First of all, if you have metallic bars, dissipation is normally negligible. There are basically two things you need to do for metallic bars:

- measure the true speed of sound in the bars: this is done by performing a single bar experiment and measuring the delay for a wave to travel hence and forth in the bar (well, normally to the middle of the bar, as the strain gages are typically at that location)

- checking proper alignment and friction effects by doing a bar-to-bar experiment

These are the only things that need to be done for metallic bars (there is no 'callibration correction' necessary for the actual measurement data).

If you have polymer bars, then it becomes more complicated as you will have to take into account dissipation (i.e. attenuation) of the wave inside the bars. This is done best by using data coming from several strain gage locations on the bar. A lot of information on this can be found in the proceedings of the Dymat association (www.dymat.org). As posted previously, the ASM handbooks also have a chapter dedicated to SHPB

On a metallic bar, the strain was measured by the mean of strain gauges glued at 1 m from the surfaces where the concrete samples were glued. It was then possible characterizing the incident and transmitted stresses waves via the knowledge of the Young's modulus of the Hopkinson bar and the wave velocity. It was then possible to plot Stress / Strain curves after having measured the strain directly on the concrete sample by the mean of strain gauges.

Reference:

F. Toutlemonde, C. Boulay, J. Sercombe, F. Le Maou, S. Renwez., R. Adeline, Characterization of reactive powder concrete (RPC) in direct tension at medium to high loading rates, 2nd Int. conf. on concrete under severe conditions: environment and loading, CONSEC’98, 1998, June 21-24, Tromsø, Norway.

Hello all,

First, thank you for your contributions but doesn't calibration means expressing strain according to the recorded voltage of gauge for both bars.

Also, i want to know which material is suitable for bars to test Epoxy/glass composite.

Best regards.

Your spec sheet for the strain gages should give you a calibration factor. If not, then you're in trouble! Concerning the epoxy/glass composite: aluminum bars should do well. The difficulty will be in attaching the samples to the bars. In my opinion, glueing them works best.

Yes you're right, Frederik. We took the gauge factor for a coarse approach of the sensitivity of the aluminium bar (Ø 74 mm). We also verified the bar sensitivity with weights at the bottom of the equipment (two half bars of approx. 5 m each). The concrete samples (Ø 74 mm, 100 mm in height) were glued between the two bars with a methyl-metacrylate resin. The ratio between the poisson ratio and the Young's modulus of both the concrete and the aluminium are approximately similar, so that transversal strains are almost identical. It means that shear stresses are minimum in the glue between aluminium and concrete.

Aluminum bars are just what i had used, Frederik. For HBM gages, you are right i had received a spec sheet with the pack but i want to know how to get this calibration factor. In some manuscripts, authors carried out a compression test for a specimen, made from the same material as the bar's one, on which they had glued a strain gage. By comparing voltage-time signal recorded by the gage and stain-time curve given by the machine, they can get a relationship between Strain and voltage. This latter isn't necessarily linear due to type of used bridge. Regarding the attachment of samples, what about using a thin film of grease at interfaces. The thin film can maintain the specimen but it must be waves transparent .

- I apologize for my English because I'm not a native English.

Concerning the sensitivity, I attached a file explaining the calculations.

The relations are linear for elastic behaviors. Concerning the fixation of the sample I'm not sure to understand your meaning. If you work in compression, may be a layer of oil could work. In tension, only glued samples are required.

Thank you Boulay for the attached file, yes i'm working in compression.

About your posted article Gerlach, why high strength composite materials require triangular pulse shapes and by what means we may have them.

@ Larbi: actually it aren't 'high strength' composite materials that require a triangular pulse, but rather 'brittle' materials. As the SHPB (Split Hopkinson Pressure Bars) have problems in getting accurate results in (typically) the first 5-6% of deformation (due to the Pochammer-Cree oscillations and time to get force equilibrium), you will get unusable results if you have a composite that has a maximum failure strain of a few percent, as the sample will break before having force equilibrium. This problem can be circumvented by having a longer rise time for your loading pulse by adding a 'wave shaper' between the striker and the impact bar (typically a small, soft copper disk).

