Post by thread:[PIC]: Measuring weight

I'm thinking about ways to measure weight, on the cheap. The idea is to have a rack in a cooking oven, with one rail of the rack support supported on a spring, to measure the weight of the object being cooked ( never mind why we are interested in that... long story) So the problem of measuring weight comes down to measuring the position of this rack, as it rides up or down on it's spring. Now we can't tilt too much, or we'll dump Mrs. Smith's pie all over the inside of the oven. 1/4"[6mm] or 3/8" [9mm] would be about the maximum we could tolerate, for maybe a 5 or 6 LB [2.5 kg] food item. I'd like to be able to measure to 1/4 Lb accuracy So we are trying to measure how far this oven rack drops after the item is placed on it. Here's what I've thought about so far: 1. Actuate a lever, which cranks a small potentiometer. Probs: Pots travel round in a circle, you might not get more than 45 degrees of rotation out of this scheme, limiting your abitlity to discriminate small weights. Small pots tend to have a lot of hysteresis, and this would show up as error if you are not moving the pot many degrees. Any slop and tolerance in your original mechanism gets multiplied by the lever arm ration (assuming mechanical disadvantage) and results in more errors and hysteresis at your output. 2. Actuate a lever, which cranks a liner pot. problem: Linear pots are kinda expensive. Cheap ones tend to skip and have dead spots. (i.e. 1,2,3,4,5,0,0,0,0,7,8,9,10 like that) 3. Ala' digital micrometers: print a fine gauge grid of lines on an otherwise clear piece of plastic. Place it between an optical switch (led/phototransistor) shielded through a small slit. Count the number of light/dark variations as it moves. Problems: Where is zero? Place your item in the oven, then plug it in - how does this sensor know where zero was? Did Rube Goldberg think of this method? 4. Actuate a low-pitch screw thread which rotates a potentiometer, say the whole thread turns 270 degrees for 3/8" travel, thus driving the pot through it's range. Hmmm. Gut feeling says this would hang up somehow. This might work OK for longer travel - 5-6 inches. 5. Run a springy shorting bar across a series of fine-pitch silver plated tracks on a printed circuit board, and measure the position of the shorting bar by various methods. (port input, shorting out a series of resistors, keypad-like array, etc. etc. ) So far I can't think of too many problems wioth this, except I suspect it might be expensive to implement and the tracks might wear out. Any other ideas? -- Lawrence Lile -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email spam_OUTlistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

Lawrence, What about some sort of strain guage setup? Regards, Jim On Mon, 30 April 2001, Lawrence Lile wrote: {Quote hidden}> > I'm thinking about ways to measure weight, on the cheap. The idea is to > have a rack in a cooking oven, with one rail of the rack support supported > on a spring, to measure the weight of the object being cooked ( never mind > why we are interested in that... long story) > > So the problem of measuring weight comes down to measuring the position of > this rack, as it rides up or down on it's spring. > > Now we can't tilt too much, or we'll dump Mrs. Smith's pie all over the > inside of the oven. 1/4"[6mm] or 3/8" [9mm] would be about the maximum we > could tolerate, for maybe a 5 or 6 LB [2.5 kg] food item. I'd like to be > able to measure to 1/4 Lb accuracy > > So we are trying to measure how far this oven rack drops after the item is > placed on it. Here's what I've thought about so far: > > 1. Actuate a lever, which cranks a small potentiometer. Probs: Pots > travel round in a circle, you might not get more than 45 degrees of rotation > out of this scheme, limiting your abitlity to discriminate small weights. > Small pots tend to have a lot of hysteresis, and this would show up as error > if you are not moving the pot many degrees. Any slop and tolerance in your > original mechanism gets multiplied by the lever arm ration (assuming > mechanical disadvantage) and results in more errors and hysteresis at your > output. > > 2. Actuate a lever, which cranks a liner pot. problem: Linear pots are > kinda expensive. Cheap ones tend to skip and have dead spots. (i.e. > 1,2,3,4,5,0,0,0,0,7,8,9,10 like that) > > 3. Ala' digital micrometers: print a fine gauge grid of lines on an > otherwise clear piece of plastic. Place it between an optical switch > (led/phototransistor) shielded through a small slit. Count the number of > light/dark variations as it moves. Problems: Where is zero? Place your > item in the oven, then plug it in - how does this sensor know where zero > was? Did Rube Goldberg think of this method? > > 4. Actuate a low-pitch screw thread which rotates a potentiometer, say the > whole thread turns 270 degrees for 3/8" travel, thus driving the pot through > it's range. Hmmm. Gut feeling says this would hang up somehow. This > might work OK for longer travel - 5-6 inches. > > 5. Run a springy shorting bar across a series of fine-pitch silver plated > tracks on a printed circuit board, and measure the position of the shorting > bar by various methods. (port input, shorting out a series of resistors, > keypad-like array, etc. etc. ) So far I can't think of too many problems > wioth this, except I suspect it might be expensive to implement and the > tracks might wear out. > > Any other ideas? > > -- Lawrence Lile > > -- > http://www.piclist.com#nomail Going offline? Don't AutoReply us! > email .....listservKILLspam@spam@mitvma.mit.edu with SET PICList DIGEST in the body jimKILLspamjpes.com -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email .....listservKILLspam.....mitvma.mit.edu with SET PICList DIGEST in the body

