Contact an Expert

Thread: Spectrum Analyzer L.O. Issue


Permlink Replies: 5 - Pages: 1 - Last Post: Aug 2, 2012 9:36 AM Last Post By: tabbott Threads: [ Previous | Next ]
SOLT_guy

Posts: 268
Registered: 05/30/09
Spectrum Analyzer L.O. Issue
Posted: Jul 14, 2012 7:48 AM
Click to report abuse...   Click to reply to this thread Reply
Dear Sir:

I own a spectrum analyzer and I use a HP 8593E spectrum analyzer (SA) where I work.

Both units that I use have the same issue. The SA local oscillator appears on the screen and sometimes it is difficult to make measurements with the LO signal so close to my test signal in frequency.

I am seeking to purchase another Agilent SA in the future and I want to know when did HP/Agilent begin making SAs that supress the L.O. signal from the display? I am assuming that this problem has been eliminated over the years Agilent has been making SAs.

My next question is an immediate question which I need to have answered:

The HP 8593E SA has a bandwidth (BW) of 9KHz to 22GHz. This is my problem. I needed to measure a signal at approximately 10 KHz. I input a signal out of my function generator and output the signal to the SA and I saw a blank screen on the display. I was saying to myself, "where is my signal."

I read the manual about measuring low frequency signals, or signals that are close in frequency to the LO signal, and the manual instructed me to use a "coupler" in my test set up. The manual does not specifically assert which type of "coupler" to use. There are lots of devices that use the term, "coupler", in the electrical industry.

Please tell me exactly how to correctly measure a low frequency signal using the HP 8593E SA. In your reply, please specify the parts to use in the test set up , as well as, the test set up. Measurement mode, or button presses, for low frequency measurements would be appreciated.

It may be possible that my unit is not working correctly. If you suspect this, please tell me how to test to see if it is not operating correctly.

This is what leads me to believe that there may be a problem:

The LO signal came up when I began to play with the frequency and span buttons. I noticed a large LO signal at various settings and I know that this is nuisance signal.

If this is possible, please tell me how I can eliminate this signal and its harmonics from my measurements (or, at least, detect the harmonics, so that I know I am looking at a false reading).

Thank you for your time.

Edited by: SOLT_guy on Jul 14, 2012 7:49 AM

Edited by: SOLT_guy on Jul 15, 2012 12:44 PM

Edited by: SOLT_guy on Aug 2, 2012 9:36 AM
dhamilton

Posts: 144
Registered: 02/01/08
Re: Spectrum Analyzer L.O. Issue
Posted: Jul 16, 2012 9:46 AM   in response to: SOLT_guy in response to: SOLT_guy
Click to report abuse...   Click to reply to this thread Reply
Soltguy,

The 8593E can measure your 10 KHZ signal pretty well if you have both the options 130 and 004. This combination was also known as Option 140. The option 130 is the narrow resolution bandwidth filter option to take your measurement to 30 Hz bandwidth vs. 1 KHz without. Without the option it is very difficult to make quality measurements below 100KHz.

The option 004(OCXO) will have you signal frequency accurate.

If you would like to get those options installed contact me offline.
SOLT_guy

Posts: 268
Registered: 05/30/09
Re: Spectrum Analyzer L.O. Issue
Posted: Jul 20, 2012 4:14 PM   in response to: dhamilton in response to: dhamilton
Click to report abuse...   Click to reply to this thread Reply
Dear DHamilton;

Thank you for your reply.

I am surprised to hear that an option is required to measure low frequency. Measuring low frequency shouldn't be difficult to integrate into the SA. I would like to know the design tradeoff which made designing low frequency measurement untenable for this unit. Hopefully an Agilent engineer can answer this question.

You are a really knowledgeable guy.

I was wondering if you could tell me if any of the latest SA models have completely eliminated the LO signal from the display?

Thank you for your input, it was most welcome.

Edited by: SOLT_guy on Jul 20, 2012 4:14 PM
dhamilton

Posts: 144
Registered: 02/01/08
Re: Spectrum Analyzer L.O. Issue
Posted: Jul 20, 2012 4:32 PM   in response to: SOLT_guy in response to: SOLT_guy
Click to report abuse...   Click to reply to this thread Reply
SOLTguy,

The newer analyzers have capability down to 3 HZ because the have Resolution B/W filters and the RF front end hardware designed to operate low enough to discern the the intelligent signal you are measuring, see harmonics and other parameters you may need to measure. However, they are quite expensive.

The 859xE series can measure fair @ 10 KHz w/o option 130. It measures good w/option 130 because the narrowest B/W filter is 30 Hz vs. 1 KHz. With the 1 KHz B/W filter it is difficult to make very good measurements below 100 KHz.

All spectrum analyzers do have a LO bleed through. Some more than others.