What about wave shaper dimensions and material (why exactly soft copper disk).

For exact wave shaper you will have to look through the relevant literature. Soft copper is used because it is soft (if the wave shaper is too strong, the pulse will not be altered i.e smeared out enough), but its acoustic impedance is still close enough to the metallic materials typically used in SHPB (otherwise you would simply crush the wave shaper without any relevant interaction).

For a calibration table you can simply move the input bar in touch with the output bar. As soon as your pulse is usually longer than 300 microseconds, all non-equlibriums (on grease or wave shaper) affect the voltage on gauges only for a less than a microsecond. However, if you need in your table the ramping section of the input pulse, I believe, you need to use the same conditions as will be in your real experiments. If you wish to provide an 'ideal' contact between the rods, you have to keep in mind that it is impossible neither in the case of direct contact of input and output bars or in the case of the bars with a sample. You cannot calibrate the time component of your table anyway, because it involves the sound velocities in a sample that is different from those in the case you put a steel (aluminum) sample instead of a real one.

The article, "Evolution of Specimen Strain Rate in Split Hopkinson Bar Test," by H Shin and J-B Kim reports a method of validating the experimental result of the split Hopkinson bar. It cal also be used for the calibration of the instrument. The article can be downloaded at: https://doi.org/10.1177/0954406218813386

This conversation is helping me a lot in caliberation of SHPB,

please continue.

Dear Larbi,

To achieve a high strain rate, the split Hopkinson pressure bar (SHPB) technique has proven suitable. For high strain rate, it need more developing the split Hopkinson pressure bar (SHPB) technique. Earlier it was considered, 1D model is not much suitable is insufficient to the geometry of the cross section, so an approximate 3D wave propagation model for bars with arbitrary cross-section can be developed using Hamilton’s principle. 3D model provides a dispersion relation (wavenumber vs. angular frequency) and average values for displacements and stresses over the bar/specimen interfaces.

DFT technique which is shown to be suitable for calibration of the experimental set-up, for the technique of redundant measurements and for the evaluation of SHPB tests. A calibration procedure for the experimental set-up using an ordinary SHPB projectile and an optimizing procedure can be developed using wave propagation model. The optimizing procedure can be used to find Young’s modulus and the axial positions of the measuring sections to optimize all of the parameters in the dispersion relation. Poisson’s ratio is determined first. Estimation of poisson’s ratio, young modulus, redundant measurement, comparing possible options when evaluating SHPB tests would be essential elements. An advantage of both procedures (dispersion relation and poisson’s ratio) is that the actual bars to be used in the SHPB tests can be also used in the calibration, i.e., for this no special test-pieces of the bars have to be manufactured and testing required. The procedures are independent of the use of the 1D or the 3D formulation, if a few Fourier components with low frequency and high energy content are used solely when estimating Poisson’s ratio. From the SPHB tests, it was observed that two-pulse method gives relatively good results for both dispersive and non-dispersive wave propagation, with only small oscillations appearing on the stress-strain curve. It shows that the two-pulse method is rather insensitive to errors in the calibration and suggests that the two-pulse method is robust and may often give comparatively good results. The three-pulse method in combination with non-dispersive wave propagation gives unsatisfactory results not suitable. The method of mixing the 1D formulation with the dispersion relation obtained from the 3D formulation gave the same results as using the complete 3D formulation. This indicates that the effects of dispersion are much more pronounced than the effects of warping of the bar cross-section in the tests considered here.

Reference:

1. Al-Mousawi M.M, Reid S.R. and Deans W.F. (1997). The use of the split Hopkinson pressure bar techniques in high strain rate materials testing. Proceedings of the Institution of Mechanical Engineers 211(C): 273-292

2. Gong J.C, Malvern L.E. and Jenkins D.A. (1990). Dispersion investigation in the split Hopkinson pressure bar. Journal of Engineering Materials and Technology 112: 309-314

3. https://www.diva-portal.org/smash/get/diva2:162134/FULLTEXT01.pdf