Matlab QPSK exmaple for Adalm Pluto

Greetings everyone!

I have encountered a problem which I have tried to solve for a long time but eventually failed.

I think that it is due to my lack of knowledge for signal processing.

I have used a single Adalm Pluto module as a Transceiver and two of them as a Receiver and a Transmitter.

I am working within Simulink under Ubuntu.

As for the learning process - I have tried to launch examples for QPSK in Communication Toolbox Simulink.

In both cases I do not see the parcel as it is described in the Receiver QPSK example. If I comment both the Pluto Tr module and Pluto Rx module and than connect QPSK transmitter with QPSK Receiver, I do see a good picture on the Constelation Diagram and I do see all messages as they should be (Hello World 000, Hello World 001 etc). When I am using Pluto modules - I am not getting anything at all, and I see garbage on the constelation diagrams (I do see that something is going on there but it looks bad).

Could anyone help me with this please?

P.S.

I have tried to change Pluto Modules parameters but I did not help.

Answer 1

[Alexey Petrenko]

Finally I got a result.

Following through Kevin's advices and my own ideas I ended up having all parcels got by the receiver.

The correct way was to lower down the sampling rate withtin pluto TX and RX down to 100k samples.

Although it is totally not the end because as I understand it - I should be able to receive up to 20MHz which is by definition pluto's rf band.

[Kevin]

This is great news. Good to know that finally it is working perfectly for you.

You said that lower sampling works (but not higher sampling rate). This seems to point to bandwidth of your signal. I am guessing you are sending your signal over the air. Try using cable connection and see if this problem goes away. I think it is the wireless channel causing you problem.

Have fun!!!

Answer 2

[Kevin]

Hi Alexey Petrenko,

It can be tricky debugging hardware problem.

So let's step back and do something really simple.

Can you use the transmitter pluto to send a single tone and use the receiver pluto to receive it? Plot the received signal in time domain to make sure that the signal is not too large or too small (use the tx gain and rx gain to adjust). Also plot the received signal in frequency domain to see if you get a tone.

There are a few other things we can try. But let's do this first and see what we get.

Kevin

[Alexey Petrenko]

Greetings Kevin!

Thank you very much for your feedback.

As for this first step - I have successfully used Pluto for transmitting audio signals using audio toolbox already - made kind of a radio voice repiter:) So the problem as I see it is really in signal processing.

Thanks in advance,

Alex

[Alexey Petrenko]

Dear Kevin!

Following your advices dedicated to frequency deviation, I have tried to manually change the Rx carrier.

As a result I have seen a little difference in const. diagrams but inspite of this good result I am catching this "bursts" anyway.

I have tried to do something with Tx and Rx Pluto parameters but it did not help at all.

So now it's all about this bursts I am catching and I have no idea of thier nature...

Maybe it has something to do with me using only one Pluto? Maybe I should use two of them - one for Rx and one for Tx?

Regards,

Alex

[Kevin]

Hi Alex,

Yes, this can be something related to using just one pluto, especially (I think) changing rx center frequency will also change tx center frequency.

So let try the experiment with two Plutos.

There is also one thing we should be careful. When using one Pluto to transmit, I am assuming that you are using transmitRepeat, then you ideally the signal should not have glitch between the end of the signal and the beginning of the signal. If there is glitch, it may (or may not) cause problem to the receiver recovering the bits. This is something you need to keep in mind. But for now, assuming that this is not a problem.

Kevin

Answer 3

[Kevin]

I see. OK. Let's go back to your QPSK example.

Do you use a cable to connect transmitter and receiver? If not, can you try this first?

[Alexey Petrenko]

Greetings Kevin - seems that I have made a global comment, please see it below:)

[Alexey Petrenko]

Still I do not get all the messages clearly. I am trying do something with damping factor of symbol and carrier sync and with the normalised loop bandwidth but still the effect I get is questionable... I mean that I am not sure that tuning of those parameters does anything at all.

[Alexey Petrenko]

Dear Kevin!

Many thanks for your responses and help.

Seems that I finally got it - I have got 100% results dealing with sample rate. For some reason 400k sampling rate is big in the way that I am having the result I presented to you later. Lowering sample rate to 100k gave me 100% of parcels coming to be correct.

I will look forward to deal with it later anyway.

Thank you very much!

Answer 4

[Alexey Petrenko]

Greetings Kevin!

I have tried the coax connection earlier but the results on const diagrams were totally the same as it was with antenna connection.

So now I am using antennas.

Also I have tried to manage rx parameters after square root block and was able to get a result :) so now I am getting a part of a parcel or a whole parcel from time to time. I have changed the timing error detection parameter within symbol synchronizer block to zero crossing now I got something like this (see pic).

[Kevin]

There is one more thing to check: the frequency offset between the transmitter pluto and receiver pluto. The frequency offset (or error) on one pluto can be as large as 25 ppm (i.e. 25 * 1e-6 * center frequency). In the worst case scenario, the frequency error between tx and rx = 2 * 25 * 1e-6 * center frequency.

I don't remember how much frequency offset the example matlab receiver can handle. But you can manually frequency shift the received signal to see if you get better results (this would involve some trial and error). Or you can use tx a tone and look at the received signal to estimate the frequency offset.

Hope this works.

[Alexey Petrenko]

Greetings Kevin!

Happy New Year to you!

I have tried to manually change the central frequency of the Pluto RX, seems that somehow it helps - I will rely on more accurate results in a few days.

Thanks in advance!

[Winfried Johannes]

Hello,

I have a similar problem.

I use the Matlab QPSK example for the data transmission between two plutos. If I use the internal clock for both Plutos it works very good with a cable connection between transmitter and receiver.

Now I use an external clock (40 MHz) for the transmitter Pluto and the internal clock for the receiver Pluto. Again with a cable connection between both Plutos. The result is, that the error rate is very high and the constellation is bad. I believe it is a synchronisation problem.

Do anybody tried similar things? Any ideas?

I tried the Matlab Frequency Correction Example (Transmitter Pluto with external clock connected with a cable to the receiving Pluto with internal clock). The result is attached. I see two strong tones (60 khz and 80 khz) instead of three

Do anybody know, what the problem is?

Thanks

Winfried

[Alexey Petrenko]

Greetings Winfried!

I am no guru as you have probably noticed from my other posts, but if I may suggest - try to lower down sample rate as I did and try to manually adjust the carrier within your PlutoRX.

Hope that helps - please feel free to tell of your results.

Regards,

Alex

[Winfried Johannes]

hello. thanks for your help. it works now. i adjusted the clock of the adalm pluto.

Answer 5

[Winfried Johannes]

hello everybody

i have a problem to understand the matlab code of the qpsk example for adalm pluto. i have some questions and i hope you can help me:

what is the samplerate of the adc of the adalm pluto used by this example?

what is the samplerate of the data transfering over usb to the pc? what is the decimation factor of the adalm pluto?

what is the data format and the packet size of the data transfering over usb to the pc?

hopfully you can help me!

thanks.

best regards

winfried