well my final project for my mechatronics class this semester was to build a scale. we pretty much had free range on what we could do and what materials we could use.
specs/limitations/goals:
must have a mechanical and electrical aspect to it
must have digital readout
must have a tare function
should weight 0-1500 grams within a 1% accuracy
must have a count function. meaning you put one thing on there (say a penny) have it weigh it, then you put more on, and it tells you how many are there
must read out weight in grams/kilograms and lbs/oz
my professor didnt really expect people to be that accurate, it was just a goal.
basically what our initial idea was get a force sensor that outputted a certain voltage for what weight is on it. so we started on the program based on that. but when we got the sensor in, it had 4 pins on it. i searched the manufactures website and found out that there were 2 outputs. one outputted a voltage(V_2), and the other ouputted a voltage(V_4), and the total output voltage (V_o) was the difference between the two.
V_o = (V_1)-(V_2)
heres the sensor we used
http://content.honeywell.com/sensing/prodi...ce/008028_0.pdf
so basically i have no clue on how to get one ouput that was equal to the difference of the two. luckily the manufacturer had a wiring diagram on how to do this.
here it is.
we used two 9 volt batteries to power this. the op amps require +/- power rather then +/ground so thats why the batteries are setup like that. the sensor itself cant handle more then 12 volts and works best at 10 volts so we used a voltage divider to cut it down from 9 volts to 5 volts. the only problem with this is that it drains the batteries pretty fast. to help cut down on this, i wired in some power switches inline with the (+) power and the (-) power. a few hours and they are drained enough to cause innacuracy problems. this is because the sensor's output is dependednt on the voltage it gets. same thing with the op amps. the more voltage they get, the more they amplify. so calibration was a bit of a problem. also the resistor with the arrow through it is a variable resistor or a potentiometer. this is what we used as our gain resistor so we could manually adjust the gain (explained later). we used a potentiometer to be able to account for resistor innacurracies. we actually used 2 of them because we set our circuit up to work on 2 different ranges or weight (0-150g, and 0-1500g) for maximum accuracy at smaller weights. we switched between the 2 with a couple of resistors that were powered with our handy board. basically we set our program up to automatically sense when it was over the 150g, and if it was it would click on the relays and automatically go to the larger range.
it was hard to get the right range. the sensor itself only puts out from 0-.24 volts or something like that. but our 8-bit ADC (analog to digital converter, 0-255) works on a 0-5 volt range. thats what the amplifiers do, they turn 0-.24 into 0-5 for the best possible accuracy for those who dont know this is called gain. after the adc recieves the input from the circuit it turns the voltage into a bianary number. so for 0V it would output 0, and for 5V it would output 255. and everything inbetween with increments of 1. so thats like .0196 volts/digital output. i think the best we could get after the adc was 12-220, rather then the ideal 0-255. this really only affects the amount of weight we can weigh.
we had lots of trouble with our tray causeing accuracy problems. and then with the constant trouble shooting draining the battery thus causing more accuracy problems, then having to keep adjusting the program to the right numbers, it was a nightmare. i think we easily spent 20+ hours just trying to get everything working properly. for our presentation we screwed it up and it didnt do so good, butwhen we got it working, we had a best accuracy of about 2% we measured a 1000g weight at 1020g and a dime 2.5g, at 2.8g.
heres some pics of the finished box. the box itself looks like shit but im most proud of the wiring (cause i did it all myself)
the two black boxes are the relays, the two blue things are the potentiometers, and the 3 chips are the op amps.
all in all it was great learning how to do this stuff, although the whole thing was way more complicated then it should have been. even my professor was stumped for a while and he had to take it home and figure it out.
and there were lots of groups that had really great even more complex scales then our group. some looked really well put together. one group had a platform that compressed a piston, the piston compressed gas in a cylinder and a pressure sensor would sense the pressure created, then go to the computer to calculate the weight. another group had a basket out on the end of a plastic beam, and they had a bending guage to see how much the beam bent, and would calculate the weight from that. lots of cool stuff. althought i think our group had the best wiring
most everyone just shoved wires and resistors into a breadboard and shoved it inside their scale.