From: James Paul [EraseMEjimspam_OUTTakeThisOuTJPES.COM] > Lawrence, > > What about some sort of strain guage setup? or a capacitive force/pressure sensor. it sounds like you already have a spring, you're just trying to measure how muchit moves, right? make a capacitor with one side fixed in place to the oven, the other to the rack. separated by the spring, of course. As the spring compresses, the plates move closer together, and the capacitance changes. I did this a while ago, except for 2500 PSI. It used a CMOS version of the 555 to generate a 50% square wave whose frequency was dependent on the capacitor. one side of the capacitor plate was made out of PCB track on the same board as the 555. The other plate was was also made from a PCB trace, but on a separate board. The plates were mounted in an aluminum housing, separated by a stack of belville (i never remember how to spell that) washers. This system only requires very small movement for capacitance to change. The change in frequency is easy to measure with a PIC or other micro. If this is a one-off, this should work for you. If it's for production, you're going to have all sorts of fun with mfg. tolerance stackups. Plan on having to calibrate each unit :( For your oven application, you probably want to charge/discharge the cap instead. Or at least have the ciruit portion outside the oven :) Duh! I guess I didn't have to say that... -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listservspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body

On Mon, 30 Apr 2001 12:46:28 -0700, O'Reilly John E NORC wrote: >I like #5. > >A variation on this idea could be to use a binary "barcode" made by having a >clear piece of glass or plastic with clear and opaque areas. Use LEDs on >one side and phototransistors (or whatever) on the other side. All clear = >0 - 1/4 lb, 00001 = 1/4 - 1/2 lb, etc. You wouldn't be touching the >barcode, so it wouldn't wear out. Just watch out for the pie splattering >apples all over it! > And yet another auto-zeroing variation... Use the same technique as above, but place binary codes radially on a glass disk which is placed on a shaft. A wire could be wound around the shaft and turn it as the rack moves up an down. Auto-zeroing is accomplished by reading the binary code from the disk before an item is placed on it. The weight is calculated by using the distance between (number of binary codes between) the initially read value and that after the weight is added. A high pass filter could be used to detect movement when weight is added or removed and the amount if weight is obtained by comparing the binary codes before and after the weight change. Since the disk is driven by a wire, the disk will not get apples splatter all over it. I wire would need to be chosen that would minimize the amount of expansion and contraction induced by the heating and cooling of the oven. Dan -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email @spam@listservKILLspammitvma.mit.edu with SET PICList DIGEST in the body

No, mount LEDs horizontally, move the interruptor vertically (probably with a lever to amplify the mechanical movement). Read the weight code out parallel from the PTs. On Mon, 30 Apr 2001, Eisermann, Phil [Ridg/CO] wrote: {Quote hidden}> Good idea, except for the 1/4" of movement restriction. Might be tough to > get all those LEDs in there. Unless you have a lever that amplifies the > movement. > > What about having a magnet on your moveable arm, and reading the position of > the magnet with a Hall Sensor? Nothing to wear out here, either. And you can > get pretty good results with those things over short distances (think > millimeters) > > > > {Original Message removed}

Hmmm - never heard of this idea before. Yeah, the weight measured will depend on whether the food item is placed in the center of the rack, or on one end or the other. We're using this weight to give an early prediction of another number we can measure preceisely later, so this is just to get us in the ballpark, so to speak. If the whole thing sufferes from a 25% error due to the consumer placing the item wrong, that's OK. I just don't want my sensor to introduce any errors bigger than, say, 5%. on top of that How come you know fiberoptic cables do this (I suppose the hard way - sounds like an interesting story, Bob..) -- Lawrence {Original Message removed}

