EB5AGV's Workbench: HP-8566A repair

A 100Hz to 22GHz Spectrum Analyzer

HP-8566A

This page is devoted to the repair of my HP-8566A spectrum analyzer. I hope you enjoy it!

HP8566A MICROWAVE SPECTRUM ANALYZER


Unit as listed at eBay

The unit was sold as two sepparate auctions, one for the DISPLAY SECTION and another for the RF SECTION. I won both ;-)

It was listed as working DISPLAY SECTION but unworking RF SECTION, as it generated a YTO UNLOCK message.

Here you have some pictures of the unit which were put on the eBay listings:

HP-8566A

HP-8566A

HP-8566A

HP-8566A


Unit as received... ouch!!!

Packaging was extremely POOR as you can see in this picture

HP-8566A

So I was sure there would be some damage... and there was :-(

HP-8566A

HP-8566A

HP-8566A

HP-8566A

HP-8566A

HP-8566A

I did my best to fix the unit

HP-8566A

After checking for loose cards (which there were some!) and loose screws, I connected and powered the unit. It showed the STAND-BY LEDs

HP-8566A

But then, after toggling the POWER, all I got were all LEDs ON and a flashing ADRS'D LED

HP-8566A

HP-8566A

Only good thing so far is that RF unit seems to be an updated one, as you can barely see in this picture

UPDATE: it seems someone swapped the 'B' control boards, rear panel and front panel for the 'A' version. But, as I found later, some PCBs, as the YTO/POWER SUPPLY ones, are 1986 vintage, so they correspond with this label. All in all, what I have is an hybrid 'A'/'B' version which I hope to convert to a full 'B' version if I can get the needed parts. Any help on this ;-)?

HP-8566A

Here you have some more pictures of the inner works... it is a really wonderful engineered instrument!

RF UNIT, TOP SIDE

HP-8566A

RF UNIT, BOTTOM SIDE

HP-8566A

DISPLAY/IF UNIT, TOP SIDE

HP-8566A

DISPLAY/IF UNIT, BOTTOM SIDE

HP-8566A

RF UNIT, without keyboard

HP-8566A

DISPLAY/IF UNIT, REAR PANEL (partial)

HP-8566A

DISPLAY/IF UNIT, REAR PANEL (partial), showing interconnecting sockets

HP-8566A

RF UNIT, REAR PANEL (partial)

HP-8566A

RF UNIT, detailed views

HP-8566A

HP-8566A

HP-8566A


Working on the DISPLAY SECTION (I)

Well, I couldn't do anything to fix the disastrous trip, so I needed to move on and start troubleshooting!

I began with the display section. There is a way to operate it without the RF unit, by placing a couple jumpers, one to start the unit and another one to select TEST mode. Once done, pressing RESET button you should get a test display... this is what I got:

HP-8566A

Not too pretty :-(

I took some cards out of the display area, which is this:

HP-8566A

An then, once reinserted, I got a better picture!

HP-8566A

HP-8566A

Wiggling somewhat the cards I got an still better one

HP-8566A

And then... surprise!

HP-8566A

Right image come and went by barely touching the cards, specially the A6 one. So I got it out and found some oxide on its edge connector

Image is not focused and FOCUS control does nothing... but well, that is an analog problem which will take care of later.

HP-8566A

Once cleaned, display unit showed a nice test pattern consistently so I decided to connect it again with the RF unit... but not, it didn't work. So I decided to get all cards out and carefully clean them and the card cage

HP-8566A

I am now reinserting them...


Working on the RF SECTION (I)

As the DISPLAY SECTION seemed to work (at least it displayed the TEST pattern), I went back to the RF module and took a similar approach: remove all digital boards:

HP-8566A

Here you have all of them, one by one:

HP-8566A

HP-8566A

HP-8566A

HP-8566A

HP-8566A

I carefully cleaned the edge connectors, added some contact cleaner, reinserted them and hoped for the best... and it happened!!!

HP-8566A

GREAT NEWS!!!. So I pressed the INSTR PRESET button... and got this:

HP-8566A

As oven was still heating, there were some errors:

HP-8566A

I hooked the calibrator and... voila!!!

HP-8566A

So unit kind of works!!!. It shows the YTO UNLOCK error, which was already present in the display of the eBay listing and now, besides, it can't be focused... but, so far, so good :-)!!!


Working on the DISPLAY SECTION (II)

I decided to go for the FOCUS problem. Checking the service manual, I found the circuit which managed it. It is an small PCB, just over the CRT. It is the Z-axis circuit.

I got it out and found a resoldered transistor, just in the FOCUS circuit. Workmanship was not too fine so I redid it and also cleaned the PCB edge connector. Once I got the unit back to the system, I got this nice display:

HP-8566A

And now FOCUS control works as it should :-) !


Working on the RF SECTION (II)

There was just a pending error, the YTO UNLOCK. Everybody suggested it was a matter of failed capacitors in the YTO control boards. So I took out all boards in that section.

BTW, a curious thing is that most of the ICs in ALL these boards have manufacturing dates up to 1986... so the 85660B in the rear label has some truth on it...

HP-8566A

HP-8566A

HP-8566A

HP-8566A

HP-8566A

I was looking for bad capacitors, so this one seemed very suspicious (look the discoloration in one side):

HP-8566A

So I took my trusty ESR meter to check all of them:

HP-8566A

And, yes, all were OK except for the suspicious one, which was open :-)!!!. It was a 10uF/60V capacitor. All what I had in stock were 10uF/50V 105C, so I replaced it with one of them

HP-8566A

I did a thoroughly visual inspection of the boards, and found a power resistor with overheated solder joints:

HP-8566A

So I cleaned and resoldered it carefully... this was a problem waiting to show up in the future!

