Remember Me?

#1
February 12th 05, 09:46 AM
 Jason Posts: n/a
IP3 calculation and estimation

Hello All

May I know if there is anyone here really have experience in
calculating or estimateing IP3 intercept point for any mixer real
case,(cmos or bipolar mixer)?
I need some advice on this topic
Thank you and hear from you all

rgds and thanks
Jason

#2
February 12th 05, 11:36 AM
 Airy R.Bean Posts: n/a

I do NOT have experience of this, but using one's
background training as an aspiring engineer, I would
use the output of a calibrated signal generator as the input
signal, together with an accurate - and variable - attenuator.

Connect an RF Voltmeter to the output of the mixer.

Now you can plot input power against output power, assuming
that it is within your capabilities to assess the output impedance of the
mixer.

This should give you sufficient points to plot the response curve up
to the 1dB compression point.

However, there is a problem here, and that is, at the point at which
the straight line starts to curve off, you are not just measuring the
output power of the desired output, but also are including that of the
newly-arrived distortion products.

However, as I said, I don't have the answer for you (We are, after all,
amateurs and not professionals), but perhaps I have given you food for
thought?

"Jason" wrote in message
oups.com...
Hello All

May I know if there is anyone here really have experience in
calculating or estimateing IP3 intercept point for any mixer real
case,(cmos or bipolar mixer)?
I need some advice on this topic
Thank you and hear from you all

rgds and thanks
Jason

#3
February 13th 05, 05:22 PM
 Frank Posts: n/a

The procedure is explained in detail, in an application note on
"Mini-Circuits" web site,
at: http://www.minicircuits.com/appnote/mxr1-18.pdf

Regards,

Frank

"Airy R.Bean" wrote in message
...
I do NOT have experience of this, but using one's
background training as an aspiring engineer, I would
use the output of a calibrated signal generator as the input
signal, together with an accurate - and variable - attenuator.

Connect an RF Voltmeter to the output of the mixer.

Now you can plot input power against output power, assuming
that it is within your capabilities to assess the output impedance of the
mixer.

This should give you sufficient points to plot the response curve up
to the 1dB compression point.

However, there is a problem here, and that is, at the point at which
the straight line starts to curve off, you are not just measuring the
output power of the desired output, but also are including that of the
newly-arrived distortion products.

However, as I said, I don't have the answer for you (We are, after all,
amateurs and not professionals), but perhaps I have given you food for
thought?

"Jason" wrote in message
oups.com...
Hello All

May I know if there is anyone here really have experience in
calculating or estimateing IP3 intercept point for any mixer real
case,(cmos or bipolar mixer)?
I need some advice on this topic
Thank you and hear from you all

rgds and thanks
Jason

#4
February 14th 05, 12:23 AM
 Jason Posts: n/a

Hello All(Airy, Caveat, Frank, Richard...)

Do allow me some time to check that application notes out before
posting back to you .
I hope we can discuss more sooner
Kindly be patient

Thank you all

best rgds and thanks
Jason

#5
February 14th 05, 02:44 PM
 Crazy George Posts: n/a

Jason:

Now that you have been directed to the theoretical discussion, let me alert you to some of the practical problems in
doing these measurements. Because of the very nature of the effect to be measured, the dynamic range required of the
test equipment approaches and sometimes exceeds the limits of even the very most expensive laboratory equipment.
Getting accurate measurements using average grade equipment requires a lot of caution, novel approaches, and it is often
simply not possible to make measurements with accuracy and repeatability of better then +/- 3 dB. Some quick examples:
power summers (dividers) are non-linear, intermodulation occurs in the most unexpected places, filters are either too
lossy or too broad, and so on.

--
Crazy George

the ATTGlobal.net is a SPAM trap. Use the att dot biz account.
"Jason" wrote in message oups.com...
Hello All(Airy, Caveat, Frank, Richard...)

Do allow me some time to check that application notes out before
posting back to you .
I hope we can discuss more sooner
Kindly be patient

Thank you all

best rgds and thanks
Jason

#6
February 16th 05, 07:53 AM
 Jason Posts: n/a

Hi George

Thank you for sharing with me your experience.
I will keep that in mind .
Currently I am still reading some related materials in order to
understand more of the basics

Will messsage to the board again soon

rgds and thanks
Jason

 Posting Rules Smilies are On [IMG] code is On HTML code is OffTrackbacks are On Pingbacks are On Refbacks are On

All times are GMT +1. The time now is 12:57 AM.