HX711 load cells giving wrong readings

Question

I have four HX711 cells in a Wheatstone configuration. I use them from home automation to measure bed presence. It worked correctly for a few weeks, but suddenly stopped working (cannot get any useful readings anymore.) I suspect one or multiple wires has a broken connection.

I measured the resistance across all the pins, but they seem incorrect. Measured values are in kOhm. Real readings were 1.48, 0.99 etc, but I rounded this in the listing below. I don't think this should be an issue.

As far as I know, the resistance value between all the pins should be 1 kOhm.

E- | A- | 1.5
E- | A+ | 1.5
A+ | A- | 2
A+ | E+ | 1
A- | E+ | 0
E- | E+ | 0

Could someone tell me from these figures which wire is the problem?

Answer 1

Hopefully you are not measuring the resistances with the wires still connected to the ADC board. The 0 resistance reading from E- to E+ (and A- to E*) looks very suspicious. Be sure to lift at least one of the wires to be measured off the ADC board to get an accurate reading.

Initially (assuming your measurements are correct) it seems the problem might be a shorted wire pair. Is it possible that two wires are being crushed together, or shorted to a metal frame?

Measuring the voltages with the circuit powered, the voltage measured from A+ to E-, and from A- to E- should equal about 1/2 the value measured from E+ to E- .

Another idea: Draw out your circuit as shown in the 4th picture in the link you provided. Mark down your specific wiring, add the resistance values as known for the load cells you have (1k each for each resistor shown?), then troubleshoot from there. It should a bit easier to diagnose the problem looking at it that way.

Answer 2

I think you've just learned that serviceability is a thing to keep in mind in your designs :)

The load cells can be in enclosures with connectors, and you'd use external cables to connect them into a circuit you want. That way you can disconnect the cables and measure each load cell individually.

Another approach is to have "pigtail" connections: the load cell end of the cable is soldered inside the enclosure to the load cell, but the other end of the cable has a connector and still allows measurements to be made.

The bridge connections would be made inside the interface box, and the cables from individual load cells to the interface box would be all identical and of same length and construction to maintain the bridge balance.

You can use 3.5mm stereo cables for load cells connections - cheap and do what you need: they have two wires in a shield, or sometimes two individually shielded wires. In either case it's OK.

If you want more robust connectors, XLR would be the next best thing - again, the benefit is that you'd be buying premade cables, and even fairly cheap ones will suffice and take less effort than making them yourself.

If you are 3D printing the enclosures, then 9 pin connectors would work as well, since you won't be laboriously filing or milling the mounting holes.