AnsweredAssumed Answered

HP 8591A problems

Question asked by dutchmen on Aug 22, 2010
Latest reply on Mar 18, 2011 by hardyhansen

Like • Show 0 Likes0
Comment • 18

Hello all,

I'm a owner of a HP 8591A (with option 001 and 021), but some problems occured a few days ago and i hope that someone here can help me to find out what the problem is and how i can repair it.
Some explanation about the problem: Since a few days the calibration signal is very unstable and also some of the 'BW- correction numbers in the displayed cal data' is out of range.
Because of this, i took the service manual and did some easy calibration, such as the power supply- and Crystal an LC Bandwidth calibration.
After i did this, the BW correction numbers that falled out of range are a little progressed, but not that much that they fall inside the -0,80 and +0,50 range.
Also the calibration signal is still very unstable.

I hope that someone here can help to find out what the problem is?

Thanks in advance,

Marcel.

No one else has this question

Outcomes

18 Replies

sm7ovk Aug 23, 2010 11:38 AM
Mark Correct
Correct Answer
Hi Marcel,

do you have access to another instrument, (as another spectrumanalyzer or a powermeter), so you can measure the 300MHz cal-output? If so check the outputlevel to see if it's unstable.

Regards
Jens
Like • Show 0 Likes0
Actions
MikeM Aug 21, 2014 6:44 PM
Mark Correct
Correct Answer
When you say unstable are you talking in frequency or amplitude?

If it's the frequency that's unstable does it make a difference if you have a span wider than 10 MHz vs a 10 MHz span? In spans of 10 MHz and narrower it uses the countlock board to drive the FM coil of the YTO to lock it, while spans wider than 10 MHz don't use the FM coil drive or the countlock scheme. If your problem is on only one side of this break point this will help to find the problem. With the power turned off you can disconnect the large ribbon cable going to the countlock board and then turn it back on and since it won't find the board it won't try to use it, which might then make it more stable at a 10 MHz span and narrower, which would further isolate the problem to something in the countlock/FM coil loop...
Like • Show 0 Likes0
Actions
dutchmen Aug 24, 2010 2:55 AM
Mark Correct
Correct Answer
Hello sm7ovk and Mike M,

The frequency at the top of the cal. signal is very stable. When i put the 'Cal out' into my frequency counter it says 300 mhz with very little deviation.
It is the amplitude who is very unstable. It looks like there is some modulation all the time and sometimes it looks like a lawn with big grass turning up and down.
I have tried to disconnect the cable to the counterlock board, but this doesn't make any difference.
I hope this is a clear explanation and i don't know if i can take a picture and put it down here?

Best regards,

Marcel.
Like • Show 0 Likes0
Actions
sm7ovk Aug 25, 2010 12:16 PM
Mark Correct
Correct Answer
Hi,

when you write "looks like modulation" is this when you look at the cal-signal with the same spectrum analyzer, (SA), or measured with another SA? It's important to be sure if the cal-signal is right or wrong.

First - I assume that you have already opened your SA - but remember there are dangerous voltages present! Do not do this if you haven't been working with similar things before! When the cover is removed place the SA on a firm table on the side. Printed instructions inside the SA should give you a hint how this is done.

Inside the SA a lot of the boards are interconnected with coaxial cables with SMB-connectors. I had one where the shield was only slightly connected and not firmly tighten, possibly since that SA had been opened many times. If the cal-reference isn't stable, (by frequency and level), the SA will not be able to make it's "calculations".
If you haven't any power meter or another SA, second test would be, (with the cal-output connected to the SA-input), move the SMB on the board gently back and forth to see if there is any difference. This can be very tricky to see and you should only twist the cables a little. If you find any SMB that has problem with the shield, note the number. (Be gentle with the connectors.)
If there isn't any connector that makes any difference next step is to check the levels on the different connectors. To do this you must have access to another SA and a CLIP. I'm not sure if there is any CLIP for this on the net, but you have perhaps found some? On many of the HP-SA's there is a oscillator with a chain of amplifiers for amplification and multiplication depending on the SA version. To follow such a chain you'll have to have a oscilloscope so you can see how the level and frequency changes. If you find a faulty amplifier you must be able to solder. Plan these steps already. A faulty shield or connector can easily be fixed and you can exchange a cable, but if there is a component error you must be more prepaired before you go any further. Just so you know...

