what you'll need to handle. This melt will merely radiate enough heat that it's difficult to get near. Your target melt will probably radiate enough heat (and light!) that it may be difficult even to view from some distance. You'll have to fill us in on the details, there. I'd be quite interested, at least. Again, you may see things very different from myself, but such an experience may just turn you into a minimalist when it comes to deciding how much "stuff" should be placed in the vicinity of the melt. If so, you'll want to know as early as possible. In either case, if you're the type of person who relishes seeing and learning new things, this should be a real eye opener. Best of luck & be careful out there... Wes B. ***************************************************************************** * @spam@wesbKILLspamspectra.net * "It's a magical world, Hobbes ol' buddy... * * * ..Let's go exploring." - Calvin * *****************************************************************************

1998\10\30@015816 by Mark Willis

Wes A Brzozowski wrote: {Quote hidden}> > On Wed, 28 Oct 1998, Mark Willis wrote: > > > Wes A Brzozowski wrote: > > > > > > > Yep <G> Those parts I remember (I may series connect a bunch of type > > > > S thermocouples myself, to raise the voltage, which is good.) > > > > > > One little problem there... when you connect thermocouples in series, you > > > generate a whole lot of reverse thermocouples in the process of > > > connecting them. These will mostly counteract the thermocouples you're > > > interested in, and unless you can keep the unwanted thermocouples at a > > > very controlled temperature, they'll add uncertainty to your measurement. > > > > Other end of the thermocouples are cooled to a known, stable, measured > > temperature, I thought that went without saying <G> (Chilled water, > > probably?) > > We each have our own personal style of working, and have found various > methods and procedures that have served us well. Myself, I've found that > *nothing* goes without saying when talking to someone I don't know, > and whose experience in the matter is unknown to me. Also, a peek at my > earlier comment, which you've quoted above, contains an allusion to the > controlled temperature of the unwanted thermocouples, and hints at this > as a possible disadvantage, rather than a design solution. Still, you > may want to investigate this first hand. More on that in a moment. > > It's possible that you may find otherwise, but when you consider the cost > of the plumbing, the lower reliability and cost (not to mention the > inconvenience!) of the numerous lengthy "thermocouple extension cords" > you'll need for the distance to keep your water from boiling, let alone > keep it chilled, you may start thinking instead in terms of the cost > and convenience of one hot thermocouple, possibly a single remote cold > junction. and a decent instrumentation amp. Count also the number of > commercial temperature measurement solutions that use series > thermocouples. This may prove helpful. > > Now, the fact that no one else uses a particular solution doesn't make > it a bad thing. If your solution is novel, it may be a very good thing. > But, getting back to that "personal style" thing again, I've found that > it's been only a small percentage of the time that my "novel" solutions > were truly novel. More often, I've found that the reason no one uses a > particular solution is not because no one's tried it, but rather because > they already have. Your track record may well be better than mine in this > area, and that would be great, but it's still prudent to view a novel > idea with suspicion