OK, I've done it another way: injecting a signal with high frequency, low amplitude noise (around 4mVpp), managed to narrow down the mighty 'HOLE 0x5ff 1535', with help of a little program in C,
ONE MILLION SAMPLES.
Basically there are no errors, nothing lost, just the samples are sticking to the adjacent levels.
That means an error of 1-2 ADU (Analog-Digital Unit), better than expected.
So, it's not a black hole, it's not a supernova candidate, it's just a small glitch, a smaller brick in the wall ... I expect the rest to be similar ... almost nothing.
It's a good reason to overclock the ADC like ...10x.
Maybe I will get some errors![:)]( "Smile")
ONE MILLION SAMPLES.
Basically there are no errors, nothing lost, just the samples are sticking to the adjacent levels.
That means an error of 1-2 ADU (Analog-Digital Unit), better than expected.
So, it's not a black hole, it's not a supernova candidate, it's just a small glitch, a smaller brick in the wall ... I expect the rest to be similar ... almost nothing.
It's a good reason to overclock the ADC like ...10x.
Maybe I will get some errors
Code:
Sampling number 1000000Level < 0x5fa 1530 : 0Level = 0x5fa 1530 : 0Level = 0x5fb 1531 : 1370Level = 0x5fc 1532 : 127729Level = 0x5fd 1533 : 557522Level = 0x5fe 1534 : 48374Level = 0x5ff 1535 : 3Level = 0x600 1536 : 257297Level = 0x601 1537 : 7699Level = 0x602 1538 : 5Level = 0x603 1539 : 0Level > 0x603 1539 : 1Total valid samples = 1000000Code:
Sampling number 1000000Level < 0x5fa 1530 : 0Level = 0x5fa 1530 : 0Level = 0x5fb 1531 : 4Level = 0x5fc 1532 : 3598Level = 0x5fd 1533 : 245024Level = 0x5fe 1534 : 92273Level = 0x5ff 1535 : 1Level = 0x600 1536 : 630893Level = 0x601 1537 : 28194Level = 0x602 1538 : 11Level = 0x603 1539 : 1Level > 0x603 1539 : 1Total valid samples = 1000000Statistics: Posted by gmx — Tue Oct 01, 2024 8:59 am