AnsweredAssumed Answered

PSA opt. 233 measuring receiver

Question asked by hudla on Sep 8, 2011
Latest reply on Sep 19, 2011 by hudla

Like • Show 0 Likes0
Comment • 4

Hi all,

I have a problem with remote reading of very low power levels using PSA opt. 233 Measuring receiver. For levels below -100 dBm, IF BW = 10 Hz, Accuracy High, the reading takes tens (even hundreds) of seconds, thus I need a command which waits until the reading is finished and then I want to read the level using :READ:TRFL?. Which command(s) should go before? I can use *OPC and then reading *STB? in a loop until bit 5 of the register is set to 1, but which command should I watch for completion? Which operation is actually completed, when the level reading appears on the display?

:INIT:TRFL
:INIT:CONTinuous ON
*CLS; *ESE 1                   // set event status register
:INIT:IMMediate              // restarts measurement
*OPC
read *STB? until bit 5 is set to 1

The bit 5 of status register is 1 right after the first reading of *STB? (apparently :INIT:IMM is finished quickly), but still the analyzer acquires data and a green message with XX seconds to the end of reading blinks. Which command should precede *OPC that sets the status Operation complete bit to 1 first when the reading is finished and not earlier? I tried :MEAS:TRFL or :FETCH:TRFL, but both commands end with timeout (some operation is not yet completed and the level reading is not available). Not very elegant solution is to set a very long waiting time before :READ:TRFL? to avoid the timeout, but...

Thank you for any ideas, the debugging is a bit time consuming when each reading takes minutes
Martin

No one else has this question

Outcomes

Jewe Sep 19, 2011 2:44 AM

Hello Martin,

Why not turn off the continues measurement?
You want to be in control of when the measurements starts and ends.
Before starting the measurement read out what the measuring time will be.
Set a delay until half of your bus time out(Normally this would be 10 seconds)

in short programming:
set up measurement
readout the sweep time
trigger sweep.
Wait till sweeptime - 5 sec.
Send *opc
Read *STB?

debugging tip: set the rbw bigger so the measurement is a bit faster but not too fast to see your problem.

Actions

hudla Sep 19, 2011 3:29 AM

Hello Jeffrey,

thank you for a hint. Actually I tried this, however it would work if the acquisition time would always be the same. For very low levels (< -110 dBm), I'm facing a problem that sometimes the measurement is done in e.g. 20 seconds, sometimes in 60 seconds or more (i.e., sometimes the green message goes like "Remain: 30s...20s...10s...0s" and then I can see the reading, sometimes the counter restarts once or more and goes again from the beginning, thus the measurement time is higher and not known in advance). Since I don't know the measurement time exactly, the *STB? ends with time out in some cases (continuous or single sweep, does not matter).

I'm reading the measured value e.g. 10-20 times until the reading settles and "makes sense". From Agilent app. note #5589-8161EN, p. 27, one could understand that the measured value is averaged internally and that the longer time one measures, the more accurate the result is. From my experience: for levels below -120 dBm, the first one to five readings have (sometimes) large error and then the readings converge to expected value (generator E4438C, external frequency references connected, so the frequency difference between generator and analyzer should be < 1 Hz and the analyzer should be properly tuned). I don't have the 1DS or 110 option (preamp.), so I'm trying to measure signals very close to PSA's noise floor.

Regards,
Martin

Actions

Jewe Sep 19, 2011 7:04 AM

Hi Martin,

If I remember correctly the following:
10 dB lower in the lower range is 10 times averaging
20 db lower in the lower range is 100 times averaging
30 db etc.
so 10^(x dB below the normal noise floor/ 10) times averaging.

What is your reference level when performing these measurements?
0 dBm or 107 dbuV(this is the same in 50 ohm)
Have you tried with a lower reference level?

Your best bet is getting a preamp if you want to go lower in the noise floor.

