Load cell basics – technical questions and answers part 2

Interface has hosted a series of ForceLeaders webinars over the past year on topics such as multi-axis sensors, gauges, torque transducers, custom solutions and more. We plan to continue this popular series into the new year based on the feedback we've received on topics of interest and support needs.

One of the features we like to include in our events is answering a list of top questions as well as questions we receive from attendees. In our Load Cell Basics event, we addressed many interesting inquiries as well as the most important questions about our precision load cell products.

We highlighted the first set of questions in our recent post, Load Cell Basics Technical Q&A Part One. And there's more: Below you'll find more questions we answered during the live event. A webinar summary is also online, recapping the key topics we discussed during the hour-long event with force measurement experts Keith Skidmore and Brian Peters.

Is there a better way to tare a load cell?
Tare is defined as a deduction from the gross weight of a substance and its container, taking into account the weight of the container, sometimes also referred to as the container. For the purposes of this reference, we note that the container is the actual body of the load cell.

Regardless of whether a load cell is tared electronically or mechanically, the preload taring of the system is usually carried out using instrumentation. You want to ensure that the mechanical tare load placed on the sensor does not limit the usable capacity range. As far as the load cell is concerned, there is actually no need to accommodate any type of mechanical tare. They are linear devices if you stay within the range. When it comes to mechanical preload, how you preload some attachments is important, especially with the Low Profile load cells. We recommend preloading if you can. This is how we achieve the best results so that the load path through the threads does not change, which can have a very small impact on linearity hysteresis.

Does frequency measurement affect the response speed of the load cell?
This depends on the mechanical bending design of the respective sensor. It is important to know that every sensor is a mechanical spring. Depending on the stiffness or deflection properties of a load cell, different natural frequencies arise. This information is typically documented in our product data sheets by model and configuration. In general, the stiffer the sensor, the better its behavior in cyclic applications.

Keith also notes that load cells work great at hundreds of hertz. If you're asking about tens of thousands of hertz, they're not. There is a point where very high frequencies of these standard load cells are not as effective. Most of our typical test customers who apply cyclic loads to a test specimen will confirm that low profile load cells or strain gauge based load cells are excellent for their frequency requirements. Contact our application engineers if you have an extremely high frequency requirement.

Tare and development of an error budget
This question comes from an engineer working out a calculated error budget that already has a mechanical tare load on the system. If you tare the system before starting the test, how does this affect the initial preload in the system? Does this affect your error budget? The answer is: yes, it can have some influence. You can compensate for the temperature effect at zero; However, if you have a lot of preloads, this won't work. They have to go back to zero to compensate for the effect. This can impact your error budget. Basically, going back to zero eliminates the temperature effect. You cannot make the same assumption in your budget if you have significant prior expenses.

Does Interface recommend specific amplifier instrumentation products?
We have a wide range of amplifier solutions. They range from small, built-in amplifiers to in-line options that mount on DIN rails. We also offer various digital output meters for Ethernet, Ethercat, Modbus and USB.

It is important that you understand that there are errors that can affect the performance of the meters. Choosing the right measurement device should depend on the application and changes in the test environment. Resolution is a crucial factor when doing a digital conversion. There are many parameters to consider. We encourage you to review Interface's offering and contact our application engineers to review your specific requirements.

What is the most common problem when installing a load cell?
One of the problems we get asked about is usually related to mechanical installation. It's not as simple as a nut, bolt and threaded connection. There's a lot more to it than that, and what's often overlooked is the importance of keeping things properly aligned to avoid parallel load paths. The entire load must be passed through the load cell to provide a correct measurement. For example, if you have a mounting plate and the plates sit on the load cell screw heads, the load will be bridged. Mechanical installation is often overlooked when planning the test setup. It is often not given enough importance when it comes to your exam.

We also often see that users either have insufficient torque in an assembly or perhaps too much torque for the lower capacity load cells. It is advisable to pay attention to the installation torque. If you are not familiar with mechanical installation, we recommend requesting our detailed installation instructions. We will provide you with information on the recommended tightening torques and the correct methods for attaching the sensor. A load cell can easily be over-torqued because it is designed to handle axial loads. We have seen users crank things and thereby transfer too much torsional load, resulting in performance issues that may not be immediately noticeable.

Do cables and amplifiers affect load cell calibration results?
The answer is that they can certainly impact results. That doesn't mean it has to be that way, but it can affect performance and measurement accuracy. If you e.g. For example, if you change the length of the cable or have a device with a different temperature where the cable goes from an oven to a room with a cooler temperature, these things can affect the measurement. Even if the load cell is at a constant temperature but the cable changes temperature, this will result in a change in the signal. This is something you don't want in your measurement.

You can mitigate the temperature by shielding the cable or using a thicker cable that has less impact. Some of these factors can be compensated for through calibration. If you attach a long cable and then load your load cell, it will take into account that there is a longer cable and you can fully calibrate out the effect of cable length on signal drop due to resistance. Some devices measure the voltage on the load cell with a sensor. In this case it is insensitive to any cable effects. You can change the cable length and it won't make a difference.

Is a force application of 10kn (2.250) lb. too much for a 2.000 lb. Load cell?
Yes, this is above the calibrated range and therefore performance cannot be guaranteed. Load cells typically remain linear beyond capacity, but there is no guarantee of accuracy. This is especially true if you overload the load cell with frequency. They start to stress the load cell in areas that you should avoid to keep the results accurate. In many cases it is better to use a larger sensor. Pay attention to actual performance. This is an advantage of most of our products, especially the LowProfile series.

Is there a maximum sampling frequency for strain gauge load cells?
This depends somewhat on the model of the load cell. Hundreds of hundreds of Hertz is certainly appropriate. If you come close to the natural frequency specified on our product data sheet, we recommend that you discuss the application with our application engineers.

We know engineers who use load cells for safety testing. For example, you may need to sample 10 kilohertz. The load cell makes no difference to the speed at which you sample the signal, but these cells can only respond with a response time of typically one millisecond. There are no moving parts or active circuits; However, typically something is tested until it fails. It is recommended to sample at a few thousand hertz to ensure you capture the exact point at which the system fails or spikes. If you are concerned about the introduced creep error, you should record the non-zero value as soon as you release the load. This is an accurate measure of how much creep error has been introduced since it is symmetrical.

We train our load cells before calibration. This is relatively quick and is typically performed in hydraulic frames that are automated for testing and cycle the load at short cycle intervals, about 30 seconds. It really depends on the extent of the stress, not the duration of the stress. Testing the load at 120% for an hour is no better than 10 seconds, except there is a lot of creep for an hour. During an exercise you have to wait longer until the crawling process starts again. So if the load is a fast cycle, you can quickly check the zero point.