1
$\begingroup$

I want to use the AD7401 Sigma delta modulator connected to my FPGA.

The Datasheet pg.15 has some demo verilog code which I understand alright except one thing:

They use unsigned integers for all of the registers. Which means especially for the Integrators(that only get 0 or 1 added) that they will definately overflow at some point. The Integrators of CIC Filters are designed to overflow, but all the sources I found explicitly say that this requires twos complement to work properly. Am I missing something, or why does the Decimator work with unsigned Integrators aswell(that dont wrap around "nicely").

Also am I right that switching all the registers to signed, and mapping the IO Input (0,1) -> to First Integrator(-1,1) will fix this and result in a signed output.

$\endgroup$

1 Answer 1

1
$\begingroup$

Addition and subtraction in 2's complement or unsigned is the same. The only difference is for overflows.

Assuming 16-bit arithmetic, overflows occur in unsigned addition when you go over 0xFFFF (65535). For example, if you add 1 to 0xFFFF you will wrap around to 0.

In 2's complement, the overflow will technically occur when you go over 0x7FFF (32767). If you add 1 to 0x7FFF you will wrap around to 0x8000 (-32768).

Notice however that 0xFFFF + 0x0001 = 0x0000 and 0x7FFF + 0x001 = 0x8000 whether you're in 2's complement or unsigned.

Therefore, it does not matter whether it is implemented in 2's complement or unsigned, the behaviour will be the same.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.