I purchased this board, because I thought it would be a convenient way to get an amplifier for a strain gauge bridge circuit, and probably cheaper than buying separate elements to make one up myself.
I have traced out the circuit and found it to be as in the attached diagram. (None of the capacitors has any indicated value, so I'm not sure how big they might be). It is a quarter-bridge circuit using the single strain gauge fixed to the circuit board. The amplifiers are in an LM358 IC.
I have modified the board to be a half-bridge circuit by removing R9, and the fixed strain gauge and providing connections to the strain gauges on my apparatus. The circuit connects to +5 VDC and 0 VDC from the Arduino Nano. Everything works as I expected, except the output saturates at about +3.2 V, and doesn't really get close to the supply potential. That doesn't really make the best use of the range of the A-to-D input.
Also, the output seems quite heavily damped. The trace in the attachment shows the output read by the A-to-D of the Arduino board averaged over 100 ms intervals, while I place different calibration loads on the apparatus. Between the fixed loads (10, 5, 2, 1, 0.5, 0.2, 0.1, 5, 10 N) there is no load on the apparatus, yet the indicated value falls relatively slowy at first (having several indicated points on the falling edge) and approaches zero really very slowly. [The rising edges are similar]. The project is a rocket motor test apparatus, and I need it to record reasonably faithfully changes over half the full range of the sensor in about 100 ms. I suspect that C7 isn't helping, and I plan to remove it. Before I do that (because I'm pretty sure I'll never get it back on) is there anything about the LM358 amplifiers, or the A-to-D input of the Arduino that is likely to be the reason for this overly heavy damping?
If I need to make my own amplifier to overcome these limitations of LM358, what would be a good one to pick? It still needs to work from a single supply +5 VDC & 0 VDC, but ideally swings all the way to the supply rails.
C7 is the problem - its loading the output of the second opamp, which is bad news (it could be oscillating at
high frequency). Its also causing the large time constant I think (the LM358 has a very asymmetric output
stage, hence the much slower falling edges).
I suggest removing C7.
You might then have some noise, but that's best filtered with a 1k / 1uF divider separate to the unit.
chapellane:
I purchased this board, because I thought it would be a convenient way to get an amplifier for a strain gauge bridge circuit
Why don't you use a HX711 breakout board for that, like everybody else does.
Regulated supply for the strain gauge, analogue pre-amp, and 24-bit A/D (digital out). All in one.
Leo..
chapellane:
Everything works as I expected, except the output saturates at about +3.2 V, and doesn't really get close to the supply potential. That doesn't really make the best use of the range of the A-to-D input.
You could set Aref to EXTERNAL in setup(), and connect the 3.3volt pin to the Aref pin.
Then 0-3.3volt produces 0-1023 A/D.
Leo..
Wawa:
Why don't you use a HX711 breakout board for that, like everybody else does.
Regulated supply for the strain gauge, analogue pre-amp, and 24-bit A/D (digital out). All in one.
Leo..
Probably mostly because I don't think I came across it. At 10 /s it probably isn't really fast enough for what I want, but I see it can be switched to 80 /s, and that would be. I might switch to it, if I can't be successful with modifications to the one I already picked up. Thanks.
Do you think so? I thought everything in that part of the circuit would be pretty much static, and the capacitor would do no harm there at all. (And I thought the same about C6 & C8). How will C5 be affecting the output?
It also for the other half of the differential input into the second opamp it’s not just tied to power, it affects the bridge and slows down the response. Otherwise it would not be tired to the bridge. In any case a better circuit diagram might be helpful. That top line isn’t specified. I’m guessing it’s VCC but not sure and where does it go to the left? Also what’s the power supply is it DC at 5v or split rail?
I just found a better schematic... I actually don’t see any issues with any of these caps.. they should charge pretty quickly since they are not additionally impeded. I wonder if there’s a issue with programming with something interfering or blocking...
As far as rail to rail opamps, those are easy to find if you search under the different manufacturers... such as TI or Analog Devices. I just received a shipment of LM7301 which are rail to rail and can run on single supply. I have not had a chance to test them yet.
WHile I understand this board is designed by someone selling "pre made" circuits, the design is wanting for this application.
The LM358 is a poor choice for such an application. The output will barely go to within 1.5 V of the upper rail.
So 5V - 0.5V(diode) = 4.5 4.5 -1.5 = 3V which is basically what you are getting.
You would have the same issue when the bridge is balanced and the output should be 0V. This amplifier won't go to 0V. Here the lower limit is 5 to 10 mv. Certainly much better but you will also find as the output gets near the lower rail the linearity becomes worse.
You could offset your device so the output is say 0.5 volts when the bridge is balanced and subtract that number from you Nano conversion.
Wawa:
Why don't you use a HX711 breakout board for that, like everybody else does.
Regulated supply for the strain gauge, analogue pre-amp, and 24-bit A/D (digital out). All in one.
Leo..
I took out the C7 capacitor to no avail, the output signal still seems to be damped. The rising edge is of the output always seems to take about 500 ms.
I swapped the amplifier to the HX711 board, but I still have slow response. (Even with the board set to sample at 80 /s).
The whole thing needs some more work, maybe I'll have to make my own amplifier after all.
chapellane:
I took out the C7 capacitor to no avail, the output signal still seems to be damped. The rising edge is of the output always seems to take about 500 ms.
I swapped the amplifier to the HX711 board, but I still have slow response. (Even with the board set to sample at 80 /s).
The whole thing needs some more work, maybe I'll have to make my own amplifier after all.
I started to suspect that there was nothing wrong with either of the electronic boards, and the problem lies elsewhere. I made a square-wave signal generator and applied that to the HX711 amplifier, and found that it could follow the rising and falling edges easily.
What I was measuring before wasn't anything undesirable in the amplifier. What I was seeing was an accurate measurement of an undesirable physical property of the material that the mechanical apparatus was made of.
Yes with the Arduino. With the fastest sampling possible (84 /s) there was only one sample which fell on each of the rising or falling edges of the applied wave. [I applied a signal as close as possible to what I expected should be coming from the rest of the apparatus, a differential square wave about ±15 mV, with a common mode approximately DC level of about 2.5 V]. That showed me that the amplifier on the HX711 board is capable of switching its output much more rapidly than I had observed in my earlier tests, and that there is no problem with the electronic part of the experiment. The damping I observed was physical damping by the materials of the mechanical parts of the apparatus.