and skepticism, at least until getting a little data. > That's not to say that I follow my own advice in this matter nearly > enough. The excitement of a new idea often has an amazing amount of > momentum, and I continue to learn "the hard way"... > > If the data ends up looking good, then it's full speed ahead! > > > That or I use an amplifier to raise the single > > thermocouples' voltage to near the A/D's max. range, if needed (I need > > Yep. Start with just a few "back of the envelope" preliminary designs, > just comparing the costs. It's fairly straightforeward to estimate the > amount of radiant heat you'll be dealing with at that temperature, for > different distances from a certain sized melt; consider what you'll need > to do to minimize the error you'll get from heat coupling into those > unwanted junctions. Again, you may have an interesting solution, or not. > > > to go stare at spec sheets for type 'S' thermocouples <G>) Doesn't do > > to have an A/D converter that gives 8 bits 0-5VDC and then feed it a > > signal that only changes from 0.45VDC to 0.78VDC or something, giving 1 > > bit ADC resolution with 0 bits accuracy <G> (and most thermocouples will > > source loads of current, and are quite low impedance IIRC, but do not > > source much voltage at all.) So an Op Amp insulating the couple(s) from > > Think "low offset instrumentation amp". > > > the melt are both probables unless I go a different route (I'll have a > > couple months to learn here <G>) > > Taking the time to learn is an excellent idea. I mentioned taking some > data. If you're a fairly "hands on" kinda guy, you might want to peek > through rec.crafts.metalworking, under a series of threads with titles > like "coffee can foundry". You'll find you can melt small amounts of > aluminum for almost no cost, and you could use such a setup to get a > vague feel for the kind of environment you'll need to deal with. Don't > try to actually pick up a melt using this apparatus, no matter what > they say, though. While fascinating, I do think it's quite crazy. > Just produce a melt, perhaps try a measurement or two, (MAKE SURE THAT > YOUR SENSOR IS **ABSOLUTELY** DRY!!! BAKE THOROUGHLY BEFOREHAND!) and let > it cool, undisturbed. This will of course be a very lightweight version > of what you'll need to handle. This melt will merely radiate enough heat > that it's difficult to get near. Your target melt will probably radiate > enough heat (and light!) that it may be difficult even to view from some > distance. You'll have to fill us in on the details, there. I'd be quite > interested, at least. > > Again, you may see things very different from myself, but such an > experience may just turn you into a minimalist when it comes to > deciding how much "stuff" should be placed in the vicinity of the melt. > If so, you'll want to know as early as possible. > > In either case, if you're the type of person who relishes seeing and > learning new things, this should be a real eye opener. > > Best of luck & be careful out there... > > Wes B. (Well, hoist on my own petard, I guess - I'm usually only that "goes without saying" way on safety stuff <G>) I've cast lead & so on so I do know about "pre-heating" <G> I'm looking forwards to this one, I'll be challenged <G> One person mentioned (off list) that some folks just run things on a "fourier" manner (i.e. knowing mass, BTU's in, and desired temperature, just cook things for yea long "because it's always worked before", & then turn off the heat.) Too low-tech for me - BUT, I guess I'll solve the problems the client *wants* me to solve, whatever they are <G> Mark, KILLspammwillisKILLspamnwlink.com

