That’s a major problem. Sorry.
And when one uses a nickel and a quarter and watches what happens that is major too.
When a detector fails something basic that is major problem.
Fails in both ID and Tone departments.
Like the detector is being locked up. And when it unlocks it locks back up. With targets sitting substantial distances apart.
I understand your concern.
Target identification is a very difficult thing to do on a reflected electromagnetic wave signal.
Besides the composition of the metals themselves. You have the mineralization of the ground. They speed at which the coil is moving across the target. The distances. The angle of approach. The vector of approach. The 3-dimensional position of the target. The frequency of the detector. The coil dimensions.
I don't know the garret metal detectors. But I assume the reflected electromagnetic analog signal is converted to a digital signal for analysis. So that signal will end up looking something like this.
Now their chip/software needs to classify that target as to what it is. Many of this these digital targets will look identical to our eyes. But their logic has to classify that to a known target. Not a perfect science. They will look similar and sometimes get mis-classified.
I tested over 30 thousand gold target samples last year, using different sizes at different approaches, distances, frequencies, etc. I could never get 100% accuracy. I don't think any system can. There is just so many factors involved that will cause two different targets to look very similar.
I personally, don't tout any metal detector brand. But I ask that you not base your decision on one test. I think if a metal detector can get 85+% accuracy on classifying, it did really good. There is a lot involved in classifying targets.
85% is a B+, which is good in my book.
Just my humble opinion.
Kind regards