[SOLVED] HP 8563A Spectrum Analyzer Repair - Is the YIG oscillator dead?

[1Ghz]

EDIT: June 29, 2020

TL;DR
The problem was caused by bad YIG oscillator.
Replaced the YTO (YIG oscillator) and that fixed the problem.


I recently bought an used HP 8563A spectrum analyzer.
But it's not working and it displays an error messages.

Code: [Select]

301 YTO UNLK
317 FREQ ACC
334 LO AMPL

Here is the pictures.

1. 8563A 1st LO Output (picture from the 8563E)

It's not working correctly. No signal output from low to mid frequency range.

2. 8563A CAL Output (picture from the 8563E)

It's working correctly.

3. 8563A Power Supply LEDs

I think it's working correctly.

Is the YIG oscillator dead? Should I replace the YIG oscillator?
Thanks,

[4thDoctorWhoFan]

Since you are displaying the cal signal correctly, it's not the YIG.  Usually with a YIG problem, you will not get any signals to display.

I'm no expert though. 

[MadTux]

Better look at block diagram of 8563A for YIG frequency sweep range before declaring YIG for dead.
YIG of 8561E runs between 3-6.81GHz, somewhat close to those 5-7GHz you measured, probably also depends on what band your SA is set to.
Low band 0-2.9GHz mixes it to 3.91GHz, so YIG should sweep beween 3.91-6.81GHz.
High band 2.9 GHz and above downconverts it to 310MHz, so it should sweep 3.21-6.11GHz
Harmonic mixing in bands above, probably similar to HP 8566B
http://www.simonsdialogs.com/2018/11/hp-8561e-spectrum-analyzer-30-hz-to-6-5-ghz-some-consequences-of-a-past-disaster/
http://www.simonsdialogs.com/wp-content/uploads/2018/11/8561e-block-diagram-with-annotations.jpg

Perhaps try setting the 8563A into zero span mode, then you can measure frequencies without sweep.
YIG coil can also be tuned by current from power supply for basic testing outside of instrument, if you get crazy otherwise ;-D

Your YIG is probably fine, btw, more likely the PLL is messed up, perhaps dried up capacitors or so, usually the first that goes bad if noone has messed anything up.

[1Ghz]

Normally, YIG of 8563A runs between 3 to 6.81 GHz, too.
More pictures.

1. 8563A low band 0 Hz to 2.9 GHz sweep test
1) Setup


2) Result

No low to mid frequency sweep.

2. 8563A high band 2.9 GHz to 6.2 GHz sweep test
1) Setup


2) Result

No low to mid frequency sweep.

3. 8563A CAL output 800 MHz span test
1) Picture from the 8563A

No 0 Hz LO feedhtrough displayed.
No CAL output signal displayed.

2) Picture from the 8563E


4. 8563A CAL output zero span test
1) Picture from the 8563A

No CAL output signal displayed.

2) Picture from the 8563E

[MadTux]

PLL goes crazy on that one, YIG likely is fine.
Try measuring tuning current (as voltage) on YIG tuning coil without sweeping, it should be a constant current(voltage), on your instrument, it's likely oscillating like crazy.

The spikes you see are probably when unlocked PLL matches frequency at a sweep spot. A few MHz above or below on 8563E sweep, 8563A YIG has drifted to a totally different frequency, so it generates the wrong impression of no YIG output, when in reality it's just running all over the spectrum.

[1Ghz]

Replaced the YTO (YIG oscillator) and that fixed the problem.
I bought the YTO from Test Cal Instruments.

1. -5 dBm 20 GHz CW signal. And no errors. Yay! 


2. Local FM station


3. 2.75 GHz to 20 GHz sweep test

[shakalnokturn]

Did you open the YTO assembly to see if anything could be done to it or just for the sake of autopsy?

[1Ghz]

Bad YIG oscillator photos.

Inner circuit






Coils



Maybe FM coil is burnt?

[shakalnokturn]

FM coil does look a bit suspicious.
I'm not sure what's been going on here either, looks as if it has heated enough to change the gold plating's colour.

[dietert1]

Two years ago i fixed a similar problem on our HP 8560A. As far as i remember the YIG oscillator uses a special 28V voltage that comes from a voltage doubler. Those voltage doublers were made with electrolytical caps and those don't last as well, since on power-on they get reverse voltage every now and then. Replaced one cap and the spectrum analyzer worked again.
The instrument is servicable, yet the power supply is buried deep inside, so it takes patience to get the board out. The capacitor has 3 pins with the third pin connected to a built-in shield foil. Made something similar from a normal cap and copper foil i had.

Regards, Dieter

[1Ghz]

I'm not sure what's been going on here either, looks as if it has heated enough to change the gold plating's colour.

Yes, it's suspicious! Maybe too much current? Hmm... 

Those voltage doublers were made with electrolytical caps and those don't last as well, since on power-on they get reverse voltage every now and then. Replaced one cap and the spectrum analyzer worked again.

If it fails again, I'll check the caps.

[romal76]

I'm having 8561E with bad YIG (p/n 5086-7906).
Is it possible to replace bad YIG 5086-7906 with another one 5086-7903? Which steps? How to make tuning after replace if it possible?

[1Ghz]

Is it possible to replace bad YIG 5086-7906 with another one 5086-7903? Which steps? How to make tuning after replace if it possible?

Sorry, I have no idea. I replaced it with same YTO.

[romal76]

It's working good, I was check it. 5086-7906=5086-7903
Next problem for me is SYTF, broken switch.

[1Ghz]

It's working good, I was check it. 5086-7906=5086-7903

Nice to hear that. Thanks for share your experience. 

Next problem for me is SYTF, broken switch.

Hope you can fix it soon. 

[larry540]

I am new to this forum and recently purchased a 8561e which 28v supply is running at around 7v. It turns out if I unplug the SYTF(5086-7803) the supply goes back to 28v. The SYTF looks to be pulling well over 800ma on the Tune input. I don't belief this is normal. Does anybody have any comments? Is there any hope of repair?