Re: Integrate block functionality

[Marcus D. Leech]

On 12/20/2019 10:39 AM, Joe K wrote:
> Hi all,
>
> Maybe this is a dumb question, but I didn't find any discussion on the 
> web or in the mailing list archives.  The integrate block (e.g., 
> blocks.integrate_ff) doesn't seem to operate as I'd expect.
>
> What I'm looking for is a block that integrates from time 0 with no 
> decimation.  So, an input stream of [1, 2, 3, 4, 5] would output [1, 
> 3, 6, 10, 15], etc.  A constant input K would generate a ramp with 
> slope K.  That's what I thought the integrate_ff block, with 
> decimation=1, would do, but it does not.
>
> The integrate_x blocks seem to only integrate over the previous N 
> samples, where N = decimation.  Therefore, with decimation=1, the 
> integrate_x blocks simply relay the input to the output.  If so... 
> what is the point of that?
>
> Is there any block that accomplishes a true integration as described 
> above?  It would be a quick block to implement in an OOT module but it 
> seems like this is a basic function.  Also, is the operation of the 
> integrate_x blocks intended to be how it is?  If so, it seems misnamed.
>
> Thanks,
> Joe
A IIR filter with Alpha and Beta both set to 1.0 will accomplish this.  
Although, in real life, most people using integration functions in DSP want
  something more like what the integrate block does, since without a 
"dump" type operation, the integrator output will "rail" pretty quickly.