Nice idea... say 5 LED/phototransistor pairs with a strip of glass or plastic, that gives you 32 combinations... up to 8# in quarter-pound increments. Spring-load the rack and connect it to the interruptor. For that matter, forget the glass or plastic, use a thin piece of aluminum with holes in it and forget about shock and heat damage. Dale On Mon, 30 Apr 2001, O'Reilly John E NORC wrote: {Quote hidden}> I like #5. > > A variation on this idea could be to use a binary "barcode" made by having a > clear piece of glass or plastic with clear and opaque areas. Use LEDs on > one side and phototransistors (or whatever) on the other side. All clear = > 0 - 1/4 lb, 00001 = 1/4 - 1/2 lb, etc. You wouldn't be touching the > barcode, so it wouldn't wear out. Just watch out for the pie splattering > apples all over it! > > Just a brainstorm. > > John > > {Original Message removed}

How about suspending one side with nitinol wire and measuring the current required to pick it up off a contact? Or just heat the oven & note the temp at which it floats... regards, Jack Lawrence Lile wrote: {Quote hidden}> > Hmmm - never heard of this idea before. > > Yeah, the weight measured will depend on whether the food item is placed in > the center of the rack, or on one end or the other. We're using this weight > to give an early prediction of another number we can measure preceisely > later, so this is just to get us in the ballpark, so to speak. If the whole > thing sufferes from a 25% error due to the consumer placing the item wrong, > that's OK. I just don't want my sensor to introduce any errors bigger than, > say, 5%. on top of that > > How come you know fiberoptic cables do this (I suppose the hard way - sounds > like an interesting story, Bob..) > > -- Lawrence > > {Original Message removed}

Sorry, Lawrence, this was supposed to go to the list :) At 02:24 PM 4/30/01 -0500, you wrote: >I'm thinking about ways to measure weight, on the cheap. The idea is to >have a rack in a cooking oven, with one rail of the rack support supported >on a spring, to measure the weight of the object being cooked ( never mind >why we are interested in that... long story) Someone already caught the "distance from the hinge" part. But I've read too many posts here not to be hit with the idea of a mechanism where you try to balance the weight of the shelf. You drive the screw into a spring which pushes up on the shelf. With this setup you don't need to have much movement, only enough to know when you've passed the balance point. The shelf movement can be restricted to indicate only "too heavy" or "too light". Since you drive the screw you don't have to figure out how to amplify the shelf movement into something readable. Of course, you didn't want to use a motor. Well, free advice is worth what you pay :) Barry -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email KILLspamlistservKILLspammitvma.mit.edu with SET PICList DIGEST in the body

You guys are trying to re-invent the wheel. They used to be made out of wood... We have put men on the moon. We don't need wooden wheels anymore. (Except on certain ancient Mardi Gras floats in New Orleans) Stop trying to use all your engineering skills to enumerate the myriad ways of assembling levers, cams, gears, wheels, planes(helical or whatever), ad nauseaum ad infinitum. Use ONLY the latest and cheapest of the most complex item available on the market to drive the nail. In this case, IMHO, it would have to be a high temp strain gauge. So, don't forget the temperature in your climatically controlled environment. (The optical scale would have to be etched on metal, etc...depending on the temp, which has not been divulged). An oven rack is supported by two rails and many points along the rails. (Actually, no object rests in a centrifugal field on more than 3. (Yes, I know it's a gravity field. Lesson for the day: How do you tell the difference...) But, for all practical purposes, you can't see the multiple points. So, we will assume our oven rack is resting contentedly on it's multiple contact point rails. And of course always on only three points. Yes, some of those three points may be rather long and linear, not Euclidean at all. So let's not go there, or the engineering will get so goofy we can't stand it. Forget the two point side and let it find whatever points it will scrinch, and bend, and find, as it gets hotter. (For you geometry nuts out there, NEVER more than two. But the two points changing back and forth along the whole length, more or less, as it complains bitterly by twisting and warping in response to it's global warming). The opposite side is where we place our SINGLE strain gauge. This will require placing a bar(with VERY low temp coef., to keep the geom. nuts happy)across the rails on the opposite side. In the middle of that bar we place our strain gauge and let the rack come to rest on a three point system: The Far Side (Sorry, Gary) with it's myriad of changing "two points" as it gets hotter, wherever they are, and the gauge side supporting the rack at a single point on top of the strain gauge. Measure the voltage/resistance across the gauge before and after, the placement of the evaporating concoction(center of gravity of concoction ideally in geometric center of the rack, but if not, do the trig) and let's all go home. As for the far side, you might want to place two point supports, as far apart as possible and equal in height to the low temp rail on the gauge side... Chris Cox (The other Chris) isermann, Phil [Ridg/CO]" wrote: > > Good idea, except for the 1/4" of movement restriction. Might be tough to > get all those LEDs in there. Unless you have a lever that amplifies the > movement. > > What about having a magnet on your moveable arm, and reading the position of > the magnet with a Hall Sensor? Nothing to wear out here, either. And you can > get pretty good results with those things over short distances (think > millimeters) > > {Original Message removed}

