HDL QPSK Transmitter and Receiver CFO problem

Hi

Hope you are all doing well.

I am working on QPSK design mentioned details in the given link

https://www.mathworks.com/help/comm/ug/hdlqpsktransmitterreceiver.html

when i decrease the Rsym rate from 1.92MHz to 50KHz it is working well in the simulation but when i decrese the Rsym rate less than 50KHz upto 5KHz then i am not able to reciever my sent signal with carrier frequency offset (CFO) 100KHz in the channel. I think that is Carrier frequency offset problem because when i decrease the CFO then i am able to recieve the signal correctly.

can anyone help me to solve that issue how i will recieve my TX signal at CFO =100KHz?

Answer 1

[Kranti Balaga]

Hi Ali Shan,

In QPSK single carrier model, the carrier frequency estimates in steps of coarse and fine estimates which are depends on the symbol rate. The document clearly mentioned the ranges of the estimated CFO vs Symbol rates. You can see Frequency and Time Synchronizer section

"The frequency estimator estimates a normalized frequency (with respect to symbol rate) range of -0.125 to 0.125 which corresponds to a frequency offset range of -240 KHz to 240 KHz ...."

Coarse cfo ranges +-240KHz and residual fine frequency range +-3KHz @ 1.92 MHz i.e., (coarse:1.92*0.125 MHz and fine: 1.92*0.0016 MHz)

Your case,

coarse CFO range 6.25KHz and fine freq range 80Hz @ 50 KHz

coarse CFO range 0.625KHz and fine freq range 8Hz @ 5 KHz

so it will not work for 100 KHz CFO.

The Implemetation of single carrier frequency estimatation algorithms depends on the symbol rate that you choose.

Thanks,

Kranti

[ali shan]

Hi Kranti Balaga

Thanks for your kind information.

Is it possible for us to change the frequency estimator with a normalized frequency (with respect to symbol rate) range of -0.125 to 0.125 and also the residual frequency offset after coarse frequency offset correction is with in the normalized frequency range of -0.0016 to 0.0016 to achieve our desired result for low Symbol rate? And also one thing I forget to mentioned that in simulation i mentioned the Channel CFO 100KHz but that is not the exact channel CFO during hardware testing because that depend on the Real Channel environment condition.

Sir,if you know any other possibility to achive our deired result on low Symbol rate then kindly let me know please.

Thanks in advance

regards

Ali Shan

[Kranti Balaga]

Hi Ali Shan,

As long as the cfo impairment are within the bounds as per the document, you can apply. If the CFO is beyond the limits, the model unable correct those drifts. Our coarse algorithm can find it up to 0.125*Rsym, which is the max range of our estimation algorithm.

Also, you should check, if the symbol rates are very low and the frequency offsets are more like 100 KHz, you may get inteference and may not able to correct beyond.

I suggested to check your requirements of typical ranges of CFOs to test.

Thanks,

Kranti

[ali shan]

Hi Kranti,

Thanks for your quick response.

Can you tell me about where i will find the 0.125 value in Coarse frequency compenstor ? And also tell me the purpose of 2^15 consecutive estimates in the Integrator .

Answer 2

[Kranti Balaga]

Hi Ali Shan,

The value 0.125 implied in the algorithm that we choose. Implemented correlation based estimation, as the received symbols are QPSK, raise the input signal to the 4'th power to eliminate the phase mapping dependency, then correlate and find the angle. The derived equations will yield the estimation which limits to 0.125 times symbol rate.

Integrate the 2^15 estimates and average across for better estimate. It's an averaging length and an ephirical value based on the simulation. You change as per your requirement.

Thanks,

Kranti

[ali shan]

Hi Kranti Balaga

hope you are doing well.

thanks for your response.kindly let me know if i will change the average length for better estimation then is it possible for us to achieve or recover the recieved signal with with minumum frequency offset like CFO=100khz or less?

Answer 3

[Kranti Balaga]

Hi Ali Shan,

I am fine, hope you are doing well.

The more average length gives a better/accurate estimate, but that doesn't mean estimating higher carrier drifts.

The model can estimate 100Khz, but the symbol rates should be above the defined range.

I hope, I answered your question.

Thanks

Kranti

[ali shan]

thanks for your kind response.

Actually we want to reduce symbol rate aupto 5kbps.Can we modify this design somehow to achieve this symbol rate or not?If so,then what are those customize changes required in the receiver side or we need to keep in consideration or the limiting factors to work on, in the algorithm

Answer 4

[Kranti Balaga]

Yes, simply you can provide the Symbol rate in the Inputdata mask, change the Carrier Frequency Offset (in Hz) corresponding to that symbol rate (supported ranges) in the Channel block mask and run the model, it works. You can observe the estimated CFO value.

The synchronization algorithms in the receiver design, work for 4 times the symbol rate where we introduced the impairments in the channel. If you want to estimate out of the range, you may need to introduce cfo with more sampling rates, run the synchronization algorithms at a higher rate and correspondingly you may need to modify interpolation filters in symbol synchronizer, and L&R algorithms in Coarse Frequency Estimator. As a whole, you need to redesign synchronization algorithms.

In general there is always a limit on the supported CFO (normalized with fs/2 i.e., value between +-1) with the symbol rates otherwise we may get aliasing.

Thanks,

Kranti

[ali shan]

Can you please provide any reference book or any other helping source link for detailed study of algorithm implemented in the given design about symbol synchronizer and L&R algorithms, other than the two books mentioned in reference link:-

1. Michael Rice, Digital Communications - A Discrete-Time Approach, Prentice Hall, April 2008.

2. Floyd M.Gardner, Phaselock Techniques, Third Edition, John Wiley & Sons, Inc., 2005

[Kranti Balaga]

Yes, we only referred the above two books, which are mentioned in the example document.