A reliable solution for EMI noise suppression
NDE equipment is frequently used in industrial and noisy environments. By far, impulsive EMI noise is more harmful than white or coloured noise since the filter response is indistinguishable from flaw indications. Operators ignore EMI spikes because they appear randomly in the A-scan screen.
However, in automatic inspections and, specially, when performing C or D scans, which are based on the signal amplitude or peak position, impulsive EMI noise can completely hidden the weakest signals, reducing the effective dynamic range of the system and the possibility of detecting flaws below the noise level. An effective EMI noise cancellation filter is included at the beginning of the processing chain. It is based on the principle that, in a small set of consecutive acquisitions taken a short time interval apart, noise spikes appear at random positions while flaw indications remain at the same location. To this purpose, an order statistics non-linear filtering process applied to 2 or 3 consecutive A-scans is followed.
Fig. 1 – EMI noise reduction principle
The architecture for the EMI filter is shown in Fig. 1. The process starts with the EMI-filter disabled to store a single A-scan in the FIFO. Then, the subsequent A-scans are filtered with partial results stored in the FIFO. The EMI filtering hardware is independent of the number M of A-scans processed but, in practice, there is little gain if M exceeds 3 or 4. Moreover, since this is a non-linear filtering function, there is a risk of destroying useful information if this figure is made too large.
Fig. 2. EMI filter architecture
The EMI noise cancellation is especially effective for A, B, C and D scans as shown in Fig. 3. Here, the B-scan of an aluminium test-piece with several flat bottom holes is obtained in a highly noisy environment. The effect of the EMI filter is apparent.
Fig. 3 – B-scan of an aluminium test-piece with several FBH. Top: EMI filter disabled, Bottom: EMI filter enabled. Transducer 10 MHz.
The EMI Filter is available in all of our ultrasound systems, including single-channel, multi-channel, TOFD and phased-array. It is hardware implemented, and hence, it works in real-time without any reduction in acquisition rate.
Read more at:
- A Full Featured Ultrasound NDE System in a Standard FPGA