Here are some easy-to-follow steps to help you troubleshoot potential load cell problems. Before you begin you will need a good quality digital multimeter, at least a 4½ digit ohm meter. The tests are: physical inspection, zero balance, bridge resistance and resistance to ground.
How does it look? If it is covered with rust, corroded or badly oxidized, chances are the corrosion has worked its way into the strain gauge area as well. If the general and physical condition appear good, then you need to look at specifics: sealing areas, the element itself, and the cable.
In most load cells, areas of the load cell are sealed to protect the contents from contamination by water and chemicals. To see if any seals have been degraded, get right up close to the cell and look at the strain gauge seals. Is rust concentrated on a part of the cover weld? If there is no cover, do you see any tiny holes in the potting? These are indications that there has been contamination to the gauge area. Check the load cell cable entrance for signs of contamination.
Other items to look for: metal distortion or cracks, metal rippling, cracks in the weld, or abrasions in the metal. It may be necessary to remove the load cell and check it for physical distortion against a straight edge.
No inspection would be complete without thoroughly inspecting the cable. Cable should be free of cuts, crimps and abrasions. If your cable is cut and in a wet environment, water or chemicals can "wick" up the cable into the strain gauge area, causing load cell failure. If your physical inspection fails to uncover any identifiable damage, a more detailed evaluation is required.
This test is effective in determining if the load cell has been subjected to a physical distortion, possibly caused by overload, shock load or metal fatigue. Before beginning the test, the load cell must be in a "no load" condition. That is, the cell should be removed from the scale or the dead load must be counterbalanced.
Now that the cell is not under any load, disconnect the signal leads and measure the voltage across the negative signal and positive signal. The color code for determining negative- and positive-signal leads is provided on the calibration certification with each load cell. The output should be within the manufacturer’s specifications for zero balance, usually ± 1 percent of full scale output. During the test, the excitation leads should remain connected with the excitation voltage supplied by the digital weight indicator. Be certain to use exactly the same indicator that is used in the cell’s daily operation to get a reading accurate to the application.
The usual value for a 1 percent shift in zero balance is three-tenths millivolt, assuming 10 volts excitation on a three millivolt per volt output load cell. To determine your application’s zero shift, multiply the excitation volts supplied by your indicator by the millivolt per volt rating of your load cell. When performing your field test, remember that load cells can shift up to 10 percent of full scale and still function correctly. If your test cell displays a shift under 10 percent, you may have another problem with your suspect cell, and further testing is required. If the test cell displays a shift greater than 10 percent, it has probably been physically distorted and should be replaced.
Before testing bridge resistance, disconnect the load cell from the digital weight indicator. Find the positive and negative excitation leads and measure across them with a multimeter to find the input resistance. Don’t be alarmed if the reading exceeds the rated output for the load cell. It is not uncommon for readings as high as 375 ohms for a 350 ohm load cell. The difference is caused by compensating resistors built into the input lines to balance out differences caused by temperature or manufacturing imperfections. However, if the multimeter shows an input resistance greater than 110 percent of the stated output value (385 ohms for a 350 ohm cell or 770 ohms for a 700 ohm cell), the cell may have been damaged and should be inspected further.
If the excitation resistance check is within specs, test the output resistance across the positive and negative signal leads. This is a more delicate reading, and you should get 350 ohms ±1 percent (350 ohm cell). Readings outside the 1 percent tolerance usually indicate a damaged cell.
Now comes the tricky part. Even if the overall output resistance test was within normal specifications, you could still have a damaged load cell. Often when a load cell is damaged by overload or shock load, opposite pairs of resistors will be deformed by the stress—equally, but in opposite directions. The only way to determine this is to test each individual leg of the bridge. The Wheatstone Bridge diagram, illustrates a load cell resistance bridge and shows the test procedure and results of a sample cell damaged in such a manner. We’ll call the legs that are in tension under load T1 and T2, and the legs under compression C1 and C2.
With the multimeter, we tested each leg and got the following readings:
- T1(–Sig, +Exc) = 282 Ω
- C1(–Sig, –Exc) = 278 Ω
- T2(+Sig, –Exc) = 282 Ω
- C2(+Sig, +Exc) = 278 Ω
NOTE: When testing leg resistance, a reading of 0 ohm or eight means a broken wire or loose connection within the cell. In a good load cell in a “no load” condition, all legs need not have exactly equal resistance, but the following relationships must hold true:
- (C1 + T1) = (T2 + C2)
In this damaged load cell, both tension legs read four ohms higher than their corresponding compression legs. The equal damage mimics a balanced bridge in the output resistance test (3 above), but the individual leg tests (1, 2 above) show that the cell must be replaced.
NOTE: On multiple-cell applications for matched millivolt output, excitation resistance values may be higher than 110 percent.
Resistance to Ground
If the load cell has passed all tests so far but is still not performing to specifications, check for electrical leakage or shorts. Leakage is nearly always caused by water contamination within the load cell or cable, or by a damaged or cut cable. Electrical shorting caused by water is usually first detected in an indicator readout that is always unstable, as if the scale were constantly "in motion." The wrong cell in the wrong place is the leading cause of water contamination. Almost always, these leaking cells are "environmentally-protected" models designed for normal non-washdown, not the "hermetically sealed" models that would have stood up to washdown and other tough applications.
Another cause is loose or broken solder connections. Loose or broken solder connections give an unstable readout only when the cell is bumped or moves enough so the loose wire contacts the load cell body. When the loaded scale is at rest, the reading is stable.
To really nail down electrical leakage problems though, test resistance to ground with a low-voltage megohmmeter. Use caution—a high-voltage meter that puts more than 50 volts of direct current into the cell may damage the strain gauges. If the shield is tied to the case, twist all four leads together and test between them and the load cell metal body. If the shield is not tied to the case, twist all four leads and the shield wire together and test between them and the body. If the result is not over 5,000 megohms, current is leaking to the body somewhere.
If the cell fails this test, remove the shield wire and test with only the four live leads to the metal body. If this tests correctly (over 5,000 megohms), you can be reasonably sure current is not leaking through a break in the cable insulation or inside the gauge cavity.
Minor water infiltration problems can sometimes be solved outside the factory. If you are sure that water contamination has occurred and if you are sure that the cable entrance seal is the entry point, try this remedy: Remove the cell to a warm, dry location for a few days, allowing the strain gauge potting to dry. Before putting the cell back into service, seal with silicone around the cable entry point in the load cell body. This prevents the reentry of water vapor into the cell.