'High temperature sensor'
1998\11\01@155552 by Dennis Plunkett

At 22:56 29/10/98 -0800, you wrote: {Quote hidden}>Wes A Brzozowski wrote: >> >> On Wed, 28 Oct 1998, Mark Willis wrote: >> >> > Wes A Brzozowski wrote: >> > > >> > > > Yep <G> Those parts I remember (I may series connect a bunch of type >> > > > S thermocouples myself, to raise the voltage, which is good.) >> > > >> > > One little problem there... when you connect thermocouples in series, you >> > > generate a whole lot of reverse thermocouples in the process of >> > > connecting them. These will mostly counteract the thermocouples you're >> > > interested in, and unless you can keep the unwanted thermocouples at a >> > > very controlled temperature, they'll add uncertainty to your measurement. >> > >> > Other end of the thermocouples are cooled to a known, stable, measured >> > temperature, I thought that went without saying <G> (Chilled water, >> > probably?) >> >> We each have our own personal style of working, and have found various >> methods and procedures that have served us well. Myself, I've found that >> *nothing* goes without saying when talking to someone I don't know, >> and whose experience in the matter is unknown to me. Also, a peek at my >> earlier comment, which you've quoted above, contains an allusion to the >> controlled temperature of the unwanted thermocouples, and hints at this >> as a possible disadvantage, rather than a design solution. Still, you >> may want to investigate this first hand. More on that in a moment. >> >> It's possible that you may find otherwise, but when you consider the cost >> of the plumbing, the lower reliability and cost (not to mention the >> inconvenience!) of the numerous lengthy "thermocouple extension cords" >> you'll need for the distance to keep your water from boiling, let alone >> keep it chilled, you may start thinking instead in terms of the cost >> and convenience of one hot thermocouple, possibly a single remote cold >> junction. and a decent instrumentation amp. Count also the number of >> commercial temperature measurement solutions that use series >> thermocouples. This may prove helpful. >> >> Now, the fact that no one else uses a particular solution doesn't make >> it a bad thing. If your solution is novel, it may be a very good thing. >> But, getting back to that "personal style" thing again, I've found that >> it's been only a small percentage of the time that my "novel" solutions >> were truly novel. More often, I've found that the reason no one uses a >> particular solution is not because no one's tried it, but rather because >> they already have. Your track record may well be better than mine in this >> area, and that would be great, but it's still prudent to view a novel >> idea with suspicion and skepticism, at least until getting a little data. >> That's not to say that I follow my own advice in this matter nearly >> enough. The excitement of a new idea often has an amazing amount of >> momentum, and I continue to learn "the hard way"... >> >> If the data ends up looking good, then it's full speed ahead! >> >> > That or I use an amplifier to raise the single >> > thermocouples' voltage to near the A/D's max. range, if needed (I need >> >> Yep. Start with just a few "back of the envelope" preliminary designs, >> just comparing the costs. It's fairly straightforeward to estimate the >> amount of radiant heat you'll be dealing with at that temperature, for >> different distances from a certain sized melt; consider what you'll need >> to do to minimize the error you'll get from heat coupling into those >> unwanted junctions. Again, you may have an interesting solution, or not. >> >> > to go stare at spec sheets for type 'S' thermocouples <G>) Doesn't do >> > to have an A/D converter that gives 8 bits 0-5VDC and then feed it a >> > signal that only changes from 0.45VDC to 0.78VDC or something, giving 1 >> > bit ADC resolution with 0 bits accuracy <G> (and most thermocouples will >> > source loads of current, and are quite low impedance IIRC, but do not >> > source much voltage at all.) So an Op Amp insulating the couple(s) from >> >> Think "low offset instrumentation amp". >> >> > the melt are both probables unless I go a different route (I'll have a >> > couple months to learn here <G>) >> >> Taking the time to learn is an excellent idea. I mentioned taking some >> data. If you're a fairly "hands on" kinda guy, you might want to peek >> through rec.crafts.metalworking, under a series of threads with titles >> like "coffee can foundry". You'll find you can melt small amounts of >> aluminum for almost no cost, and you could use such a setup to get a >> vague feel for the kind of environment you'll need to deal with. Don't >> try to actually pick up a melt using this apparatus, no matter what >> they say, though. While fascinating, I do think it's quite crazy. >> Just produce a melt, perhaps try a measurement or two, (MAKE SURE THAT >> YOUR SENSOR IS **ABSOLUTELY** DRY!!! BAKE THOROUGHLY BEFOREHAND!) and let >> it cool, undisturbed. This will of course be a very lightweight version >> of what you'll need to handle. This melt will merely radiate enough heat >> that it's difficult to get near. Your target melt will probably radiate >> enough heat (and light!) that it may be difficult even to view from some >> distance. You'll have to fill us in on the details, there. I'd be quite >> interested, at least. >> >> Again, you may see things very different from myself, but such an >> experience may just turn you into a minimalist when it comes to >> deciding how much "stuff" should be placed in the vicinity of the melt. >> If so, you'll want to know as early as possible. >> >> In either case, if you're the type of person who relishes seeing and >> learning new things, this should be a real eye opener. >> >> Best of luck & be careful out there... >> >> Wes B. > > (Well, hoist on my own petard, I guess - I'm usually only that "goes >without saying" way on safety stuff <G>) > > I've cast lead & so on so I do know about "pre-heating" <G> I'm >looking forwards to this one, I'll be challenged <G> > > One person mentioned (off list) that some folks just run things on a >"fourier" manner (i.e. knowing mass, BTU's in, and desired temperature, >just cook things for yea long "because it's always worked before", & >then turn off the heat.) Too low-tech for me - BUT, I guess I'll solve >the problems the client *wants* me to solve, whatever they are <G> > > Mark, RemoveMEmwillisTakeThisOuTnwlink.com > > Can't do that if you want to make aluminum! Dennis

