EEVblog Electronics Community Forum

Electronics => Repair => Topic started by: Dago on September 05, 2013, 05:08:54 pm

Title: Spectrum analyzer repair
Post by: Dago on September 05, 2013, 05:08:54 pm
Hey! Today I received a HP 8592A spectrum analyzer I got pretty cheap on ebay. Seems to have survived the long shipping.

On first boot it said "ADC-GND FAIL". Not sure what this was about, pressed the "Preset" button and the error disappeared.

The spectrum analyzer _seems_ to be working fine as far as I can tell. Attenuator seems to work etc.

The problem is that when I try to run the calibration routines. With the "CAL FREQ" routine it's doing its thing for 20-30s and clicking away on the relays but the routine ends up with "CAL: SPAN SENS FAIL". The "CAL AMPL" routine also fails, it just straight up says cal signal not found (but there it is), but this might be related to the other routine failing.

I found this about the error:

Code: [Select]
CAL: SPAN SENS FAIL
During the CAL FREQ self-calibration routine the main-coil span-sensitivity adjustment routine has failed, indicating that the spans controlled by the main coil (LO spans greater than 10 MHz) are not working correctly. The span-sensitivity adjustment routine checks for three displayed signals: 0 Hz, 300 MHz, and 600 MHz. If three signals are not found in a 750 MHz span, the error message is displayed.

Does anyone have any ideas what could be wrong? I haven't popped the hood open yet, maybe it is something obvious...
Title: Re: Spectrum analyzer repair
Post by: krenzo on September 05, 2013, 06:57:40 pm
Did you get it from the seller in Hong Kong?
Title: Re: Spectrum analyzer repair
Post by: PA0PBZ on September 05, 2013, 07:20:33 pm
About the ADC-GND Fail I found the following: http://www.home.agilent.com/agilent/editorial.jspx?cc=US&lc=eng&ckey=1000001789:epsg:faq&nid=-11143.0.00&id=1000001789:epsg:faq (http://www.home.agilent.com/agilent/editorial.jspx?cc=US&lc=eng&ckey=1000001789:epsg:faq&nid=-11143.0.00&id=1000001789:epsg:faq)

The other one looks like the YIG Oscillator is not behaving as expected, so time for some service manual reading I think  :)
Title: Re: Spectrum analyzer repair
Post by: PA4TIM on September 05, 2013, 08:23:53 pm
Wow, a 22GHz analyser.  First thing I do with every repair is checking the power supply rails. At least yours is more or less working. I just finnished a Tek 2710, that is only 1.9 GHz but it had no trace, no readout and a very sick psu. Took me over a months (about 1-2 hours a day) but the reward was wordth it. Hope you succeed.

Fred
Title: Re: Spectrum analyzer repair
Post by: KJDS on September 05, 2013, 09:40:17 pm
The first LO is probably the cause. Check to see what's happening with a real signal going in on narrow spans. Then increase the span width and check again.
Title: Re: Spectrum analyzer repair
Post by: Dago on September 06, 2013, 04:42:36 am
Quote from: PA4TIM on September 05, 2013, 08:23:53 pm
Wow, a 22GHz analyser.  First thing I do with every repair is checking the power supply rails. At least yours is more or less working. I just finnished a Tek 2710, that is only 1.9 GHz but it had no trace, no readout and a very sick psu. Took me over a months (about 1-2 hours a day) but the reward was wordth it. Hope you succeed.

Fred
Quote from: KJDS on September 05, 2013, 09:40:17 pm
The first LO is probably the cause. Check to see what's happening with a real signal going in on narrow spans. Then increase the span width and check again.

Thanks for the ideas!

I need to check the PSU and the LO. Especially the LO sounds like a good candidate for investigation...

Does anybody have a repair manual or any documentation relating to these analyzers? So far I've found "8590 Series Analyzers Assembly-Level Repair" but pretty much every topic seems to apply to the E/L-series and no mentions about the A-series. No idea how different they are.