Chris Cox wrote: > (Actually, no object rests in a centrifugal field on more than 3. The three point lesson is for non-deformable objects, most of the oven racks I know deform like crazy when respectable loads are placed on them. I think your strain gage is possibly a very good choice, but I would place one near each of the four corners of the rack so they are the ONLY possible contact points. Then regardless of where the load is placed the total weight is the sum of the loads on the four strain gages (even if only three are making contact). I also assume a thermometer is present to measure the temperature near each of the strain gages to calibrate them vs temperature. Unfortunately the original post did not specify whether this mechanism had to fit in an existing oven or the oven could be modified appropriately to handle the measuring apparatus. If the rack is removed and placed on another level within the oven, is the measurement expected there as well? Is this going into a commercial product where it will be used and abused, or a one-off for some engineering study that can be handled "gently"? Without knowing some of these answers, a true solution is unobtainable. David W. Gulley Destiny Designs -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu

>other ideas: Make your own LVDT using mains frequency (leakage from a power transformer and a standard inductor used as pickup coil with rack as moving core between them). Make your own variable capacitance sensor using a HF source (PIC clock osc), the rack as shorting armature and a third sense electrode. Optically sense rack position (many ways). Put the rack on a spring with short travel and add a single position sensor with narrow band (switch can be ok, or Hall sensor proximity sensor or variable reluctance sensor using rack proper). Make sure that the rack swings shortly when the door is closed (is there a motor that moves the rack ? You could turn it on and off briefly and measure inertia). Measure period using PIC to obtain weight. I like #2 best, it is sure to work, followed by #4. Peter -- http://www.piclist.com hint: To leave the PICList spamBeGonepiclist-unsubscribe-requestspamBeGonemitvma.mit.edu

Peter L. Peres wrote: > > The nitinol idea gave me another one. If you use some sort of bimetal > device (or even nitinol, which also works with heat, not electricity), and > a single on/off sensor then you could rely on the heat in the first > minutes to move the grate (or the sensor) and find out where the grate is. > This would take a couple of seconds after you turn on the heat. Yeah, that's what I was getting at - I like the bi-metal strip approach. Nitinol came to mind because of its neg- ative expansion tempco, and the temp range over which it exhibits elastic deformation - Possibly an advantage in the app. regards, Jack -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

You are correct of course about summing the 3 or 4 or whichever points are making contact with the rack. Since the original problem was to monitor the loss of mass during and after heating, you would only have to know the tare of the container, then weigh the whole before placing on the rack and then after placing on the rack and monitor the change during heating. The difference is a percentage of the original whole... "David W. Gulley" wrote: {Quote hidden}> > Chris Cox wrote: > > > (Actually, no object rests in a centrifugal field on more than 3. > > The three point lesson is for non-deformable objects, most of the oven > racks I know deform like crazy when respectable loads are placed on > them. > > I think your strain gage is possibly a very good choice, but I would > place one near each of the four corners of the rack so they are the ONLY > possible contact points. Then regardless of where the load is placed the > total weight is the sum of the loads on the four strain gages (even if > only three are making contact). > > I also assume a thermometer is present to measure the temperature near > each of the strain gages to calibrate them vs temperature. > > Unfortunately the original post did not specify whether this mechanism > had to fit in an existing oven or the oven could be modified > appropriately to handle the measuring apparatus. > > If the rack is removed and placed on another level within the oven, is > the measurement expected there as well? Is this going into a commercial > product where it will be used and abused, or a one-off for some > engineering study that can be handled "gently"? Without knowing some of > these answers, a true solution is unobtainable. > > David W. Gulley > Destiny Designs > > -- > http://www.piclist.com hint: To leave the PICList > piclist-unsubscribe-requestEraseME.....mitvma.mit.edu -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads