thanks Ron! I wasn't sure if I needed one there.

 

thanks Ron! I wasn't sure if I needed one there.

Your schematic does not show pin 4 connected to ground.

 

if your truly looking for a +/- 3000 range, you should retain the negative supply (with cap). This assumes that your DAQ allows provides for a +/- input.

If so, make sure your polarity on the cap is correct with respect to ground (+ terminal to ground).

 

if your truly looking for a +/- 3000 range, you should retain the negative supply (with cap). This assumes that your DAQ allows provides for a +/- input.

If so, make sure your polarity on the cap is correct with respect to ground (+ terminal to ground).

Good catch. I didn't notice the ± in the first post.

LM7910 has a dropout voltage of 2.5V, so you need to choose a different negative regulator.

 

...if your truly looking for a +/- 3000 range...

I alluded to this question in post #12, and the OP didn't challenge measuring only a positive load. But maybe he didn't get the point of my less-than-direct question.

 

I alluded to this question in post #12, and the OP didn't challenge measuring only a positive load. But maybe he didn't get the point of my less-than-direct question.

noted,

nulling the output would require a negative bias, unless full scale reduction and compensation was employed with the ADC.

 

noted,

nulling the output would require a negative bias, unless full scale reduction and compensation was employed with the ADC.

I helped on a very similar project about a month ago. I suggested a null scheme in post #61. It's not perfect, but it might be good enough.

More accurate output nulling could be done by temporarily applying a voltage divider to the reference input. The exact value is not very important, but a low voltage will keep common mode rejection (or lack thereof) from being an issue. Adjust the null pot while monitoring the voltage between the ref input and the output, and set it to zero. Then switch the reference input back ground. See the attachment.

 

if your truly looking for a +/- 3000 range, you should retain the negative supply (with cap). This assumes that your DAQ allows provides for a +/- input.

If so, make sure your polarity on the cap is correct with respect to ground (+ terminal to ground).

Eventually I need it to read both compressive and tensile forces so I will need the full +/- 3k range (though the mechanism will not ever exceed 1.5k lbs). I was thinking of amplifying it so that 10V = 1.5k lbs in order to get better resolution but I'm not sure if that makes sense.

I will check out the capacitors.. I guess mine are backwards? I'm not sure how you can tell which way its supposed to go, some clarification on this would be much appreciated. I assumed its the direction of current flow as in voltage drop so the positive end gets connected to the higher voltage and the negative end to the lower, or ground.

 

I helped on a very similar project about a month ago. I suggested a null scheme in post #61. It's not perfect, but it might be good enough.

More accurate output nulling could be done by temporarily applying a voltage divider to the reference input. The exact value is not very important, but a low voltage will keep common mode rejection (or lack thereof) from being an issue. Adjust the null pot while monitoring the voltage between the ref input and the output, and set it to zero. Then switch the reference input back ground. See the attachment.

I wish I understood what you meant here. I am just reading the voltage and have a software tare that accounts for the voltage at zero load then in subsequent measurements I just subtract that voltage before figuring out the actual load. Am I talking about the same issue here or am I completely off?

 

If so, make sure your polarity on the cap is correct with respect to ground (+ terminal to ground).

Um... I just switched it and the capacitor blew up in my face... Are you sure about this? I'm hoping it didn't damage any other part of my circuit.

Could you also please specify which capacitor? Do all of them have to be reversed? Please refer to my last schematic. I'm scared of changing anything now as I'm not sure if I'll blow up another circuit.

So after spending the last 15 minutes reading about this.. I had it right. Positive to higher voltage, negative to ground. Everything specifically warns against putting the positive terminal into ground. I'm not sure why you told me to switch it? Hope it wasn't deliberate.

 

I wish I understood what you meant here. I am just reading the voltage and have a software tare that accounts for the voltage at zero load then in subsequent measurements I just subtract that voltage before figuring out the actual load. Am I talking about the same issue here or am I completely off?

If you have a single power supply, you may need offset adjustment to get the output off the ground rail, if it happens to be stuck there with no strain on the load cell. However, since you are doing software tare adjustment, you could just add a little offset to the Ref input if it is needed. If your output is positive with no strain, then you need to do nothing. Keep in mind that offset wastes dynamic range.

 

Um... I just switched it and the capacitor blew up in my face... Are you sure about this? I'm hoping it didn't damage any other part of my circuit.

Could you also please specify which capacitor? Do all of them have to be reversed? Please refer to my last schematic. I'm scared of changing anything now as I'm not sure if I'll blow up another circuit.