Krenzo: yes.
Title: Re: Spectrum analyzer repair
Post by: rf-loop on September 06, 2013, 06:55:26 am
Quote from: Dago on September 06, 2013, 04:42:36 am
Does anybody have a repair manual or any documentation relating to these analyzers? So far I've found "8590 Series Analyzers Assembly-Level Repair" but pretty much every topic seems to apply to the E/L-series and no mentions about the A-series. No idea how different they are.

Also I have only original (paper manuals) repair service manuals buyed from Agilent (including CLIP) for 8590E/L series) becouse I have repaired these but of course they are not compatible for A.    (E/L models service manual can today find also downloadable becouse today it is discontinued model but I'm not sure if all CLIP is included)

Some  A model (not remember exactly model number) I have repaired but in this case it go without manual and just using experience. And your problem (or failure) indications are very different. Somewhere I have seen A models some service manual but I can not remember where.  I will PM if I remember.
Title: Re: Spectrum analyzer repair
Post by: alm on September 06, 2013, 02:02:06 pm
Artek Manuals (http://artekmanuals.com/manuals/hp-manuals/) sells a (likely high quality) PDF scan of the 8590A service manual with schematics for $15.
Title: Re: Spectrum analyzer repair
Post by: krenzo on September 06, 2013, 02:42:11 pm
Quote from: Dago on September 06, 2013, 04:42:36 am
Krenzo: yes.

Ahh, I had my eye on that one too and kept wondering why it was going for so low.  Best of luck in getting it in complete working order.
Title: Re: Spectrum analyzer repair
Post by: Dago on September 07, 2013, 06:52:57 pm
I got the analyzer to pass calibrations :)

But now I get intermittently errors which are most likely related to EEPROMs going bad (reloading by pressing reset usually gets rid of the error).

Does anyone have any information or even hunches on what kind of EEPROMs these are?

(http://www.dgkelectronics.com/storage/electronics/8592A_repair/eeproms.jpg)

Any idea what I could use to replace them? "Component level information packet" -manual lists them as the types that are printed on the stickers (for example 08592-80034), made by HP.

Edit: Ah sorry, these are EPROMs and not EEPROMs so these most likely contain the firmware, not calibration values...
Title: Re: Spectrum analyzer repair
Post by: Kjelt on September 07, 2013, 07:03:24 pm
Quote from: Dago on September 07, 2013, 06:52:57 pm
Does anyone have any information or even hunches on what kind of EEPROMs these are?
AMD mark and 28 pins DIL my bet it are either 27C128, 27C256 or 27C512.
Check the connections with the datasheets and see which matches or try an intelligent eprom programmer but I am not sure if they have manufacturing data already in in 1988  :)
If I am not mistaken it is possible to read out old eproms using different kind of a little bit lower voltages, store the data each time on harddrive and binary compare the found data's. Those bits that toggle are your main suspects.
Title: Re: Spectrum analyzer repair
Post by: PA0PBZ on September 07, 2013, 07:12:28 pm
1818-3448    AMD 27C512-200DC

http://www.sphere.bc.ca/download/hp_xref-free.pdf  page 32

Title: Re: Spectrum analyzer repair
Post by: free_electron on September 07, 2013, 08:03:32 pm
Eproms dont go intermitedly bad. They work or they are corrupt.
Title: Re: Spectrum analyzer repair
Post by: Dago on September 07, 2013, 08:03:54 pm
Thanks for the help guys!

I actually thought these were EEPROMs, I have now gotten a "FAIL: 1310 0000000000" couple of times where 1310 code was marked (by hand!) on the support manual to indicate EEPROM read error which is indicative of failure. I checked the component level information packet and service guide for the 8590A model but I was actually unable to find any mention of an EEPROM. Maybe I just missed it, need to look harder...

