Impact of LO spurs on heterodyne chain

Most of us know from customer specs the level of spurs are defined (in e.g. from a LO fractional-N synthesizer), but I would like to know if someone could provide any paper showing the impact of high spurs on a heterodyne chain.

We know, the LO spurs can be translated (in e.g. up-converted) to a rx chain and affect the phase error, or increase the BER, specially when it hits the non-linear elements such as DBM-mixer, so the mixer, as multiplier, in addition with other desired signals can cause intermodulation effects on it.

Any paper is welcome. Thks.

http://www.nt.tuwien.ac.at/fileadmin...beneder_da.pdf
5.4. The Frequency Calibration of HIFI:
http://www.ee.ucla.edu/~brweb/papers...es/RASIC96.pdf
http://www.analog.com/static/importe...es/AN-0996.pdf
http://drum.lib.umd.edu/bitstream/19...i-umd-4746.pdf

Thank you dear, I have read the articles which are quite good, however it does not bring the information I want. I am sorry, I was referring more specifically to the spurs caused by a fractional-N synthesizer.
I am looking for a relationship between the amplitude of the spurs and impact on the receiver / transmitter up/down conversion chain.
As we know, there are specs, for example, spurs lower than 60 dBc, spurs lower than 70 dBc, but the questions is why ? Why it does need to be at that level ? It must be a justification why -60 or -70 dBc are needed
It will offcourse impact on BER, phase error but what is the proportion between spurs in dBc <-> BER or spurs dBc <-> EVM, etc ?
Otherwise an article showing measurements for a certain type of modulation of the level of spurs in dBc versus BER or EVM would help !

Enrico

you can do the analisys in two ways:

1st: (inaccurate & fast), consider the spur as an additional noise to your local oscillator. For example, if the integrated phase noise is -40 dBc and spur level is -40 dBc, the equivalent noise can be considered -37 dBc. Apply this in all your system analisys for TX and RX.

2nd: (accurate but slow): the study has to be very detailed and depends on your system architecture, the frequency of the spurs, the modulation you use and a general rule cannot be given. At the end you have to consider (for RX) the SNR at demodulator (and the spur CAN affect it, as it is downconverted, filtered and processed in baseband as the RX signal) and (for TX) the impact of non ideal local oscillator to the modulation.
In this case the use of a system simulator is strongly suggested and sometimes the system design can require a similar effort of the transistor level design!

I hope it can help.

Mazz

Mazz,

Thanks for your reply, however I am not looking myself to do this analysis, a paper would save me this time !

Please reply in case you have any papers to share. This is what is needed.
Thanks