MPS20N0040D - Lib in micropython to use this sensor.

[pierregraciano]

Hello everyone.
I bought a MPS20N0040D sensor, and Im searching ways to test it on my ESP32, using micropython. Do you have some website or code to help me to test this sensor? The module has HX710b.
Thank you.

[robert-hh]

The MPS20N0040D sensor itself is just a passive resistor bridge. It requires additional circuitry for connecting it to a ESP32. There are breakout boards available which use this sensor. Do you use such a board? Then the type of this board and the interface chip is important.

[pierregraciano]

What limits this pressure range is the HX710B?

[robert-hh]

Yes. The input range of the HX710 is about +/-12mV (1.65V / 128) at 3.3V supply. At 5V supply it's about +/- 18mV, but at the end of the range it get's nonlinear, so you cannot use the full range.

[pierregraciano]

Do you think is a good idea buy some HX711 and use in place of HX710? The footprints are different, but some wires can fix it.

[robert-hh]

If you need the larger range, then you have to change your approach. You can get as well just the MPS20N040 sensor itself in a DIP package, 2.54mm pitch, and use these with a HX711 breakout at a gain of 32 or another sensitive A/D converter breakout like the ADS1115, which has +/-256,V as lowest input range. You can test that easily on a breakout board.

[robert-hh]

There is another quick hack: if you have good eyes and a steady hand, you can solder a resistor across the sensor output. The source impedance of the sensor bridge is about 5k, so a 1.5k resistor should increase the range by about 4. That will of course have a little bit more noise than a HX711 at gain 32, but less than the other noise sources.
At the board here it's between Pins 4 and 6 of the sensor.

[robert-hh]

I made two test series comparing the MPS20N040D sensor and a calibrated DLV0105D sensor attached to the same pressure volume. The volume has a small leak, so the pressure falls within a suitable time, giving a nice range test. I used two variants of the MPS20N04D sensor: a) raw sensor attached to a HX710B, b) Sensitivity reduced sensor by attaching a 1.5k resistor across pins 4 and 6 attached to a HX710B.

a) Raw sensor: The resulting conversion factor of MPS20N04D at HX710B reading to kPa was 710000. The pressure range is about +/- 12 kPa. A second MSP20N040D sensor had a conversion factor of 710000.
b) Sensitivity reduced sensor. The resulting conversion factor of MPS20N04D at HX710B reading to kPa was 174380. The pressure range is about +/- 48 kPa.

In the test, the MPS20N040D was sampled first, then the DLV015D. Some diagrams of the test series below. The diagrams show, that both sensors run pretty synchronous.

a) Raw sensor:

Detail at the low pressure end of the test:

b) Wider range sensor using a 1k5 resistor across the sensor output.

Detail at the low pressure end of the test:

[robert-hh]

By the way: the data sheet of the MPS20N040D is inconsistent. It mentions 540PSI and 40kPa, but 40kPa is equivalent to about 5.4 PSI, or the other way around: 540PSI is equivalent to about 4000 kPa or 40bar.

[robert-hh]

As another option to extend the range of the MPS20N0040D, the voltage across the bridge can be reduced, lowering the output range. A test of setting the voltage to 1.25V using a voltage reference results in a pressure range of +/-24kPa. The test results are good. Still, lowering the output with a resistor seems the easier way, and as shown in the diagrams above, the results are fine.

[robert-hh]

Some figures of the sensor with the voltage at 1.25V.

Full test series:

Detail at the lower range:

[robert-hh]

I got samples of the MPS20N0040D sensor and wired them up to a) and ADS1115 A/D converter in the +/-256mV range, and b) an HX711 at the B inputs with a gain of 32. Both variants cover the full range of the pressure sensor, and the results show good tracking with the DLV-015D sensor, as above, with higher resolution than the DLV-015D.

Test series with HX711:

Detail at low pressure:

[pierregraciano]

Hello Robert. Thank you so much, you will help me a lot! I'll mark your answer as correct. Thank you again!!

[robert-hh]

Maenwhile I made another test using the raw MPS20N004D sensor connected to a ADS1115 ADC at a pressure up to 100kPa. Diagram below, as always compared to the DLB015d. Since the pressure falls almost in a e**-t/T series, a log/lin scaling for the results should be close to a straight line. The output of the mps20N0040D is shown as mV. The yellow line the is ratio between the output of MPS20N0040D and the DLV015D. It should be mostly constant. Since the DLV015D has a lower resolution, about 10.5 digits, the results at low pressure get worse.