Best regards,

Jeffrey

Actions

hudla Sep 19, 2011 10:13 AM

Hi Jeffrey,

my reference level for the measuring receiver calibration is 0 dBm, but I also tried -10 dBm. I've chosen the level according to the manual, i.e. the level should be within the dynamic range of my N5532A head (-20 dBm to +30 dBm). The absolute level for Range1 was compared with our reference etalon power meter and the error is within +/- 0.15 dB in the whole frequency range.

Finally I solved the initial problem such that I set the PSA's time out to 600 seconds (more than a single measurement can take) and all the commands are written directly after each other, without testing the status bit. Without a prior knowledge of the measurement time, this seems to be the most efficient solution.

Best regards,
Martin

Actions

4 Replies

Jewe Sep 19, 2011 2:44 AM
Mark Correct
Correct Answer
Hello Martin,

Why not turn off the continues measurement?
You want to be in control of when the measurements starts and ends.
Before starting the measurement read out what the measuring time will be.
Set a delay until half of your bus time out(Normally this would be 10 seconds)

in short programming:
set up measurement
readout the sweep time
trigger sweep.
Wait till sweeptime - 5 sec.
Send *opc
Read *STB?

debugging tip: set the rbw bigger so the measurement is a bit faster but not too fast to see your problem.
Like • Show 0 Likes0
Actions
hudla Sep 19, 2011 3:29 AM
Mark Correct
Correct Answer
Hello Jeffrey,

thank you for a hint. Actually I tried this, however it would work if the acquisition time would always be the same. For very low levels (< -110 dBm), I'm facing a problem that sometimes the measurement is done in e.g. 20 seconds, sometimes in 60 seconds or more (i.e., sometimes the green message goes like "Remain: 30s...20s...10s...0s" and then I can see the reading, sometimes the counter restarts once or more and goes again from the beginning, thus the measurement time is higher and not known in advance). Since I don't know the measurement time exactly, the *STB? ends with time out in some cases (continuous or single sweep, does not matter).

I'm reading the measured value e.g. 10-20 times until the reading settles and "makes sense". From Agilent app. note #5589-8161EN, p. 27, one could understand that the measured value is averaged internally and that the longer time one measures, the more accurate the result is. From my experience: for levels below -120 dBm, the first one to five readings have (sometimes) large error and then the readings converge to expected value (generator E4438C, external frequency references connected, so the frequency difference between generator and analyzer should be < 1 Hz and the analyzer should be properly tuned). I don't have the 1DS or 110 option (preamp.), so I'm trying to measure signals very close to PSA's noise floor.

Regards,
Martin
Like • Show 0 Likes0
Actions
Jewe Sep 19, 2011 7:04 AM
Mark Correct
Correct Answer
Hi Martin,

If I remember correctly the following:
10 dB lower in the lower range is 10 times averaging
20 db lower in the lower range is 100 times averaging
30 db etc.
so 10^(x dB below the normal noise floor/ 10) times averaging.

What is your reference level when performing these measurements?
0 dBm or 107 dbuV(this is the same in 50 ohm)
Have you tried with a lower reference level?

Your best bet is getting a preamp if you want to go lower in the noise floor.

Best regards,

Jeffrey
Like • Show 0 Likes0
Actions
hudla Sep 19, 2011 10:13 AM
Mark Correct
Correct Answer
Hi Jeffrey,

my reference level for the measuring receiver calibration is 0 dBm, but I also tried -10 dBm. I've chosen the level according to the manual, i.e. the level should be within the dynamic range of my N5532A head (-20 dBm to +30 dBm). The absolute level for Range1 was compared with our reference etalon power meter and the error is within +/- 0.15 dB in the whole frequency range.

Finally I solved the initial problem such that I set the PSA's time out to 600 seconds (more than a single measurement can take) and all the commands are written directly after each other, without testing the status bit. Without a prior knowledge of the measurement time, this seems to be the most efficient solution.

Best regards,
Martin
Like • Show 0 Likes0
Actions

PSA opt. 233 measuring receiver

Outcomes

Related Content

Recommended Content