Questions about TWTA bwhavior, modeling

[charan langton]

TWTA Input & Output Backoff

Hi Everybody,
I'm trynig to calculate the IBO & OBO of a Saleh HPA in a OFDM system. I need to know if to calculate IBO, do I need to have just the amplitude of the OFDM signal or a complex OFDM signal(The output of the IFFT Operation)?

Thank you.

Regds,
Bavithi.

[charan langton]

The IBO is the power of the input signal normalized to the maximum input power. So 0 dB IBO means you are at maximum input power of the amplifier.

Lets say that your amplifier can handle a maximum of 1 watt input power, and can produce 10 watt output power. The IBO at this point is 0 dB, and so is the OBO, which is normalized to 10 watts. Since these are normalized to their maximum values, you can not determine the absolute gain of the amplifier from the IBO, OBO values.

Now as you reduce the power, the IBO will decrease. 3 dB IBO means you have input half a watt. The OBO will also go down, depending on your IBO/OBO relationship, which is different for every amplifier. Before any kind of analysis of an amplifier, you need to know this relationship, which is sometimes given in IBO/OBO and sometimes in actual power out such as dBms. To get IBO/OBO from dBms. read my tutorial.

IBI/OBO are usually referred to as positive numbers because they are backoffs from the maximum number.

There is no such thing as a Saleh amplifier. This is a way of characterizing the IBO/OBO relationship by way of curve fitting.

So to measure your IBO, just measure the power out of the modulator and divide by maximum power. For an OFDM signal this is just the amplitude squared.

Charan Langton

[bavithi]

Hi Admin, Thankyou for the explanation. So,I'm using the saleh's model for characterizing IBO/OBO.I calculate that for the model I used,I'm getting an IBO of 8.21dB & OBO of about 5dB. However,when I calculate the Symbol Error Rate after demodulation,its very high,about 45%.I assume that I'm operating in a highly non-linear region.Is there a way to reduce it further,I haven't manage to work on controlling the IBO & OBO in my model yet.However,I see that I will have to work with an IBO of about 6 to 8 dB only.

Thank you very much.
Cheers,
Bavithi.

[bavithi]

Hi All,

I'm trying to calculate the PAPR of OFDM symbols that I generated.According to the defn,for N=16,PAPR=(Amplitude of OFDM signal)^2/Power of the signal.
Should I find the RMS value of the Numerator?

Also,when I pass the OFDM signal through an amplifier,it clips the signals above the saturation value. So in order to reduce PAPR should I make changes to the OFDM signal before it goes into the HPA?

Thank you.

Bavithi.

[charan langton]

Bavithi,

The reason your BER is so high is not because of the non-linearity but because you have not not synced your output and input signals. You need to determine the delay through the model and then delay the input signal by that amount before calculating the BER. If you have done that and the BER is still that high, then you have some other problem. One can operate at saturation and still have useable BER such as 10-3 or so. Your BER says, you have a mistake in your model.

Charan Langton

[charan langton]

Yes, you do need to find the RMS value of the signal. The peak value which is N times is rarely seen.

There are many different methods of controlling PAR, you need to research the literature and decide what method you want to investigate. The simplest idea is of course to keep the backoff high.

Charan Langton

[saxofon]

Hi,
I'm looking for a simple IBO vs OBO formula for lingbudget calculation for TWTAs, linearized TWTAs and SSPAs. I cannot find these things in satellite datasheets, one time I found this in an eutelsat handbook:
If (ibo <= 13) Then OBO = 1.7 + 0.0313 * ibo^2
Else OBO = ibo - 7 (Linear Part of TWTA)

Can I take this for all tubes (and SSPAs?) or are significant differences between satellites?
Are amplifiers normally operated in the linear part? Satellite operators are speaking of "nominal power", but hardly any one can tell me how much that is. Which backoff is normally used?

With highfrequent regards..... Matthias

[charan langton]

IBO vs. OBO is a characteristic of A particular type of tube. Ka band would be different than s-band and one manufacturer's TWTA would have different IBO vs,. OBO than another. If it is linearized, the tube then has an entirely different IBO vs. OBO.

So an equation is NOT what you want to use in a link budget. You need to go to the satellite manufacturer and get the actual IBO vs. OBO for that system. That's the only way and the only correct way. Everythig else is pure guessing.

[saxofon]

So you wrote in your tutorial that there are two tube-manufacturers, I will ask them. I also red that SSPAs behave similar (not linear), so I have to ask SSPA manufacturers too.
I wonder how the Satmaster program calculates IBO vs. OBO. You have to choose between TWTA and SSPA in this prog. and that's all.
Why are these things not in the satellites' datasheets?

Thanks and regards Matthias

[charan langton]

Satmaster probably uses a "representative" IBO vs. OBO curve. The SSPA curve is a bit more linear than a Linearized TWTA. But the IBO vs. OBO for linearized TWTAs is very different than ones for non-linearized TWTAs. I dont see how a single curve can be used for a tube in a program. Perhaps you should contact the developers of Satmaster program.

I wrote the tutorial in late 90's. The manufacturers have changed. Lot of consolidation in the industry since then.

[fikky]

Hi admin or everybody....
How to make IBO in An adaptive predistorter for HPA with 16 QAM? I need to know if to calculate IBO, do I need to have just the amplitude of the 16 QAM signal or a complex 16 QAM(The output of the square root raised cosine filter)? If anysomebody have a source code or example program
about it (using matlab) please send me.
Thank you....

[anvar]

Hi Admin and everybody,
I’m looking for any information about relationship between measured IBO vs OBO of a satellite HPA for a single carrier and expected IBOE (each carrier) vs OBOE (each carrier) for two QPSK carriers. I’d like know a reducing range each of both carriers. RCA Review, vol. 41 of 1980yr describes it for two rf carriers but I am in great need for nowadays information.

Sincerely,
Anvar

[charan langton]

It depends on the actual IBO/OBO as measured. For a 2 carrier case. the OBO follows the singel carrier case and is about 3 dB lower at saturation. It is starts same as high IBI and then departs to a 3 dB lower point.

[singhofmpl]

Hi everybody
can anyone tell how to incorporate IBO or OBO in AM-AM and AM-PM characteristic equation for TWTA or SSPA power amplifier? Assume that IBO and OBO is given in dBs.

Please reply at the earliest.

[charan langton]

Create a table of IBO vs. OBO and phase shift. Then use this table lookup to find the gain and phase at each sample point. For each point the actual IBO, which is the ratio of the amplitude squared and the average power. Use this value to look up the output fom the table.