AnsweredAssumed Answered

8590B no calibration after memory loss

Question asked by EremitaAnalogico on Nov 6, 2019
Latest reply on Dec 10, 2019 by Tardz

Like • Show 0 Likes0
Comment • 25

Hi, I have at work one 8590B that has suffer of boring problem of backup battery dead.

I have replaced the battery (both lithium and supercap) and I tryed to recalibrate the instrument.

When I performed early attempts, the 8590 says that not found the calibration signal.

In effect it was displayed with a considerable offset.

I have tryed a -37Hz procedure, and with this both amplitude and frequency calibration starts, but the process end with span sens fail.

The display obtained is strange, I show it in attachment.

Naturally, these errors don't manifestate prior to battery fail.

I hope that someone can help me to restore this analyzer, thanks in advance.

No one else has this question

Outcomes

25 Replies

Tardz Nov 11, 2019 9:10 AM
Mark Correct
Correct Answer
Hi,

You lost the calibration constants because of the faulty battery. I would first try to recall the default calibration constants.

First, enter the PASSCODE
FREQUENCY
- 37 HZ
CAL
DEFAULT CAL DATA (CAL menu page 3 of 4)

After trying again the CAL automatic calibration amptd. If it works well do not forget CAL STORE. Then redo the other calibrations.

Yves
Like • Show 0 Likes0
Actions
- EremitaAnalogico @ Tardz on Nov 12, 2019 2:46 AM
  Mark Correct
  Correct Answer
  Hi, thanks for reply, but I have already do the passcode entry with -37 Hz
  Then the cal procedure not start (cal signal not found), or start with "span sens fail" error at end if I provide an external cal signal slight decentered in frequency to fool the analyzer.
  The internal cal frequency and amplitude is correct.
  Like • Show 0 Likes0
  Actions
  - Tardz @ EremitaAnalogico on Nov 12, 2019 8:38 AM
    Mark Correct
    Correct Answer
    Hi,
    I have already experienced this problem of losing all calibration constants, and in my case it was worse because I could not read on the screen because it was unsynchronized. In short, in my case I placed the display settings first, of course, then the reference 300 MHz and finally I had the calibration flatness manually. After that I could use the other automatic calibrations.
    The internal calibrator is at 300 MHz, it comes from a multiplication of a reference to 10 MHz. You must first adjust this calibration constant:
    FREQUENCY
    -2001 HZ
    CAL
    MORE 2 of 3
    CAL SERVICE
    CAL TIMEBASE
    you have value from 0 to 255
    you must adjust frequency to 300 MHz ± 75 Hz (± 0.25ppm)
    If you have no other instrument to accurately measure 300 MHz frequency, set the middle value of 128 to begin, you can change it later.
    Try to do the frequency calibration CAL FREQ.
    If the frequency calibration PASS, try calibration in AMPL. If ever the FAIL calibration, do the FLATNESS calibration, then retry the amplitude calibration.
    
    Yves
    Like • Show 0 Likes0
    Actions
    - EremitaAnalogico @ Tardz on Nov 12, 2019 11:55 PM
      Mark Correct
      Correct Answer
      Hi, my analyzer also had out of sync monitor, but I have solved by erasing one more time the entire memory. When powered up, luckily the monitor was in sync.
      Regarding the calibration signal out of analyzer, the frequency is out less than 1 KHz below 300MHz
      The amplitude is -19,59 dBm:
      
      I have entered the commands that you have suggested, but I can't find CAL TIMEBASE menu item in Service Cal menu.
      Here my screenshot. Perhaps I have another version of firmware?
      
      Like • Show 0 Likes0
      Actions
      - Tardz @ EremitaAnalogico on Nov 13, 2019 8:36 AM
        Mark Correct
        Correct Answer
        OK, you don't have to do CAL TIMEBASE calibration for HP8590B model, sorry. This spec below
        
        go to FLATNESS DATA (edit DATA), probably the cause of this step in level because you have to chart, one below 50MHz and another above 50MHz. After this step, freq & ampl calibration normally will work.
        
        Yves
        Like • Show 0 Likes0
        Actions
        EremitaAnalogico @ Tardz on Nov 13, 2019 11:19 PM
        Mark Correct
        Correct Answer
        Hi Tardz, thanks for your assistance.
        I can do very few things when I enter Flatness Data menu.
        If I select Edit Flatness I get an error on the left of screen
        Store Flatness and Init FLT only give an OK quickly after I press respective keys.
        
