The op amp should have gain of 1000.
NE5534 is exactly 10Mhz and it has a good specs at all. It have been used in many commercial pulse induction detectors.
Ok thanks for the explanation on the pulse - that is exactly what we need to know, otherwise just writing mess
The bandwidth BW of an OPAMP depends on the Gain (called BANDWIDTH AT UNITY GAIN), thus 1000x gain and "10MHz bandwidth at GAIN=1" gives you 10kHz bandwidth at gain 1000.
10kHz BW means you can hardly see a 50us pulse.
BTW your gain is set to 500 based on the schematics..
In order to get gain=1000 with a few MHz BW you must use a series of say 3 of your opamps with gain=10 each (thus 1MHZ BW each and in total 1MHz BW, with total GAIN=1000). I doubt there is an OPAMP with GAIN=1000 and real bandwidth=10MHZ at that gain available on the market..
Generally, designing amplifiers with such high gains in a single stage is not recommended (for many reasons, especially DC coupled!!).
Also mind the input's OPAMP offsets will be multiplied by the GAIN too - you must compensate it (with R7 disconnected you have to set the output of the opamp to 0.0V). BTW the input offsets are temperature depended, a coin detected in your lab must not necessarily be detected at -5degC or +35degC in the field..
What you currently measure with your arduino is just an "average of something" at the OPAMP output, which may or may not indicate the golden nugget hidden beneath the coil..
PS: below a picture from the simulation:
1. Input -1mV to +1mV, 100us period, 50us pulse, rising/falling edge 1us
2. OPAMP is 70MHz BW, 1000V/usec slew rate, so by far better params than yours
4. See the output offset +1V due to high gain - must be compensated in your design..
5. The 5pF in your schematics decreases the bandwidth as it creates an low-pass, the corner freq will be aprox 1/(1000k * 5pF) = 200kHz, but it helps with stability a bit..