also great fun stayin in the lab all night playing with the six-axis robotic arm
well this turned int a mamoth of a post so tell me what you think
specs/limitations/goals:
must have a mechanical and electrical aspect to it
must have digital readout
must have a tare function
should weight 0-1500 grams within a 1% accuracy
must have a count function. meaning you put one thing on there (say a penny) have it weigh it, then you put more on, and it tells you how many are there
must read out weight in grams/kilograms and lbs/oz
my professor didnt really expect people to be that accurate, it was just a goal.
basically what our initial idea was get a force sensor that outputted a certain voltage for what weight is on it. so we started on the program based on that. but when we got the sensor in, it had 4 pins on it. i searched the manufactures website and found out that there were 2 outputs. one outputted a voltage(V_2), and the other ouputted a voltage(V_4), and the total output voltage (V_o) was the difference between the two.
V_o = (V_1)-(V_2)
heres the sensor we used
http://content.honeywell.com/sensing/prodi...ce/008028_0.pdf
so basically i have no clue on how to get one ouput that was equal to the difference of the two. luckily the manufacturer had a wiring diagram on how to do this.
here it is.
we used two 9 volt batteries to power this. the op amps require +/- power rather then +/ground so thats why the batteries are setup like that. the sensor itself cant handle more then 12 volts and works best at 10 volts so we used a voltage divider to cut it down from 9 volts to 5 volts. the only problem with this is that it drains the batteries pretty fast. to help cut down on this, i wired in some power switches inline with the (+) power and the (-) power. a few hours and they are drained enough to cause innacuracy problems. this is because the sensor's output is dependednt on the voltage it gets. same thing with the op amps. the more voltage they get, the more they amplify. so calibration was a bit of a problem. also the resistor with the arrow through it is a variable resistor or a potentiometer. this is what we used as our gain resistor so we could manually adjust the gain (explained later). we used a potentiometer to be able to account for resistor innacurracies. we actually used 2 of them because we set our circuit up to work on 2 different ranges or weight (0-150g, and 0-1500g) for maximum accuracy at smaller weights. we switched between the 2 with a couple of resistors that were powered with our handy board. basically we set our program up to automatically sense when it was over the 150g, and if it was it would click on the relays and automatically go to the larger range.
it was hard to get the right range. the sensor itself only puts out from 0-.24 volts or something like that. but our 8-bit ADC (analog to digital converter, 0-255) works on a 0-5 volt range. thats what the amplifiers do, they turn 0-.24 into 0-5 for the best possible accuracy for those who dont know this is called gain. after the adc recieves the input from the circuit it turns the voltage into a bianary number. so for 0V it would output 0, and for 5V it would output 255. and everything inbetween with increments of 1. so thats like .0196 volts/digital output. i think the best we could get after the adc was 12-220, rather then the ideal 0-255. this really only affects the amount of weight we can weigh.
we had lots of trouble with our tray causeing accuracy problems. and then with the constant trouble shooting draining the battery thus causing more accuracy problems, then having to keep adjusting the program to the right numbers, it was a nightmare. i think we easily spent 20+ hours just trying to get everything working properly. for our presentation we screwed it up and it didnt do so good, butwhen we got it working, we had a best accuracy of about 2% we measured a 1000g weight at 1020g and a dime 2.5g, at 2.8g.
heres some pics of the finished box. the box itself looks like shit but im most proud of the wiring (cause i did it all myself)
the two black boxes are the relays, the two blue things are the potentiometers, and the 3 chips are the op amps.
all in all it was great learning how to do this stuff, although the whole thing was way more complicated then it should have been. even my professor was stumped for a while and he had to take it home and figure it out.
and there were lots of groups that had really great even more complex scales then our group. some looked really well put together. one group had a platform that compressed a piston, the piston compressed gas in a cylinder and a pressure sensor would sense the pressure created, then go to the computer to calculate the weight. another group had a basket out on the end of a plastic beam, and they had a bending guage to see how much the beam bent, and would calculate the weight from that. lots of cool stuff. althought i think our group had the best wiring
most everyone just shoved wires and resistors into a breadboard and shoved it inside their scale.also great fun stayin in the lab all night playing with the six-axis robotic arm
well this turned int a mamoth of a post so tell me what you think