        Like • Show 0 Likes0
        Actions
        Tardz @ EremitaAnalogico on Nov 14, 2019 7:42 AM
        Mark Correct
        Correct Answer
        Hi, try this procedure:
        Yves
        Like • Show 0 Likes0
        Actions
        Tardz @ Tardz on Nov 17, 2019 10:59 PM
        Mark Correct
        Correct Answer
        Hi,
        I've done yesterday, the FLATNESS calibration procedure for my 8595E and this is result below:
        
        Now this sequence to do for your 8590B is below:
        
        First, setup on your 8590B
        - CAL 2 of 4 CORRECT OFF
        - Center FREQ at 300 MHz
        - CF step 50 MHz
        - span 10 MHz
        - resBW 1MHz
        - marker peak
        - signal track ON
        - ajust RF signal of your generator to have -14dBm on your 8590B
        - This is the REFERENCE point, if you have a relative mode on your precision milliVoltmeter, enable it now => reading must be 0.0 dB at 300 MHz.
        - For each Frequency in the chart, note the relative level on your precision milliVoltmeter when the RF level is -14dBm on your 8590B. Use step CenterFREQ and peak marker at each frequency for your help.
        
        When you chart is completed, it's will be time to store these calibration factors.
        PRESET
        and follow these steps below.
        
        Yves
        
        Like • Show 0 Likes0
        Actions
        Tardz @ EremitaAnalogico on Nov 17, 2019 8:35 PM
        Mark Correct
        Correct Answer
        Hi,
        "I can do very few things when I enter Flatness Data menu.
        If I select Edit Flatness I get an error on the left of screen
        Store Flatness and Init FLT only give an OK quickly after I press respective keys."
        
        Flatness calibration is not an automatic calibration, you must manually perform this calibration following the procedure and after enter calibration factor with EDIT FLATNESS.This procedure is not complicated, if you need help I can summarize it and you can try a test.
        
        When you press the Edit Flatness key, you see the word "error", it's normal, it's the error or calibration factor to enter not a problem. You can press INIT FLT to erase the contents (around 2 second to respond), or simply enter the points you have measured according to the procedure. However, this key allows you to view the current content without modifying it. By pressing the UP and DOWN arrows or using the Knob, you will see the current calibration factors. The first point that will be asked when you have all the measurement points described in the table, will probably be the 4 MHz based on the image you have provided. When you have finished entering all the data, press STORE FLATNESS and wait the answer (around 10 sec).
        
        The table of calibration points is taken from document 5958-7098 for models 8590B and 8592B. The second image shows the table corresponding to 8591A, a calibration table with a starting point of 4 MHz. By checking the other points with the UP and DOWN keys, you will see if this corresponds to this table. If so, then it is this table that you must follow to find your calibration points ... the version of the table depends on the doubt the version of the installed firmware.
        On the first image at right, it's my 8595E spectrum, and the on-screen plot shows the calibration factors for my spectrum. In your case, the trace is flat.
        
        Finally on the last image, we see the data that I picked up in the table, displayed on a graph and we can see that it corresponds to the image on my 8595E spectrum.
        On the same image at the top left is the calibration check at table frequencies only. Theoretically it must give a straight line centered on 0dB error. And finally still on the same image at the top right, the errors for frequencies starting at 10 MHz up to 6.5GHz per step of 10 MHz. We see here that the calibration worked and the spectrum is in the specifications.
        
        Yves
        
        Like • Show 0 Likes0
        Actions
        EremitaAnalogico @ Tardz on Nov 19, 2019 12:16 AM
        Mark Correct
        Correct Answer
        Hi Yves, thanks for all this informations.
        I returned today at work but I have to refurbish some THR880. I try soon your procedure.
        Like • Show 0 Likes0
        Actions
        EremitaAnalogico @ EremitaAnalogico on Nov 21, 2019 1:03 AM
        Mark Correct
        Correct Answer
        Hi Yves, I tryed to insert new flatness data, but the analyzer seem to not accept new values.
        
        Edit: I have performed the -2001Hz passkey prior to insert new values
        
        All current values are 18.000 in all frequencies.
        If I insert another value, i.e. -1 or any other, the stored value is always 12.000
        I tryed i.e. 20, 20.000, 1, -1, .01, etc.... always followed by push +dBm or -dBm keys
        If I push Init FLT the values comes all back to 18.000
        What it happens?
        Thanks again.
        Like • Show 0 Likes0
        Actions
        Tardz @ EremitaAnalogico on Nov 21, 2019 9:02 AM
        Mark Correct
        Correct Answer
        Hi, try -37HZ passcode instead of -2001Hz. Tell me is the table starting at 4 MHz, followed by 41MHz, 78MHz, 115Mhz ... is the right table.
        
        Yves
        Like • Show 0 Likes0
        Actions
        Tardz @ Tardz on Nov 21, 2019 1:49 PM
        Mark Correct
        Correct Answer
        It would take a test to find out if the device is able to keep the information. We will use CAL ATT ERR calibration to do this test:
        CAL
        on page 2/4 or 3/4
        DIAG SERVICE
        DISPLAY CAL DATA
        You will see normally on the first page, in the list CAL ATT ERR. Note the 5 values displayed.
        
