Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

- Community Forums
- :
- Forums
- :
- Hardware Development
- :
- Versal and UltraScale Architecture™
- :
- Modulating the adc/dac signal to separate from LO

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Mute
- Printer Friendly Page

Highlighted
##

francocapraro12

Visitor

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

07-08-2020 01:37 AM - edited 07-08-2020 01:51 AM

184 Views

Registered:
09-27-2018

Modulating the adc/dac signal to separate from LO

Hi. Im working with a AD9361 that use I/Q signals in FPGA.Im doing the OSC to separate the signal from the LO, after that i have a filter, then i have again the same OSC to put the signal in the same place of the ADC ( assumed attenuation for multiplying). The OSC is a DDS Compiler IP core from Xilinx. Im doing a Sine/Cosine of 8MHz. The scheme are this:

So if i want to do frequency translation i need a complex tone and cross multiply with both paths (I/Q). This will shift the spectrum.

How can i do the cross multiply ? Because if i do a normal muliply i got this spectrum (when i inyect a tone):

I draw an example of what im trying to do. If i use an OSC of 1Mhz and the LO is 419MHz :

In this case i have just 1 tone. But in real aplication i have several signals close to each other.

1 Reply

Highlighted
##

dgisselq

Scholar

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

07-14-2020 05:39 AM

71 Views

Registered:
05-21-2015

I can think of a couple of solutions. Perhaps the most common is to use a CORDIC. CORDICs are known for being able to rotate vectors around the complex plane without using any multiplies, although there is a scale factor inserted associated with the rotation. Often this operation is simplified so that the CORDIC just produces a sine and cosine, but the underlying operation is a full complex rotation. Further, even though they don't use multiplies, they aren't necessarily cheap when it comes to logic.

The next most obvious is to generate a sine and cosine pair and to perform the complex multiply requiring 4 DSPs. There's also a fun little trick you can use to drop this to 3 DSPs, as used and demonstrated within this FFT.

Dan