Regards
Jens
Like • Show 0 Likes0
Actions
dutchmen Aug 26, 2010 8:34 AM
Mark Correct
Correct Answer
Hello Jens,

Thanks for the advise. I know that there are high voltages inside the SA and i have a oscilloscope and a soldering iron, but not a second SA.
In the meanwhile i have checked the smb cables and tryed some new cables, but this doesn't make any difference. Also i take a few srews out of the unit where te calibration signal is come from and when i gently push it a little bit sideways there is a lot of interference on the screen, so i think the problem is in that unit...

Best regards,

Marcel.
Like • Show 0 Likes0
Actions
hardyhansen Sep 3, 2010 7:32 PM
Mark Correct
Correct Answer
Hello Jens
Clips for 8591a are on the net.
Hardy
Like • Show 0 Likes0
Actions
sm7ovk Sep 5, 2010 4:42 AM
Mark Correct
Correct Answer
Hi Marcel,

Sorry for not answering you earlier but I've been on a business trip all this week.
Hardy say that there should be a CLIP, (Component Level Information Package), on the net. I did a search but didn't found any free package. If you are about to buy a CLIP for your SA please check that the one you are buying covers your SA. There are some differencies and often you must give the dealer your serialnumber to be sure.
I haven't any 8591 myself so I can't open and check, but if you find a free CLIP please let me know - send the link, and I'll look at it to see if I'll be able to give you some more hints. Please don't try to adjust anything inside the SA yet.

You might be able to fix this with your oscilloscope - how high frequencies can it and your probe measure? Your scope might handle 40-50MHz but your probe must also be capable of handling 100-600MHz to be able to measure on these signals. You could make a passive resistor devider that will easily go up to these frequencies. An easy check is to see if you see anything when you connect the CAL output to your scope? (If your probe is marked 100MHz or less there is no reason to test this since the probe will act as a good LP-filter...)

In the 8590 there is a "comb generator" producing a 100MHz signal and it might be done in the same way in the 8591A. Do you have a small unit numbered something like 08569-...?

From what you're saying regardning the screws - it might be a grounding problem. (Please be careful and not unscrew anything with the power on.) Strange things can happen if the units aren't properly grounded. Check all connections to the boards so there isn't any bad connection. Look at the connectors with magnifiying glass and look for cracked solderings.

Regards
Jens
Like • Show 0 Likes0
Actions
hardyhansen Sep 5, 2010 4:57 AM
Mark Correct
Correct Answer
Hello Jens
I have downloaded the clip for both 859xa and 859xe analyzers,believe the a model came from agilent manuals site.
I can upload via dropbox if You wish.
Yours
Hardy
Like • Show 0 Likes0
Actions
sm7ovk Sep 5, 2010 12:03 PM
Mark Correct
Correct Answer
Hi,

thanks to Hardy I now have the CLIP - just hope that it's the same as for your SA.

Try to find J2. This should be the 300MHz output to the CAL-connector on the front. Measure on the output direct to see if it's the same behaivor as on the front. It might be a bad connection on this cable. This would one possible explaination why the CAL-signal looks strange. Have you any other source to check? Preferable a stable oscillator to see if it's only the CAL-signal or if it's all signals.
Look at 08590-60192-1, see folder K23. Here you see the oscillator-chain with the output for the CAL. This is one version, there might be several versions!

Please let me know how it looked on your scope as I asked you in my last posting..

Regards
Jens
Like • Show 0 Likes0
Actions
dutchmen Sep 6, 2010 1:33 AM
Mark Correct
Correct Answer
Hi Jens,

When i'm at home again i will see if there is a 'comb generator' with a number like you said in the post above. I also check the solders and do the test with the scope than, but i think the phrobe goes up to 75 Mhz and the scope to 120 Mhz.
I allready tryed a cable directly from the output on the board to the input of the SA, but the signal didn't change anything.
The adjusting that i have done is the power supply- and Crystal and LC Bandwidth calibration like i sayed in the beginning, but this wasn't helping anything.

When i'm at home i let you know if i can see anything on the scope.

Best regards,
Marcel.
Like • Show 0 Likes0
Actions
sm7ovk Sep 6, 2010 9:54 AM
Mark Correct
Correct Answer
Hi Marcel,

I don't think you'll see much with a 75MHz probe. Try to use a cable instead. If the level is unstable put an attenuator in front of the scope or make a 10:1 with resistors.