        PRESET
        FREQUENCY
        -2001
        HZ
        CAL
        CAL SERVICE
        SET ATTN ERROR
        enter 5 different values from previously read values.
        
        PRESET
        CAL
        on page 2/4 or 3/4
        DIAG SERVICE
        DISPLAY CAL DATA
        See if the values have changed and if they match the values you entered in the previous step.
        
        If YES, then repeat the exercise and set all values to 0, you will have to calibrate the attenuators later.
        If NO, you have a problem with the NVRAM memory (battery-related) .
        
        Yves
        Like • Show 0 Likes0
        Actions
        EremitaAnalogico @ Tardz on Nov 26, 2019 2:51 AM
        Mark Correct
        Correct Answer
        Hi, I tryed The -37 Hz passcode instead of -2001 Hz but the result is the same.
        I also tried to change the CAL ATT ERR values and I can.
        My screen is different, but I can do the procedure and the 5 values has changed from 0.00 to 13.00 in accordance with what I entered.
        To answer at you question, yes, the right table is starting at 4 MHz, followed by 41MHz, 78MHz, 115Mhz...
        
        A strange thing that I notice is the "random" value of Ref Level. If you look at my previous photos from the top of topic, You see that these values at top left corner of screen are different from each other. Do you think that this is relevant? How I can correct it?
        Thanks.
        Like • Show 0 Likes0
        Actions
        Tardz @ EremitaAnalogico on Nov 26, 2019 9:45 AM
        Mark Correct
        Correct Answer
        Hi,
        Based on the CAT ATT ERR test you have done, the NVRAM seems to work well. Yes, I had noticed the big difference on the REFlevel, I'm not sure but probably caused by the big flatness error, and the automatic amplitude calibration routine can not converge to a stable value. Each time you attempt this calibration you will have a different value. It is imperative that the flatness calibration is functional.
        
        It should be seen now if it is possible to put the values of the calibration FLATNESS to 0 at first. Here are normally the commands to get there:
        
        - Frequency -2001Hz
        - CAL
        - page 2/4 CORRECT OFF
        on last page 4/4 (for me)
        Amp cor must be to OFF (if you have this functionnality)
        - CAL
        - page 3/4
        - CAL Service
        - Flatness DATA
        - Init FLT (normally set all values to 0.0dB)
        - MORE
        - EXIT
        To verify (no need -2001Hz ... just to verify data)
        - CAL
        - page 3/4
        - CAL Service
        - Flatness DATA
        -Edit Flatness DATA
        
        The low line on the picture normally represents the 0dB value. If you still have 18dB, there is probably a problem. In my case it is currently 8595E and the problem similar to yours was with an 8591A. But as you mentioned in a previous post, any write attempt gives you 12dB, and the INIT FLT function gives you 18dB. I tried to compare your Calibration settings with mine (even if the devices are different versions), and I do not see anything particular except for unidentified numbers.
        
        I'm curious, if you put correction to OFF, does the amplitude jump below disappear?
        - CAL
        - page 2/4 CORRECT OFF
        If nothing works, I would try to start again with DEFAULT CAL DATA
        
        Yves
        Like • Show 0 Likes0
        Actions
EremitaAnalogico Nov 26, 2019 11:27 PM
Mark Correct
Correct Answer
Ok, the big jump was disappeared when I set all flatness data to equal values (18 or 12 dB)
If I set first 2 or 3 points in flatness data to 12 dB and I leave all others to 18 dB the big jump came back and disappear if I set Correct OFF.
All seem to work correctly, except for flatness data that can be only 18 dB (init value) or 12 dB (result of any data insertion).
In the Edit Flatness, If I set all values to 12dB, the respective segment of line fall below both graticule and bottom edge of screen (like big jump in waveform if I keep the topmost portion of waveform in the bottom portion of graticule).
The only another discrepance that I can see is that the reference frequency (that measured externally is exactly 300 MHz) is seen by 8590B at 352,46 MHz
Perhaps I could try to do another hard reset of NVRam by removing backup supply?
Like • Show 0 Likes0
Actions
- Tardz @ EremitaAnalogico on Nov 27, 2019 12:31 AM
  Mark Correct
  Correct Answer
  I thinks it's a good idea to do a hard reset by removing backup supply. But FREQ CAL use 300MHz and harmonics, I would try to do the following experiment: CORRECT OFF and FREQ CAL only (no calibration in magnitude)... to see if the spectrum will correct the frequency at 300 MHz.
  
