The Designer's Guide Community
Forum
Welcome, Guest. Please Login or Register. Please follow the Forum guidelines.
May 31st, 2020, 12:18am
Pages: 1
Send Topic Print
Two tone test for transmitter / PA (Read 13161 times)
Faisal
Community Member
***
Offline



Posts: 90

Two tone test for transmitter / PA
Sep 28th, 2007, 2:28am
 
Hi,

Does the two tone test make sense for transmitter?  After-all the signals are coming from inside the chip and their relative power levels known ?  


Back to top
 
 
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: Two tone test for transmitter / PA
Reply #1 - Sep 30th, 2007, 10:16am
 
Hi.

> Does the two tone test make sense for transmitter?

Yes, it does. The two tone test is useful for transmitter.

> After-all the signals are coming from inside the chip and their relative power levels known ?

You can relate ACPR requirement to IP3 and IP5.
 
Back to top
 
 
View Profile WWW Top+Secret Top+Secret   IP Logged
Faisal
Community Member
***
Offline



Posts: 90

Re: Two tone test for transmitter / PA
Reply #2 - Oct 1st, 2007, 8:17am
 
Thank you for responding. Could you elaborate a little further as to how I can relate the IP2, IP3, IP5 parameters to Adjacent Channel Power/Leakage Ratio?  Is there a way to simulate APCR in cadence?

The spectral mask for the transmitter is also important. But I think one needs to generate baseband data for this outside Cadence using tools like Matlab etc. Is there some utility in Spectre for this ? my system transmits  FSK signal i.e. constant envelope
Back to top
 
 
View Profile   IP Logged
vivarf
Community Member
***
Offline



Posts: 35

Re: Two tone test for transmitter / PA
Reply #3 - Oct 1st, 2007, 5:29pm
 
The ACPR can be simulated by Envelop Analisis (envlp of cadence spectre), you can find the example at the source link of Cadence.
In the example, the modulated base-band signal is cdma,  this signal is created and save into a file; Could someone give any ideas how to create this kind of file.
Back to top
 
 
View Profile veeater   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: Two tone test for transmitter / PA
Reply #4 - Oct 2nd, 2007, 12:39am
 
Hi.

You have to begin with top-down system study using behavioral model.

http://www.designers-guide.org/Modeling/modeling-rf-systems.pdf

Get relatioships between ACPR and IP3(IP5) using  behavioral model.

Back to top
 
 
View Profile WWW Top+Secret Top+Secret   IP Logged
Faisal
Community Member
***
Offline



Posts: 90

Re: Two tone test for transmitter / PA
Reply #5 - Oct 10th, 2007, 5:47am
 
Hi,

In a narrow-band RF transmitter, two tones could represent the complete BW power? Is this is the only reason for carrying out two-tone or multi-tone test on transmitter (yes it is non-linear) ?  

I will go through this document shortly that you have mentioned, to see the quantitative relationship betwen IP3 and APCR.

Back to top
 
 
View Profile   IP Logged
RFICDUDE
Community Fellow
*****
Offline



Posts: 323

Re: Two tone test for transmitter / PA
Reply #6 - Oct 15th, 2007, 8:32pm
 
The real problem with correlating IM3 to ACPR is that the amplitude variation of real signals is not the same as a two-tone signal. The nonlinear "response" of a circuit is different when signals with different amplitude distributions are applied to the amplifier. The nonlinearity of the circuit doesn't change; rather, it is the response of the circuit to a different stimulus that generates different levels of distortion.
It is possible to correlated IM3 to ACPR, but the correlation will change between signals of differing amplitude characteristics. Therefore, the correlation will change from signal to signal.

If you are lucky enough that the amplitude distribution of the signal is Gaussian or Rayleigh, then you stand a chance of developing an analytical correlation. This might work out for WCDMA/CDMA basestation signals or WiFi/WiMAX signals, but probably not for handset CDMA, WCDMA, or EDGE signals.
Back to top
 
 
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: Two tone test for transmitter / PA
Reply #7 - Oct 15th, 2007, 10:25pm
 
Hi.