I think I even saw some function in the manual/menus which had something to do with replacing the EEPROM but I'm unable to find it right now. Doesn't help that all the manuals are scanned without OCR so you can't search through them...
Title: Re: Spectrum analyzer repair
Post by: KJDS on September 07, 2013, 08:09:54 pm
Quote from: free_electron on September 07, 2013, 08:03:32 pm
Eproms dont go intermitedly bad. They work or they are corrupt.

That is usually true, however I've known them lose some output drive ability so the levels would be iffy, sometimes working and sometimes not, usually obvious on a scope. I've more commonly seen a dodgy decoupling cap cause similar issues.
Title: Re: Spectrum analyzer repair
Post by: PA0PBZ on September 07, 2013, 08:56:37 pm
Quote from: free_electron on September 07, 2013, 08:03:32 pm
Eproms dont go intermitedly bad. They work or they are corrupt.

I just happen to have one here on the desk that toggles one bit on one location while reading it a couple of times in the programmer, but I agree that it does not happen very often.
Title: Re: Spectrum analyzer repair
Post by: Orange on September 08, 2013, 08:53:23 am
Hi,

For the EPROM content you can have a look at www.ko4bb.com/manuals/index.php?dir=04%29_ROM_Images_and_Drivers (http://www.ko4bb.com/manuals/index.php?dir=04%29_ROM_Images_and_Drivers)

There is a ZIP file with the binaries, They are indeed 27C512
Title: Re: Spectrum analyzer repair
Post by: SeanB on September 08, 2013, 09:28:26 am
Take out of the sockets, clean the pins and reinsert them, and while they are out check the sockets for broken pins or dry joints. AMD EPROMS are pretty reliable in the most, often the problem is the socket or the solder joints or traces breaking as they are inserted.
Title: Re: Spectrum analyzer repair
Post by: Kjelt on September 08, 2013, 09:52:50 am
1988 that is 25 years since they were programmed.
My personal experience with eproms esp. the ones that are uv erasable is that the original charge decays over time due to leakage.
For my own equipment I always use 20 years as a rule of thumb and then reprogram them to give a full charge (no need to erase them just reprogram them if you are sure you have the right data).
I must say that these experiences are mostly with 27C32 eproms perhaps the later eproms are better but still there is in my opinion not such a thing as an everlasting datakeeping uv erasable eprom since that would mean there is no leakage at all: infinite resistance on the isolation of the gate.
But you people have different experiences? I really would like to hear about them.
Title: Re: Spectrum analyzer repair
Post by: Dago on September 08, 2013, 05:47:54 pm
Quote from: Orange on September 08, 2013, 08:53:23 am
Hi,

For the EPROM content you can have a look at www.ko4bb.com/manuals/index.php?dir=04%29_ROM_Images_and_Drivers (http://www.ko4bb.com/manuals/index.php?dir=04%29_ROM_Images_and_Drivers)

There is a ZIP file with the binaries, They are indeed 27C512


Good spot! Currently (I think) no problems with the EPROMs but good to have a backup.

I located the EEPROMs on the motherboard and they are made by Xicor, model X2864AD. I found some one eBay and ordered some + a ZIF socket. I'll desolder them, read them out with an arduino or something, reprogram the new ones and pop them in.

The EEPROMs still seem to work fine but sometimes on boot I get "FAIL: 1310 0000000000" which means problems with EEPROMs but the error goes away after pressing "Preset". I also sometimes get "FAIL: 0310 0000000000" but I have no idea what that means, the error code table on the support manual is uhh kind of odd... Does anybody have an idea what error code 0310 means?

(http://www.dgkelectronics.com/storage/electronics/8592A_repair/error_code_table.jpg)
(http://www.dgkelectronics.com/storage/electronics/8592A_repair/error_code_table2.jpg)
Title: Re: Spectrum analyzer repair
Post by: Bjoern on July 09, 2018, 01:30:28 pm
Were you able to fix the CAL: SENS SPAN FAIL problem ?
I have the same.

Title: Re: Spectrum analyzer repair
Post by: Bjoern on July 14, 2018, 05:55:25 pm
Have you found the error ?
I got the same issue and trying to get closer to the cause.