AnsweredAssumed Answered

OIP3 Measurement at 1MHz to 5MHz

Question asked by YP on Sep 27, 2015
Latest reply on Oct 29, 2015 by mirek

Like • Show 0 Likes0
Comment • 2

Test Setup: As per attached figure.

IP3 measurement:
F1 = 2MHz
F2 = 2.9MHz
Power combiner isolation: 6dB
Attenuation = 20dB
Input power to DUT =-20dB

Based on current setup, after optimising the attenuation settings on the MXA, I get about 35dBm OIP3 using a thru. ( I would need to get to get the test setup to be able to measure >40dBm)
I have no problem when measuring about 10MHz. Could this be due to the PSG is having more harmonics at this low frequency?
My question is if you have any suggestions to improve my test setup?

Attachments

OIP3 Test Setup.jpg18.3 KB

No one else has this question

Outcomes

2 Replies

Dr_joel Oct 18, 2015 8:29 PM
Mark Correct
Correct Answer
You could try increasing your attenuation on one or both sides. If you have trouble reaching your desired power, then you could look to get a combiner with higher isolation. One possibility is to use a directional coupler or hybrid. But remember the isolation of a combiner is never better than the return loss of what follows, so the return loss of the SA will limit the isolation.

You might do some tests to see if the OIP3 is limited by the source or the SA. Add 10 dB attenuation to the SA, and see if the OIP3 improves; use a lower RBW if the noise gets too bad. If it does not improve you know you have source problem; if it does improve you have an SA problem.
Like • Show 0 Likes0
Actions
mirek Oct 29, 2015 2:06 PM
Mark Correct
Correct Answer
Well you can do several things to improve the performance.

It look like you are using resistive combiner/splitter. If you have any other type splitter combine or hybrid (like Dr.Joel suggest) used it. Most of them will have much better isolation then 6dB (more like 25dB). I will recommend to use coupler. Coupler usually have more then 30dB isolation. The problem with coupler is that ports are unbalanced (In-out maybe 1dB loss and coupling port 6, 10, 16...dB loss). To balance the loses I usually putting attenuator on output of the coupler with value equal to the coupling value. If you have 10dB coupler put 10dB attenuator and so on. Adding that attenuator will also improve isolation.

All combiners are sensitive to DUT input mutch (as Dr Joel mention), so if you have DUT that show very low return loss (high VSWR) do not expect that the isolation of the combiner is the same as the manufacture data show. That some time lead to false assumption that you setup is good and you getting low performance from DUT.
If that the case you need to put additional attenuation between combiner and DUT.

Setting of the MXA Spectrum Analyzer:
1. Check if the DC coupling is enable (usually by default , and it apply for frequencies below 10 MHz), if so disable.
2. Check if "Sweep IF Gain" and "FFT IF Gain". If Auto change to Low.
3. Depend of power that come from DUT output (each F1& F2) to Spectrum Analyzer (SA) set the internal SA attenuator (or attach external attenuator) so the power to first mixer of SA is around -36dBm. I usually connecting 10 dB external attenuator just after DUT and then adjust internal SA attenuator to get that -36dBm in to SA receiver.
4. Change the Span (keep RBW & VBW Auto) to like 1KHz or lower to get good dynamic range.
5. Use trace averaging (in most cases 5 averaging is OK)

Regards.
Like • Show 0 Likes0
Actions

OIP3 Measurement at 1MHz to 5MHz

Attachments

Outcomes

Related Content

Recommended Content