The 859xE series should have the LO about 11 to 15 dB below full scale. If you apply a known -10 dBm signal @ 10 KHz and with the spectrum analyzer set to a RF level of -10 dBm and RF input attenuator set to 0 dB, CW set to a center of 10 KHz and a span of 25 KHz. The LO should be approximately 11 - 15 dB lower than your RF carrier. If it is close the same level or higher than your RF signal, you have defective hardware which is causing this failure. The unit requires repair.

If your unit does have the failure and/or you need the option 130 installed. Contact me offline.

Dave
Frank_BR

Posts: 60
Registered: 07/04/07
Re: Spectrum Analyzer L.O. Issue
Posted: Jul 22, 2012 4:59 PM   in response to: SOLT_guy in response to: SOLT_guy
Click to report abuse...   Click to reply to this thread Reply
SOLT_guy wrote:
My next question is an immediate question which I need to have answered:

The HP 8593E SA has a bandwidth (BW) of 9KHz to 22GHz. This is my problem. I needed to measure a signal at approximately 10 KHz. I input a signal out of my function generator and output the signal to the SA and I saw a blank screen on the display. I was saying to myself, "where is my signal."

I invite you to take a look at a post I wrote in this same forum some time ago :
http://www.home.agilent.com/owc_discussions/thread.jspa?messageID=59007&#59007

The trick is to use an external mixer to up convert the low-frequency to something around 10 MHz, for example. The 10 MHz reference that almost all spectrum analyzers have is a convenient source for the LO for the external mixer. Even a cheap general purpose mixer is so good and well balanced at 10 MHz, that the LO leakage in the external mixer isn't an issue.

SOLT_guy wrote:
I am surprised to hear that an option is required to measure low frequency. Measuring low frequency shouldn't >be difficult to integrate into the SA. I would like to know the design tradeoff which made designing low >frequency measurement untenable for this unit. Hopefully an Agilent engineer can answer this question.

Yes, with today's technology, it is, conceptually and practically, very easy to integrate low- frequency spectrum analysis into a RF/microwave spectrum analyzer. Most modern SAs already contain the necessary components to do low-frequency analysis. These spectrum analyzers are based on digital IF, which uses A/D conversion followed by FFT, so the analysis of low-frequency signals would be straightforward. In particular, the various mixers and filters used in a high-frequency analysis wouldn't be necessary as the A/D would convert directly the "bandbase" signal.
tabbott


Posts: 1,483
Registered: 07/24/09
Re: Spectrum Analyzer L.O. Issue
Posted: Jul 23, 2012 3:32 PM   in response to: SOLT_guy in response to: SOLT_guy
Click to report abuse...   Click to reply to this thread Reply
Attachment SCREN019.GIF (15.8 KB)
Attachment SCREN016.GIF (16.3 KB)
Attachment LO2.png (84.1 KB)
Attachment LO1.png (74.4 KB)
Hi -

Even a sophisticated spectrum analyzer with L.O. suppression will still display a signal at 0 Hz. When the analyzer is tuned to 0 Hz and the L.O. feedthrough is present on screen, that is when the 1st L.O. is at the frequency of the 1st I.F. For example, I believe the 1st I.F. in the 8593E in low band is 3921.4 MHz. When the 1st L.O. feedthrough is present on the analyzer display, the 1st L.O. is tuned to 3921.4 MHz and goes through the down conversion system as a “normal” signal and is displayed. The main reason for suppressing the L.O. on a spectrum analyzer is to gain better dynamic range at low frequencies. The Agilent PSA and PXA high performance spectrum analyzers both have the L.O. suppression feature.

Since your 93E is specified to function from 9 kHz to 22 GHz, you should be able to narrow the span and RBW to measure a 10 kHz signal. The attachments of my E4407B, ESA and my N9030A, PXA illustrate that a 20 kHz span and a 300 Hz RBW should work (see LO1.png, LO2.png of the PXA and SCREN016.GIF, SCREN019.GIF of the ESA).

You can actually enter a 300 Hz RBW on your 93E. What is the input power level of your 10 kHz signal in dBm from your function generator? You do not need a coupler to measure this, simply inject the signal into the RF input of the analyzer. If the 10 kHz signal does not measure correctly, change the function generator frequency to 100 kHz and measure that on the 93E. If you suspect a problem with the 93E, you can manually check the absolute amplitude accuracy and the frequency response of the analyzer as long as you have the correct calibrated test equipment to do so. You can find the manual procedure online at www.agilent.com Enter “8593E” and in the product results select “manuals”. You will see a link to the Calibration Guide. That guide has these manual tests.

If your L.O. feedthrough signal varies with span and RBW, it is possible that the 1st Mixer is damaged most likely due to overpower, excess DC or ESD damage.

I hope this helped -

Point your RSS reader here for a feed of the latest messages in all forums