EnglishFrenchGermanItalianPortugueseSpanishSwedish

SHPB / Kolsky Bar Frequently Asked Questions

  • Setup
  • Testing
  • Data Processing

Q:

I am having difficulty obtaining a long enough pulse in order to fail my samples in one pulse during tension testing. How can I increase the length of this pulse with the strikers & equipment I currently have and what is the maximum striker length I can shoot on my current system if I purchase longer strikers?

A:

Here at REL’s high strain rate laboratory, we use tension strikers up to 6 feet in length for various test setups. Your incident bar length will be the limiting factor on max striker length. We currently sell tension strikers up to 3ft in length. For temporary setups you can also tape multiple striker bars together to generate longer pulses. If they are precision ground strikers from REL and you have them tight and taped well you can get very clean results.

Typically you would take your longest strikers, hold them tight together on the incident bar and wrap the joint well with a dozen turns of high quality electrical tape. Offsetting the wraps left and right on each pass to create a layered effect with diagonal overlap. After several shots you may need to re-apply the tape if you start seeing notches in your incident pulses. You can often stack up 3-4 strikers in this manner if needed with good results. This method can also be used to generate a step loading response by placing various types and sizes of pulse shapers between the taped striker joints for special loading applications.

Q:

What is the best method for calibrating bar modulus on tension SHPB bars?

A:

The best way to calculate the modulus is to use two identical strain gauges mounted to each bar with a significant separation distance between the gauges (greater than 30”). With the pulse time delay between these two gauges and an accurate material density you can calculate modulus effectively [(dist/time) = c = (modulus/density)^.5)]. This is the only recommended method on tension bars and is a good substitute on compression bars. As always make sure the bars are clean, well lubed for steel, or dry for aluminum and slide smoothly and freely in the bushings. While there is other equipment and test procedures for measuring material modulus this method validates the entire system with gauges and bars during an actual test firing. The extra gauges placed on the bars for this type of calibration can also be used as redundant signals during sample testing and are very helpful for monitoring the consistency and quality of your gauge setup. Strain gauges are consumable items. They can de-bond, burn out, tear off or loose connection during these dynamic events. Sometimes it is not completely evident that a gauge is compromised and this can lead to inaccurate data collection. Redundant gauges at multiple locations provide a level of confidence in your transmitted signals.

Q:

What sizes of striker bars does REL typically sell and what kind of velocities can be achieved with REL’s standard launchers?

A:

REL can make any striker bar length within reason. For a standard .75” diameter compression or tension system, 3, 6, 9, 12 & 18 inch lengths are common. Up to 36” single piece tension striker bars are often used and stacked together to create longer pulses. With long systems we use striker bars up to 6ft in length regularly. For direct impact or momentum transfer applications we often run up to 12ft striker bars.

The launchers for these systems will be capable of shooting small striker bars up to 200-300 ft/sec. They can actually launch at higher velocities than this but even with C350 bars you will begin to yield the bar material if you continuously impact much faster than this.

Q:

For very low signal levels what are my options for collecting high frequency Split Hopkinson Bar load data?

A:

This is a common problem for soft rubber or polymer materials and can be remedied in a few ways. Switching to aluminum bars of a smaller diameter will boost your sensitivity on weak materials but this is not often a good solution for large samples with high deflections. Another option is to use piezo load cells in place of a transmission bar. Units from Kistler and PCB are readily available and they make some nice forcelink tension/compression units in various sizes that mount to threaded SHPB bar at a reasonable price. None of them will respond accurately past 100khz though. They can have high fundamental frequencies close to 100kHz but usable bandwidth is only 5-20kHz. That being said we do use them a lot for polymer, foam and low speed testing. If you take a close look at your raw transmitted signals and run them through a range of filters you will likely find that 5-20kHz or even lower can be very reasonable for soft material testing. Even at higher rates we rarely consider anything past 80kHz or so from a transmitted bar strain gauge signal to be useful data.

Q:

I am using a set of steel bars with a full bridge setup, which is then input into a 2310B signal conditioning amplifier. I’m pretty sure I’ve hooked it all up correctly, and I can see some strain signals with a high enough amplification. However I am getting some very high frequency noise which is screwing with our measurements. The noise is at 370 kHz, it is a very distinct peak, my frequency resolution is 1Hz. I’m just wondering if you’ve had any experience with something of the such?

If you guys have no idea, could you give me a couple tips on wiring considerations? I can get a good signal through the amplifier when I have a full bridge on a polycarbonate sample, of course I’m getting a much larger strain with that, but there is no high frequency noise. I’m wondering if the metal SHPB rods are causing some type of ground loop or something.

A:

This issue sounds a lot like a grounding problem. It is definitely not in your signal. Everything at those frequencies would be attenuated by the vishay unit itself if it was.

I would make sure you have good grounding and shielding between your conditioner, oscilloscope and bars/stands. If the problem still persists start looking at other items that are plugged into the same power source. Devices such as compressors, welders, motors, transformers, power supplies, etc can cause noise in the system that can be difficult to remove until you find the problem unit and unplug it.

Email any questions to relinc@relinc.net


Q:

If you have steel bars, is there a unique need for aluminum bars as well? In other words, can steel bars only be used with a certain range of high strength/hardness materials or can they be used for the same materials aluminum bars are used for as well?

A:

There is an overlap in the capabilities of aluminum and steel but they do have unique applications.

For materials with high elongations, low strengths or lower strain rates aluminum bars allow you to generate pulses with larger displacements. You can run lower striker velocities with a higher sensitivity on transmitted load strain gauges. As you transition to higher strain rates and stronger samples the aluminum can get overworked. Higher strain rates require higher striker bar velocities which can plastically deform the ends of aluminum bars. Stronger samples also require more energy to fail and need longer/heavier strikers which can yield aluminum bars. At this point steel bars become more attractive. Maraging C350 has phenomenal strength to take high velocity impacts with lots of energy. However, it requires these conditions to generate usable strain signal levels due to the higher stiffness of steel.

As you’ve noticed I’ve used very qualitative terms to differentiate the two. There are no definite velocities, loads or strain rates that define their capabilities. It really depends a lot on each material response at a specific strain rate. We test a large variety of materials here at REL and do have some rules of thumb, but often we have to adjust our bar selections during calibration testing for each new material.

Q:

What is the energy consumption of the testing system?

A:

A 120Volt 15Amp service is sufficient for most systems to run a compressor and data acquisition.

Email any questions to relinc@relinc.net

Email any questions to relinc@relinc.net

REL Inc.
rel-logo-2