Regards
Jens
Like • Show 0 Likes0
Actions
sm7ovk Sep 11, 2010 8:49 AM
Mark Correct
Correct Answer
Hi,
I thought a little about what you should measure if you don't have any signalgenerator to check if it's the CAL output that is unstable or something else. I think the easiest thing is to look at the shorwave part of the frequency-band. Make a simple antenna. Take a couple of meters of thin wire and form a loop from the mid connector on the input to the ground of the same connector. (Remember that several of SA's do not handle any DC voltages on the input! If the input is marked with 0VDC max then it will not cope with any DC voltages! Therefore it's safe to use a loop from the mid pin to the ground.)
Look at about 10MHz and try to find an AM station. You see them clearly as a carrier with two sidebands looking similar. Look at the carrier and try to find a really strong signal. If you don't touch the "antenna" this carrier should be very stable. Try to narrow the bandwidth to see if there are any changes in the behaviour.

Regards
Jens
Like • Show 0 Likes0
Actions
dutchmen Sep 15, 2010 5:22 AM
Mark Correct
Correct Answer
Hi Jens,

Excuse for the late respons, but i'm quite busy at the moment. I looked at the text near the input and it says '100 VDC max', so i'll try the expiriment as soon as possible and let you know what i will see on the SA.

Best regards,
Marcel.
Like • Show 0 Likes0
Actions
huebnerie Sep 24, 2010 1:16 AM
Mark Correct
Correct Answer
Hi,

hardyhansen, you wrote that you've downloaded the 859xA clip.
I have found the 8590A clip on Agilent's web site, which is not what I'd call 859xA series - 8591A, 8593A and so on would be 859xA for me.
Is the clip you got actually for the 8591A and upwards, or is it just the 8590A clip I've already found?
I bought a 8591A recently with only minor problems, but when I think back to repairing a 54200A's cpu board on component level without any factory information, I don't want to do this a second time on the 8591A.

regards
Stefan
Like • Show 0 Likes0
Actions
dutchmen Oct 3, 2010 7:27 AM
Mark Correct
Correct Answer
Hello,

Finally, after a couple of weeks i tryed to find an AM station near the 10Mhz.
I took a couple of meters of thin wire, but when i make a loop from the mid to the ground of the input there was nothing to see on the screen.
When i put the wire only in the mid of the connector there where some strong signals, but not one of them was stable enough.
I have tryed several times with several lengts of wire, but it doesn't make any difference..

Best regards,
Marcel.
Like • Show 0 Likes0
Actions
sm7ovk Oct 7, 2010 12:05 PM
Mark Correct
Correct Answer
Hi,
well it's difficult to give you a stright answer how to proceed.
Do you have an old TV, the type with CRT-tube? I'm not sure about new flatscreens but old TV-sets would have 4.43MHz, (PAL in Europe), that might be possible to receive. Even if you will receive this signal it might still be hard to say if you'll see natural differences or if you analyzer is not stable. After all the best thing would be if you could check with some type of generator, but I also understand that this can be hard to get.
Do you have any experience of building some electronics? If so it's not that hard to build a small crystal oscillator. I'll have stable output and can be usable for other things too.

Regards
Jens
Like • Show 0 Likes0
Actions
MichalKotze Mar 18, 2011 4:24 AM
Mark Correct
Correct Answer
Good day. I am currently a product engineer at Grintek ewation (Cape Town). The product i am currently using is a HP 8591A spectrum analyzer (9kHz to 1.8GHZ). My monitor stopped working and displays nothing on the screen. I wanted to remotely control the spectrum analyzer through its serial port using MATLAB. I made up an serial cable with the connections to each pin connected as shown in Figure 1. I found this connection diagram in the “HP programmers guide HP 8590 E-Series and L-Series spectrum analyser” which can be downloaded at this link http://cp.literature.agilent.com/litweb ... -90235.pdf. This diagram is on page 720 of the programming guide.

see attachment
Figure 1

When i plugged the serial cable in the serial port of the HP 8591A ( The other end of the serial cable was not connected yet) the monitor stopped working. When I the screen worked I did change some settings like restore factory defaults and aux ctrl.

I please need help in this matter, urgently.

Kind regards

Michal Kotze

Local Cell: 083 885 6969
International Cell: +27 83 885 6969
e-mail: mkotze@mtnloaded.co.za
Attachments
- HP_serial_connections.png15.7 KB
Like • Show 0 Likes0
Actions
hardyhansen Mar 18, 2011 2:20 PM
Mark Correct
Correct Answer
Hello
First try to connect an external monitor to see if analyzer is working at all.
If not I believe Your analyzer are forced into a dead lock .
This usually can be fixed with a complete calibration using the two "special" calibration calls,described in the service manual.
Will look for the pages to You
Hardy
Like • Show 0 Likes0
Actions

HP 8591A problems

Outcomes

Attachments

Related Content

Recommended Content