So after spending the last 15 minutes reading about this.. I had it right. Positive to higher voltage, negative to ground. Everything specifically warns against putting the positive terminal into ground. I'm not sure why you told me to switch it? Hope it wasn't deliberate.

Go back and read the post carefully. He was referring to the cap on the negative supply.

 

Go back and read the post carefully. He was referring to the cap on the negative supply.

I guess I misunderstood. I thought he was referring to my new (sans neg supply) schematic. Good to know though thank you.

 

Did you see post #31?

 

Eventually I need it to read both compressive and tensile forces so I will need the full +/- 3k range

That requires either returning to a dual supply or establishing a virtual ground at the midpoint voltage of a single supply.

Working with a single supply until you get it working nicely may still be a good idea, but plan to make changes before you're done.

 

Um... I just switched it and the capacitor blew up in my face... Are you sure about this? I'm hoping it didn't damage any other part of my circuit.

Could you also please specify which capacitor? Do all of them have to be reversed? Please refer to my last schematic. I'm scared of changing anything now as I'm not sure if I'll blow up another circuit.

So after spending the last 15 minutes reading about this.. I had it right. Positive to higher voltage, negative to ground. Everything specifically warns against putting the positive terminal into ground. I'm not sure why you told me to switch it? Hope it wasn't deliberate.

so, is that a sunburn or are you just happy to see me?

What we probably need to ask, is two fold, the input requirements of your DAQ and if you want the full 12 bit resolution over that input.

 

Did you see post #31?

Yeah my output at no strain is ~0.01147 V. So by what you said, is this fine since its not in the negative range?

So what I'm doing is subtracting that voltage from my signal, then multiplying the remainder by 3,000/(478*2.0918*9.95) to get my lbs value. The 3,000 is the maximum load on the load cell, the 2.0918 is the mV/V readout at 3,000 lbs, the 9.95 is the excitation voltage and the 478 is the gain of my amplifier. I chose to just go with the maximum load/readout because based on the calibration chart it is very linear with about 0.1% error which I can endure. With this I get a readout and if manually I pull on the load cell I can get some readouts and they seemed reasonable.. or so I thought. I just put a 5 lbs load on it and it fluctuates between 1.57 and 3.1 lbs, I put 10 lbs and it fluctuates between 1.57, 3.1 and 4.7 lbs. Obviously I don't have it set up right but I'm not sure what I'm doing wrong.

What we probably need to ask, is two fold, the input requirements of your DAQ and if you want the full 12 bit resolution over that input.

Ideally I would like to get the full 12 bit resolution. As far as the input requirements of my DAQ go, here they are (this is a copy paste from the specs): "4 HV Inputs have ±10 Volt or -10/+20 Volt Range." This is what I went off of by trying to tailor the gain so that the maximum voltage after being amplified would land on 10V. My calculation was 10/20.0918E-3 = 478. Though initially when I did this I didn't account for the voltage at zero load. I also assumed 10V instead of the actual 9.95V for excitation. Is that 0.01147V throwing off my readout and causing me to get the wrong values since the actual thing spans 0.01147 - 10.### V? Also what is my load resolution in lbs? I thought 3,000/2^12 but now seeing as how my DAQ spans a total of 20V I think I may have miscalculated that.

Thanks again for all the help.

 

So I went and changed the gain a bit. This time (10 - 0.0114691)/(2.0918e-3*9.95) = 479.908. Then same concept, 3,000/(479.908*2.0918e-3*9.95) = 300.345. Now I did the same calculation, subtract 0.0114691 from my DAQ's input then multiply it by 300.345 and now with the 10 lbs weight on the load cell it fluctuates between 4.7 and 6.1 lbs. I can't afford such a huge mistake. If its just a straight few lbs off that's not as big an issue but if this scales then at 1000 lbs its going to be 300-400 off which is horrible. I also just got an email that my LVDT came in.. ah.. now another piece of electronics to play with that I have idea how to use.. only good news is it comes with internal signal conditioning.. phew!

 

So I got the LVDT to work properly but I'm having a lot of trouble getting the load cell to show the right force values. I have to be finished with this by the 25th and would really appreciate any more input you guys can give.

Thanks again in advance.

 

Are you sure the fluctuation is not due to noise in your signal path, or in your ADC reference?
With a gain of 480, your equivalent input LSB is a few microvolts. A crappy ground plane, or poor power supply decoupling, can easily cause this much noise. Just allowing digital signal currents to flow through the analog ground system can cause this. Noise on the bridge power supply can cause this.
It would be difficult to know how to help you without seeing the board and the schematic. Perhaps you should post both of them here.