Of course, the original intention was to always use this with PAUT and/or FMC. We want imaging. Not having a receiver, the T-800 only works in Dual Mode, or through transmission. At first, we weren’t really sure we could get an image having only a monolithic probe to transmit with and was told it most likely wouldn’t work. So, we wired up a test array and called it the Frankenstein probe.
This was just an old array that my friends at SNI had laying around and wired it up accordingly. It was 7.5MHz in a 4×8 configuration, not ideal, but we weren’t going to destroy anything in the process.
Here it is doing an element check so far, so good.
Running out of time I didn’t really have a chance to tweak and tune like I needed, so I loaded some wedge parameters, and hooked up a ¼” round for the transmitter. Pretty dodgy work actually, but it served its purpose and revealed the world’s ugliest side drilled hole. Using a standard Delay and Sum TFM algorithm.
And now the ugliest hole:
Not pretty for sure, but it served its purpose, and the theory was working.
This was so encouraging that we decided to make a purpose-built array. After all we were targeting heavy wall Cast Austenitic Stainless Steel.
It’s understandable that when we see a dual probe set-up, we are used to seeing probes that are mirror images. The fact is there is a lot of tweaking that a person could do here. The array that we built was a 64 element 25mm x 64mm in 0.5MHz. so not exactly the same size as the monolithic transmitter.
Here are the active crystal shapes that we ended up with (to scale). As you can see, they are definitely not mirror images of themselves, and we’re wondering why we ever thought they had to be.
The obvious reason for the monolithic transmitter is that we can hit it with 800V. You simply cannot do this with an array and expect it to survive. There are reasons that array systems have relatively weak pulsers in them. All good reasons, and there are several. To begin with, most if not all, are spike pulsers. A spike pulser cannot carry the same energy content that Negative Square wave can. The T-800 is only Negative Square wave.
Quick note on pulsers:
A square wave pulser is basically a switch that connects and disconnects a voltage supply to the transducer. It is easier to get large energy content with a square wave than with a spike pulser.
Here is a square wave pulse (200Volts into 200Ω). The pulser pulls down to -200V and holds it there for a programmable amount of time, in this case 500nS, and then let’s go.
This is a 200Volt spike pulser. Spike pulsers charge a capacitor to a given voltage and then discharge it to the transducer. It pulses low and immediately returns towards zero. The width of the pulse can be controlled somewhat by component selection but the waveshape always looks like this.
Some Results:
Sinewave was fortunate to get its hands on an old CASS sample courtesy of PNNL and the US NRC. This sample was perfect as it had two different types of CASS on either side of a weld and was a segment of 28” at 2.4” thick. Equiaxed grain structure on one side, and columnar grain on the other.
To understand what we are looking at here is a picture of the block:
CASS material is notoriously horrible to look at. Its very noisy, attenuative, and anisotropic in nature. Large low frequency probes are the key to having any hope of a successful exam.
Here is an example of frequency looking at a 2.5” deep side drilled hole.
Its worth noting that the 1MHz probe had a focal depth of 2.75”, by all accounts it should have done better. It was found, a long time ago, that at least 3:1 SNR is required for any hope of success.
Here are some images of our block: notice the gain difference. We expected the T-800 to be colder on the order of 10 to 14dB versus using the native pulser, but this shows 34db difference on an ID notch. After further investigation we found that there was a grounding problem with the probe that the native pulser didn’t like. Come to find out all manufacture’s equipment is different, and after grounding out the probe to the receiver probe it settled into a more normal state. (native left, T-800 right)
Here is a comparison between the native pulser and the T-800 in a Equiaxed vs Columnar grain scenario.
As you can tell by the images, we are saving a ton of gain using the T-800. Even with the probe grounded for the native pulser, it’s a huge improvement. Additionally, the flaw (ID Notch) position seemed to change slightly with the different looks. From what I understand about CASS it makes sense.
Here are a couple of shots through the weld to the notch on the far side.
| 160V native pulser | T-800 at 800V |
Dealing with Noisy material requires a well thought out inspection plan. We cannot change the material itself, but we can definitely improve results using the right tools.