I've been having problems with my 8902A lately. In Tuned RF (TRF) it seems to get lost easily. I tune to a new frew, do INST PRESET, press MHZ to lock @ freq (typically 0 dBm no mod - 8642A, 8657B, etc.). I did rf cal and zero. Go to power, good reading, go to TRF, press cal at 0 dBm and usually okay. I step down in 10 dB increments and sometimes make it to -40 before next cal, and usually to -80 for next cal. [EDIT] - note that I have timebase out from generator externally locking the 8902A.
BUT ----- intermittently, I get errors and it gets lost (loses lock), sometimes have to re-cal at -30 and -70.
I CAREFULLY went into the adjustment routines and am trying a few of the safe ones. I found the A54 (SYNCH DET OFS) was off (supposed to be +/-0.0001 rdg; but was at 0.0054). I zeroed it which caused my TRF levels at ~0 dBm to get really unstable. So I set it back to its incorrect setting (temporarily).
Then I did ADJ # 17 TRK TUNE OFS on A20. The 455 KHz IF there is supposed to be 454.50 to 455.50 KHz. It was way off (450.5 or so). So I adjusted that (carefully lockstetp to the procedure), and then went back to the A54 SYNCH DET OFS adjustment, setting it to zero. This made my signal lock in properly (did RF power at 0 dBm, then got a good match with TRF after Cal).
I'm just doing those adjustments that may relate to TRF working properly (locking on, stable all through the range).
I get pretty stable TRF levels (I have an 8496B step atten connected to sensor module from an 8657B (don't ask about the atten on my 8657B - not dependable). But I still occasionally get these unlocks. I do a new freq, step down and do cals at 0, -40 and -80 dBm, but occasionally it gets lost.
So I'm now on Adj # 8 (ALC REF) on pg 5-16. It calls for a special code entry of: 49. S (Shift) 1 SPECIAL. It read a little low (2.0963 - vs. 2.097 - 2.103).
Anyone who's a high level guru on 8902A's please feel free to send me your thoughts or ideas.
Is is at lower frequencies? (Below 10.1 MHz I have heard of issues)
Also try 5 dB steps until CAL comes on. Sometimes you get a bad spot. Change frequencies do the CALs there and come back to the bad one.
I've not had to adjust any yet. I tried to adjust one and found I had a broken AT1 or AT2 (the programmable 10,20,40).
You shouldn't need to tie any timebases together.
Instead of MHz key which centers the tuning try keying in your frequency to tell it what is going in.
I've asked my boss about 8902's in regards to repair. He says that nobody is an expert at them because of the complexity. However your best bet is to ask everybody that has done them if they have seen your problem.
In a very rambling way - no I really can't help you but if somebody has seen it before and can remember they will know, otherwise good luck.
Also try one of the RF and Microwave forums found here
https://forums.tm.agilent.com/phpbb2/index.php
I've found that throwing a 3 or 6 dB attenuator on the 8902's input can sometimes help, and you can input the offset (don't remember how at the moment, it's Saturday :-D )
Also, if you're going through the mixer, this can cause odd readings occasionally, if you're below 1.2 GHz, go straight in.
As for adjustments on this beast, I leave it to the lead tech and go back and abuse the fiber optics.
Good luck
Mike
Quote from: mdbuike on 02-09-2008 -- 07:34:09
I've found that throwing a 3 or 6 dB attenuator on the 8902's input can sometimes help, and you can input the offset (don't remember how at the moment, it's Saturday :-D )
I can help with that. First measure the attenuator so that you have a value. Then enter the value then blue shift and the far right button in the same row as peak + , peak - , AVG . It says something like "attn" below it. This is also useful to avoid adding your zero. If zero at 100 MHz was -0.17 dB on your 436A press -0.17 shift and that button. Then run your attenuation. In this case however the offset may not really matter if you set ref with it in then run down the attenuator.
"BUT ----- intermittently, I get errors and it gets lost (loses lock), sometimes have to re-cal at -30 and -70. "
I've seen a lot of these issues before myself. Usually it was because of the TI I was working on having a slightly unstable frequency, ESPECIALLY when the frequency was manually entered into the 8902. However recaling the 8902 at -30 or -70 fixed her right up.
Just a side note, I've seen junior techs beating their head against the bench because they couldn't get a TRF reading to save their lives. Heh. Usually there was some sort of modulation coming from their source. . . .
