Forum
Please or Register to create posts and topics.

Raw IQ File Writer

Page 1 of 2Next

Raw IQ File Writer block:

RAW IQ File Writer Block

Left hand side input:

  • IQ Streams

The Raw IQ File Writer records IQ streams to the selected file in the common "raw" IQ format of interleaved complex 32-bit floating point values. For a sequence x_0, x_1, x_2, ... of complex values, these are written to disk as follows:

[Real(x_0), Imag(x_0), Real(x_1), Imag(x_1), Real(x_2), Imag(x_2), ... ]

Make sure you have a capable SSD since the data amount can be huge! You might consider to use the IQ Demodulator to reduce the data size.

To start a recording first select a target file using the file chooser, then press record to start a recording.

What is the difference between Raw IQ File Writer block with File Writer block?(File Writer – Aaronia SPECTRAN V6 Support Forum

As mentioned before it saves RAW data only without any additional information.

The Raw IQ File Writer records the IQ values into an .iq file.

Further there is also an .xml file issued with some information about the recording itself.

In the screenshot of an .xml file below a clock unit of 179999.994946 Hz is stated.

Does this mean that at the given center frequency of 331.7 MHz the signal was sampled only with a sample rate of app. 180 kHz or with 180 kilo samples per sec?

Could this be reached by selecting an IQ Rate of 92 MHz and a Span of 1/512?

The valid frequency range after the FFT would then be 90 kHz or +/- 45 kHz around the center frequency?

Is my understand correct? Sorry, still struggeling with some basics ...

Quote from Tom on 12/01/2024, 21:31

In the screenshot of an .xml file below a clock unit of 179999.994946 Hz is stated.

Does this mean that at the given center frequency of 331.7 MHz the signal was sampled only with a sample rate of app. 180 kHz or with 180 kilo samples per sec?

Could this be reached by selecting an IQ Rate of 92 MHz and a Span of 1/512?

Correct. "Clock" is the actual effective sample rate. "92 MHz" sample rate is actually 92.16 MHz, divided by 512 results in 180 kHz.

Quote from Tom on 12/01/2024, 21:31

The valid frequency range after the FFT would then be 90 kHz or +/- 45 kHz around the center frequency?

It would be +-90 kHz around the center frequency.

OK, thank you so much.

With the given centre frequency of 331.7 MHz and the selected span of 1/512, a frequency span of 331.7/512 = 0.648 MHz or +/- 0.324 MHz should be achieved.

This means that a frequency range from 331.376 to 332.024 MHz would be relevant, resulting in a downconverted or intermediate frequency range (in the best case) from 0.0 to 0.648 MHz or from 0.0 to 648 kHz.

But how does the 180 kHz sampling rate fit with the maximum frequency of 648 kHz?

Normally (according to Nyquist) the minimum sampling rate in this case should be twice 648 kHz = 1296 kHz instead of 180 kHz as stated in the .xml file above.

You misunderstood. The sample rate has nothing to do with the center frequency. The span factor always refers to the selected IQ sample rate (92, 122, 184 or 245 MHz). So in your case you get a frequency range of 331.7 MHz +- 90 kHz (331.61 MHz to 331.79 MHz).

Yes, you are absolutely right. The sampling rate of course has nothing to do with the centre frequency.

But I am not that sure if the sample rate of 180 kHz given in the .xml file corresponds 1:1 to the FFT frequency span of +- 90 kHz.

The RTSA software shows the following relationship when switching between the IQ rates @ 1/512 span:

IQ rates: 92, 122 and 184 MHz
Resulting 1/512 spans: 149.77, 199.69 and 299.53 kHz

The screenshot below shows a selected IQ rate of 92 MHz with a resulting 1/512 span of 149.77 kHz.

According to Nyquist, for a frequency range of 0 to 149.77 kHz (downconverted) to be analysed in this case, a minimum sampling rate (clock) of twice 149.77 kHz or 299.54 kHz is required, which does not seem to match the 180 kHz specified in the .xml file above.

As we need complex values in the frequency domain, we need to process the FFT externally and therefore we need to be confident that the values in the .xml files are correct.

The RTSA by default does not use the full IQ sample rate of the Spectran V6, but ignores the edges for accuracy reasons. This reduces the effective bandwidth by about 15%. You can disable that behavior by enabling the "Full Spectrum" option in the Spectran V6 block.

Not 100% sure if the Raw IQ File Writer block ignores that setting or indeed misreports the sample rate.

Not sure that i understand what is the IQ  structure

Is it 16 bits I0 , 16 bits Q0 , 16 bits I1 , 16bits Q1 ...... ?

And also what is the full scale ? 2^15 ?

Page 1 of 2Next