1998\11\01@183657 by Mark Willis

Dennis Plunkett wrote: > > One person mentioned (off list) that some folks just run things on a > >"fourier" manner (i.e. knowing mass, BTU's in, and desired temperature, > >just cook things for yea long "because it's always worked before", & > >then turn off the heat.) Too low-tech for me - BUT, I guess I'll solve > >the problems the client *wants* me to solve, whatever they are <G> > > > > Mark, spamBeGonemwillisspamBeGonenwlink.com > > Can't do that if you want to make aluminum! > > Dennis Pretty sure it's not that (Aluminum is made almost exclusively with electrical arc furnaces, there'd be no chance of considering use of other heating methods, far as I know <G> Mark

1998\11\01@201305 by Dennis Plunkett

At 15:35 1/11/98 -0800, you wrote: {Quote hidden}>Dennis Plunkett wrote: >> > One person mentioned (off list) that some folks just run things on a >> >"fourier" manner (i.e. knowing mass, BTU's in, and desired temperature, >> >just cook things for yea long "because it's always worked before", & >> >then turn off the heat.) Too low-tech for me - BUT, I guess I'll solve >> >the problems the client *wants* me to solve, whatever they are <G> >> > >> > Mark, TakeThisOuTmwillisEraseMEspam_OUTnwlink.com >> >> Can't do that if you want to make aluminum! >> >> Dennis > > Pretty sure it's not that (Aluminum is made almost exclusively with >electrical arc furnaces, there'd be no chance of considering use of >other heating methods, far as I know <G> > > Mark > > Humm, I thought that temperature was very very very important to making of the aluminum, else you would end up with a blow out and / or a multain mass of mess in the cooker. Nothing like producing 6 months of stuff that's not even good enough for muck metal castings. As for the measure and mix, yes he was quite right, that's how older foundries (Like them of the 60s in Pittsburgh, and the current Russian ones) do it. However the new foundries like Japan and the Asian countries (Notice that I left off Brazil!) do measure the temp for best continuous results. Also the actual physical content of the starter (Castoff) would need to be tightly controlled if no temperature measurements are made, just a bit to much cadmium and well rubbish in rubbish out (Well better stop) I suggest that you get in contact with G.E. (Or someone there) As they win contracts to install mills and that, I think they are currently doing the modeling for a mill in the states right at this very minute. Dennis

1998\11\02@001414 by John Griessen

My father had some experience with aluminum making and blowouts. I think aluminum plants do part of the process under a molten bath of salt that is mostly caustic. The molten aluminum is one electrode with volts applied to drive a chemical reaction to get more metal out of the alumina oxide. He described a time when an employee got splattered in the eyes with caustic when fixing something to do with an electrode, and he had to neutralize it with vinegar and drive him to the hospital. At room temperature up to boiling with water, caustic dissolves aluminum quickly, so I imagine anything that disrupts steadiness is bad bad bad for making aluminum since it could react back to oxide, or lead to blowouts, or eat electrodes too fast. For plain recycling, natural gas or propane furnaces can be used to melt aluminum and there are fluxes specialized for good results that way. John Griessen CIBOLO Metal Works http://www.aus-etc.com/~cibolo/ 512 451-8207 -----Original Message----- Humm, I thought that temperature was very very very important to making of the aluminum, else you would end up with a blow out and / or a multain mass of mess in the cooker.

More... (looser matching)
- Last day of these posts
- In 1998 , 1999 only
- Today
- New search...