HP-8566A

Once I got all cards in again, I fired the unit... and YTO UNLOCK was gone!!!!

This is the calibrator signal with several spans:

HP-8566A

HP-8566A

HP-8566A

I have checked in several frequencies and spans and, so far, YTO UNLOCK is gone for good :-)


Checking the unit

Well, today I have put back the covers of the 8566A system, as I need to move the unit in short time for an upcoming shack/lab rebuilt. But, before moving it to its temporary storage location, I have checked it with the aid of my HP-8672A generator (another nice instrument!). Well, I have found some things:

All in all, my main use for this SA will be in frequencies below 2.5GHz but, of course, it would be great if I get it working fine in all its range. So I will investigate more about this... work continues!

After some interesting exchanges with John Miles in the HP-Agilent Yahoogroup, it seems that the difference I have observed in the 2.5GHz signal can be corrected by a preselector trimmer calibration. By now, I have resorted to using PRESEL PEAK as it gets the signal almost at the correct level. But the associated DAC needs to be set at about 50, far from the factory standard of 32 (range is 0 to 63)

I have then checked the ERROR CORRECTION ROUTINE, which measures the amplitude and frequency error factors and corrects the display for them, according the the HP-8566 Manual. But it always ends with the message ADJUST AMPTD CAL and aborts the procedure. In that case, the manual calls for a Manual Calibration Procedure, explained also there. It consists of an amplitude calibration (pressing RECALL 8 and using the AMPTD CAL control) and a frequency correction (pressing RECALL 9 and adjusting by means of FREQ ZERO control), both based on the CAL signal. Here you can see the displays for both adjustments:

HP-8566A

HP-8566A

They can be successfully passed, but then, when I try again the ERROR CORRECTION ROUTINE, it fails in same point. Here you have the display showing the fault:

HP-8566A

In previous steps, signal keeps inside the graticule but in this one, it is all over the top of it... suspicious!

I will investigate what is happening exactly at that moment, as previous steps run fine, in order to try to correct the problem and have the unit finish the procedure.

As I suspected and John suggested, the extra gain was a problem. It seems gain is too much to be compensated by the auto-calibration system. I have checked the CAL signal at different bandwidths and, yes, 30kHz RBW has about 3dB more gain that the other RBWs:

HP-8566A

HP-8566A

HP-8566A

HP-8566A

HP-8566A

So now I need to find how to modify that gain...


Working on the DISPLAY SECTION (III)

After the problems found in previous section and the good advice I got from the HP-Agilent Yahoogroup (thanks to John, Vladan and all people there!), I decided to go back to the DISPLAY SECTION. Luckily, all the adjustments are on the top of the unit and it is over the RF SECTION so it was just a matter of removing top cover:

HP-8566A

This is the area with the 9 PCBs involving IF and the rest of analog processing:

HP-8566A

Once removed all boards, you can see the shiny gold-plated PCB:

HP-8566A

Here you have all the boards, one by one:

HP-8566A

HP-8566A

HP-8566A

HP-8566A

HP-8566A

HP-8566A

HP-8566A

HP-8566A

HP-8566A

Someone had been here before, as there were some test pins with globs of solder. Of course, I cleaned them. But here you have the tiny work done by the tech:

HP-8566A

HP-8566A

I got some info from Vladan regarding the 30kHz RBW problem. He told me:

Message from Vladan Hi Jose,

Just to make sure, check that the problem is only (or mostly) in the 30kHz BW setting. If it's only in 30 kHz, it's probably one of the following adjustments:

A4A4C20
A4A4C74
A4A4C73

A4A8C29
A4A8C44

This is an adjustment in the 21.4MHz IF filters (top box, spread out over two boards). There is one adjustment in each of the five filter stages. You will need to build bypass networks (4 of them). The network consists of a resistor and a capacitor. Make sure you build the correct one for your vintage boards. In addition, there are separate bypass networks for the 3MHz filters, which you will need if you decide to adjust that in the future (leave that for later, after the shack is remodeled). The network is described in the manual.

This adjustment requires you to _minimize_ the amplitude of the signal. Each stage has to be adjusted while the others are bypassed with those networks. If one of the stages is not adjusted, the amplitude will be too high.

If the amplitude is high in all bandwidths from 3kHz to 30kHz, it's something else.

Vladan

So, once all the cards had been carefully cleaned, including edge connectos, I got them back to the unit and checked again, just in case it was only a matter of bad contacts... but not, problem was still there. So I marked the current position of the trimmers Vladan commented and, while 30kHz was selected as RBW, carefully moved a bit A4A4C20... I got a nice peak distortion so it went back to original position. Then I moved to A4A4C74 and, just by readjusting it about 5 degrees, I got a more symmetrical display AND I was able to get the gain back to show a perfect -10dBm signal... Yes, I know this is NOT the correct way to deal with this and I would need to do as Vladan suggested, but I have no time now to devote so much work to the unit and, well, I was eager to have a fast solution, even if not perfect.

So I run again the ERROR CORRECTION ROUTINE pressing SHIFT-W. Here you have a link to a very low quality video of it... I hope at least you can see something!

ERROR CORRECTION ROUTINE video

After it passed :-), I pressed SHIFT-w to find out the correction parameters. They seem fine enough, as there is none which exceeds 1dB:

HP-8566A

So, all in all, it seems we are progressing :-)!


More to come soon... keep tuned!


Go back to WORKBENCH INDEX