  Yves
  Like • Show 0 Likes0
  Actions
  - EremitaAnalogico @ Tardz on Nov 27, 2019 5:38 AM
    Mark Correct
    Correct Answer
    Performing a Freq Cal the result varies from time to time.
    First, if I select CAL -> Correct OFF it turn back automatically ON when I start Freq CAL
    The end of various FREQ CAL procedure, executed subsequentially, can be:
    "No Cal Sig found"
    "Cal Done"
    "Cal: FM Span Sens Fail"
    At the End, the 300 MHz Cal Signal is seen sometimes to 352,46 MHz and less often to 314,89 MHz
    Like • Show 0 Likes0
    Actions
    - Tardz @ EremitaAnalogico on Nov 27, 2019 8:11 AM
      Mark Correct
      Correct Answer
      OK, So you have no choice, you really have to solve the problem of flatness, otherwise the other automatic calibrations will not work.
      
      What I find strange is that for a compensation of 18dB (default value), you have a 60dB jump on the screen. I hope it is not the firmware that causes the problem. I will try to make a compensation of 18 dB on my 8595E and see what happens.
      Like • Show 0 Likes0
      Actions
      - EremitaAnalogico @ Tardz on Nov 28, 2019 5:37 AM
        Mark Correct
        Correct Answer
        Well, I have erased the calibration SRAM and now all Flatness values is "0 pW" instead of 18.00
        Init flatness don't modify this strange value, and I cannot enter nothing with Edit Flatness.
        I have desoldered all four 32K SRAM chips, then I will check it this evening at my home.
        I hope that the malfunction is due to faulty RAM locations...
        I come back with news in few days...
        Goodbye and thanks.
        Like • Show 0 Likes0
        Actions
        Tardz @ EremitaAnalogico on Nov 28, 2019 10:28 AM
        Mark Correct
        Correct Answer
        Hi,
        After the erasing, have set DEFAULT CAL DATA?
        
        In my side, I did a test so I changed a value to 18 dB and the result, without surprise, was a step of magnitude of 18 dB on the screen.
        
        You probably know it already, but just a note: when you press the STORE FLATNESS button, the confirmation of the stored data takes about 30 seconds in my case ... after you have to press MORE and EXIT. If you leave this routine without waiting for confirmation, there will be no backup.
        
        I made a small calculation, if the 18 that you think are dB, was Watts, then the magnitude jump would be 62.5dB if we base ourselves on the reference of -20dBm REF 300MHz, which corresponds to pictures sent, but I do not know if this calculation really makes sense.
        
        When I had the same problem with my 8591A many years ago, it seems vaguely to have had this value of 0pW in my manipulations, but I do not remember how I remedied this problem. I remember having followed the procedure several times, I had reset the NVRAM but I also changed the Dallas DS1210 controller IC, it controls access to NVRAM. But I do not remember if it had any impact, but believe no. Anyway you already have to demonstrate that you are able to write in this memory by the test CAL ATT ERR. The SVRAM test becomes necessary.
        
        Yves
        Like • Show 0 Likes0
        Actions
        EremitaAnalogico @ Tardz on Dec 2, 2019 11:46 PM
        Mark Correct
        Correct Answer
        Good Morning, I come back with good news. I have replaced the SRAM chips and now I can edit Flatness data.
        Yes, I had set Default CAL Data after erasing attempt of last week, but the flatness data was uneditable.
        I have performed the measures of all frequencies in the table (except for those above 1 GHz... I not had any generator that go over this limit, but I will bring my old frequency doubler from home) and now both Freq and Amp calibrations ends with no errors and the Cal Signal is seen correctly at 300 MHz
        The last step I think is the ATTN ERROR settings. Do you have another useful information to do this 5 values setting?
        Thanks.
        Like • Show 0 Likes0
        Actions
        Tardz @ EremitaAnalogico on Dec 3, 2019 12:17 AM
        Mark Correct
        Correct Answer
        Wow good news!
        
        For the calibration of the attenuator it is quite simple, just follow the procedure, and the entry of the calibration parameters is similar to that of Flatness. There are only 4 tests to do (1dB, 2dB, 4dB and 8dB), for the 16dB we enter the value 0dB ... so only 5 data to save. Again, the correction must be removed before calibration.
        
        Regards
        Yves
        Like • Show 0 Likes0
        Actions
        EremitaAnalogico @ Tardz on Dec 10, 2019 6:03 AM
        Mark Correct
        Correct Answer
        Ok Yves, thanks, well done.
        Thank you for your patience and for all informations, now the analyzer is again in service.
        Best regards.
        Like • Show 0 Likes0
        Actions
        Tardz @ EremitaAnalogico on Dec 10, 2019 2:34 PM
        Mark Correct
        Correct Answer
        Hi, very good new, happy for you. Glad to have been able to help you.
        
        Best Regards
        Yves
        Like • Show 0 Likes0
        Actions

8590B no calibration after memory loss

Outcomes

Related Content

Recommended Content