Just to show my age a little, I have done the IF alignment on an HP 8552A (the IF section on the old HP 141T spec ans). So that's kind of my dues I paid (not to mention the nickname I had numerous years ago of "Captain Sneak-A-Tweak").
Back to the HP 8902A. Regarding the alignments I did, I ran it for a few hours and it does definitely seem to have improved.
However, the unit under cal with the losing lock symptoms was an HP 8642A which was a HOT unit (customer needed it back due to line down). So we were not able to attempt to re-duplicate the symptoms there.
But that was a very good point (which I did pursue) as to whether the unit being cal'd was unstable or had modulation on (simulating instability).
We'll keep watching it.
Thanks again everyone for the inputs.
Uhmm you will always have to "Recal" after every -40dB. I may be missing something in your post ( I am not feeling well today ), but I never press the "Mhz" to lock the freq, You want it auto tune the freq while doing TRFL. It can be a little tricky when using below 10Mhz too.
"However, the unit under cal with the losing lock symptoms was an HP 8642A which was a HOT unit (customer needed it back due to line down). So we were not able to attempt to re-duplicate the symptoms there. "
Don't ya hate it when that happens? Seems like everytime something "interesting" happens it just so happens to be on a HOT item or some pacing item that Joe Schmoe shoulda had cal'd months ago. . . . . . I hate being rushed on the bench. Too many days I wish I were a GS or contract, just so I could "work by the hours" (no offense guys and gals- I only wish I could have gotten away with a 40 hour work week!)
But an 8642 having issues? Those are as good as Old Faithful. . . . . . interesting indeed. Yes, I do recall those being the ones with the modulation issues causing chaotic readings.
I'm just not altogether convinced that the 8642A had a problem. It just would have been nice to verify that my adjustments (theoretically - so I hope -) to optimize some of the phase locking areas in my 8902A actually fixed a problem (as opposed to having a flaky 8642A). My working theory is that my adjustments improved some borderline adjustment points and that the 8642A is working fine. Checking it again with tuned RF would have confirmed that.
Oh well. We'll see how it goes.
Thanks everyone for all the great replies.
Quote from: flew-da-coup on 02-11-2008 -- 10:57:07
Uhmm you will always have to "Recal" after every -40dB.
this is exactly right, if you are decending in power or ascending in attenuation, at the exact points you mentioned, you will get the recal light. press the recal button, wait til recal is complete then continue up in attenuation or down in power depending on how you look at it (half empty or half full) problem comes when your attenuator goes too far past the -30 and -70 spots in one big jump. try to ease up to those points and wait for the recal light. recal and you are on your way!
Also , if you want to reduce the measurement uncertainty . do not measure the attenuation through the sensor head. Go directly to the 8902A or Down Converter ( depending on freq.).
No the problem doesn't relate to the recal/uncal lights. It showed up again yesterday afternoon then disappeared after I reset the power on the 8902A (but did not go away when I did an instr preset).
I did TRF at 0 dBm at 1000 MHz, go the uncal/recal, hit CAL, the uncal/recal went away, but my amplitude drifted by as much as +/-3dB. When in normal power mode at that same setting, reading was completely stable. When in freq counter mode with 7.4 SPCL (high res counter mode), and with timebases locked, I had a drift in frequency of +/-5 to 15 Hz or so (within specs of instruments). It was a slow drift.
I reset the power on the 8902A, then the TRF amplitude drift went away.
I am getting more convinced that my 8902A has an intermittent problem.
Quote from: Hawaii596 on 02-13-2008 -- 09:07:05
No the problem doesn't relate to the recal/uncal lights. It showed up again yesterday afternoon then disappeared after I reset the power on the 8902A (but did not go away when I did an instr preset).
I did TRF at 0 dBm at 1000 MHz, go the uncal/recal, hit CAL, the uncal/recal went away, but my amplitude drifted by as much as +/-3dB. When in normal power mode at that same setting, reading was completely stable. When in freq counter mode with 7.4 SPCL (high res counter mode), and with timebases locked, I had a drift in frequency of +/-5 to 15 Hz or so (within specs of instruments). It was a slow drift.
I reset the power on the 8902A, then the TRF amplitude drift went away.
I am getting more convinced that my 8902A has an intermittent problem.
Get rid of the Sensor Head and go directly into the 8902 form the Sig Gen. Take your reading @ 0.0dBm/1000MHz and put it in the 8902 as an offset. Then set reference and see if that works.
had an 8902 that was having intermittant problems w/autotuning. Agilent found a cold solder joint in power supply that they claimed was causing the problem. Seems OK since.