Repairing digital scales (readout not stablizing)

[Dannyx]

Good day folks. Some of you may know that this place where I work at gets its hands on all sorts of crap, from your usual stuff like TVs and PCs to fairly niche stuff like audio mixers and to downright junk like VCRs and CD players.

The device I currently have on hand falls in the second category, niche stuff, which is a set of digital scales which you'd find in a small supermarket. It's not a cash register, it just weighs stuff and displays it. It turns on alright but the problem is the readout never stabilizes: it fluctuates up/down by roughly 5 grams and it seems it tends to keep going up rather than down. I snapped some pics of the interior...nothing special: a load cell with some metal stuff attached to it for supporting the top plate and a simple board. I haven't worked with load cells before so I don't know what I should be looking for exactly and what causes unstable readings in them. I see 4 wires going to it: two of them are obvious (VCC and GND), but the other two, from what I've read are the "signal" wires which is not even a communication protocol or anything - just a set of resistors (Wheatstone bridge ?) which change value. Is it the load cell itself that is likely to be faulty in such scenario (by having something over its maximum limit placed on it for instance) or is it more likely to be an electronic issue like an op-amp ? What do you guys think ?

Two burnt resistors there, I know, but those are likely part of the battery circuit which experienced a bad battery and caused them to overheat, but I don't think they're part of the issue here...

[stj]

Re: Repairing digital scales (readout not stablizing)

starting point - why is that resistor cooked?

[Dannyx]

Re: Repairing digital scales (readout not stablizing)

starting point - why is that resistor cooked?

Haven't analyzed the circuit in depth to work out exactly what those resistors do, but I can surmise they're doing some sort of current limiting for the SLA battery that goes in a compartment on the back (not installed), so something must've gone wrong there and caused a large current dissipation on those resistors. The unit has two rocker switches on the bottom: one for power on/off and the other one to select between battery and the DC adapter (?). It powers on fine with just the adapter, so unless this is a power issue it shouldn't be of concern....for now. I shall replace them nonetheless just because they look ugly and unprofessional

[CapLeaker]

Re: Repairing digital scales (readout not stablizing)

what happens if you take the load cell off? maybe introduce some heat and or cold to certain sections of the board.
actually there are 2 resistors cooked.

[R_J]

Re: Repairing digital scales (readout not stablizing)

If that to220 is a regulator ic, it's caps could be bad and its output noisy, the processor might be seeing that noise and changing the reading.

[Dannyx]

Re: Repairing digital scales (readout not stablizing)

Desoldering the cell causes it to stay at 0, so there's something wrong with the cell IMO....

[llonen]

Re: Repairing digital scales (readout not stablizing)

The cell should be producing a signal directly proportional to the magnitude of the force being applied. So take a look at the output and see what kind of a output you are getting from the two 'signal' wires, it should be relatively stable. The cell output is rated in millivolts per volt (mV/V) of the difference voltage at full rated mechanical load so the signal you will be looking at will be quite small even at full load when excited. The bridge is excited with stabilised voltage (usually 10V, but can be less for battery powered instrumentation). The difference voltage proportional to the load then appears on the signal outputs. So any irregularities with the power source are going to affect the output 'signal'

[CapLeaker]

Re: Repairing digital scales (readout not stablizing)

If that to220 is a regulator ic, it's caps could be bad and its output noisy, the processor might be seeing that noise and changing the reading.

That is a valid point, since voltage regulators can fail in a weird way and the fact there has been some sort of a power problem on the board too.

[Dannyx]

Re: Repairing digital scales (readout not stablizing)

The one with the heatsink is a 7808 regulator and the other one with no heatsink is a 5v regulator - both read fine and someone replaced the caps on those lines before. I had a look at the circuit as well: power from the DC jack goes through the "bottom" resistor into the input pin of the 7808. The resulting 8v go through a diode to the positive terminal of the battery compartment and also to the power switch. After the power switch, the 8v go to the 5v regulator and from there it does its thing powering different stuff, including the load cell of course.

With the green/white wires desoldered, I took some measurements and although I can't remember exactly what they were, the didn't indicate any fluctuations which would appear on the display, but this of course could be because a DMM is not as sensitive as the op-amp used there. As a fun fact, when the scale is turned on without the cell it zeroes out like you'd expect, but if I then merely touch the two pads where the signal lines are, the reading goes wild and never stabilizes again until I cycle the power - is this the effect of capacitance on a very sensitive op-amp ?

[CapLeaker]

Re: Repairing digital scales (readout not stablizing)

well, the cell's output is in double digits mV scale on the DMM. Now it depends on how sensitive and the count rating of your DMM. So I am guessing the battery in there is a 6V SLA. I'd wash the board, clean and reseat the main IC (check the socket for corrosion). It kind of looks like there is something on the PCB. If that is slightly conductive, it would fool the op amp or main IC too. Hook everything back up and check the input and the output of the 5V regulator with an oscilloscope to look for noise. The caps had been replaced, I'd check them too. Usually those scales stay turned on, once it is in service, so checking or replacing them can't hurt. Op amps and other components can loose a little over time, so you could replace that, do a re calibration and see what is happening.