Re: [Discuss-gnuradio] mc4020 + microtune -- ValueError: source and destination data sizes are different

[Eric Blossom]

On Wed, Jan 26, 2005 at 07:42:42PM -0500, Ilia Mirkin wrote:
> Jamie and I are working on this together, but for the benefit of the
> list, here is what the problem was:
> 
> (after familiarizing myself with the gnuradio internals and looking at a
> bunch of other examples)
> 
> fftsink takes a complex. the mc4020 puts out a short. there is a
> short_to_float conversion going on (that's the "cvt" in the code), and
> float != complex, hence the ValueError.

As Matt pointed out, there's a make_fft_sink_f and make_fft_sink_c

A quick look at fftsink.py reveals this:

# ========================================================================
# returns (block, win).
#   block requires a single input stream of float
#   win is a subclass of wxWindow

def make_fft_sink_f (fg, parent, label, fft_size, input_rate):

[snip]

# ========================================================================
# returns (block, win).
#   block requires a single input stream of gr_complex
#   win is a subclass of wxWindow

def make_fft_sink_c (fg, parent, label, fft_size, input_rate):

[snip]


use make_fft_sink_f once you've got a stream of floats...


> BTW, my port of the mc4020 driver to 2.6 causes massive memory
> corruption - seems to make python croak a lot of the time, as well as
> gcc. If anyone has something that works better than that, please let me
> know or post a link.

Meenal, can you give us any update on on the mc4020 patch for 2.6?

Eric