ACPR can be affected by crest factor(peak factor) of modulated signal.
So ACPR is intensively dependent on modulation scheme, peak factor,
power level, backoff, etc.

Neverthless you can build relatioship between IP3 and ACPR for your specific modulation signal.

See attached figure.
Yellow and Blue curves are actual measurement results(not simulation).
Black curve is ACPR of behavioral model(simple IP3 and P1dB model without memory) simulation.
Red and Green curves are approximate polynomials from actual measuremet curves(yellow and blue).

As you see, we can build relatioship between IP3 and ACPR for target modulated signal.

Here an important issue is that you have to rebuild relationship if modulated signal scheme is changed, e.g. 3GPP --> WLAN, etc.
And if your transmitter have large frequency characteristics in passband, you have to consider their effects.
In this case you might have to build behavioral model with memory as well as nonlinearity
Back to top
« Last Edit: Oct 16th, 2007, 8:16am by pancho_hideboo »  

ACPR_BackOff.jpg
View Profile WWW Top+Secret Top+Secret   IP Logged
Faisal
Community Member
***
Offline



Posts: 90

Re: Two tone test for transmitter / PA
Reply #8 - Nov 13th, 2007, 9:33am
 
Hi,

I have a narrowband signal (approximately 5k bandwidth). It appears to me that I get different results when two tones f1, f2 are applied at the input of PA. I considered the following cases. I am running QPSS analysis and consider multiple intermodulation products.  

1) f1, f2 are 5kHz  apart (two tones are edge of channel)
2) f1, f2 are 500 Hz(two tones quite close)

Any ideas which is more realistic case of making the two tone test ? A detailed explanation is much helpful



Back to top
 
 
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: Two tone test for transmitter / PA
Reply #9 - Nov 16th, 2007, 2:57am
 
Quote:
I have a narrowband signal (approximately 5k bandwidth).
It appears to me that I get different results when two tones f1, f2 are applied at the input of PA.
I considered the following cases.
I am running QPSS analysis and consider multiple intermodulation products.  
1) f1, f2 are 5kHz  apart (two tones are edge of channel)
2) f1, f2 are 500 Hz(two tones quite close)
Any ideas which is more realistic case of making the two tone test ?
A detailed explanation is much helpful

Center frequency of your TX system is just one ?
Usually TX system have some channel frequencies, e.g. cellular, WLAN, Bluetooth, etc.
I don't think bandwidth of TX system is just same as modulation bandwidth.
Even though modulation bandwidth is relative narrow such as 5kHz,
TX system, especially PA, must have far wider bandwidth than 5kHz.

If your PA truely have very narrowband characteristics such as 5kHz
and you have interest in IM products up to 5th,
it is desirable that 5*(f2-f1) must be enough inside bandwidth of PA.
So I can't judge whether f2-f1=500Hz is appropriate or not from informations you described.

BTW what modulation is used ? FSK ?
What characteristics is important ?
Generally two tone test is alternative test.
What characteristcs of DUT do you intend to evaluate actually ?

Depending on modulation scheme and chacteristics you intend to evaluate,
sometimes a simple one tone test could be enough.

Two tone drive is equivalent to double sided AM modulation with carrier suppression.
This means an input amplitude to DUT changes dynamically.
So you can evaluate AM/AM and AM/PM characteristics under dynamic amplitude change.
This is a merit of two tone test.

You can also evaluate AM/AM and AM/PM characteristics by one tone test.
But this is under static amplitude change.






Back to top
 
« Last Edit: Nov 16th, 2007, 7:49am by pancho_hideboo »  
View Profile WWW Top+Secret Top+Secret   IP Logged
Pages: 1
Send Topic Print
Copyright 2002-2020 Designer’s Guide Consulting, Inc. Designer’s Guide® is a registered trademark of Designer’s Guide Consulting, Inc. All rights reserved. Send comments or questions to editor@designers-guide.org. Consider submitting a paper or model.