HP 8566B Debug/Rebuild

[MarshallPlexi]

Hello All!

Been a fan of this forum for a while but am just now getting in spectrum analyzer's so I have a mountain of information to process. I picked up what I thought was a fairly decent 8566B from a local shop and found after getting into it that the input attenuator was fried. Luckily I bought a spares 8566B which I thought would help only to find that it's input attenuator was fried too.

I'm picking up a new attenuator tomorrow, and until then, I have nothing I can do to help, so I thought I'd do one of the common mods that have been recommended for the A20 board, and that is to replace C7, a 10uF/25V cap. I didn't have the exact cap so I replaced it with a 220uF/50V cap thinking that more voltage is better and more capacitance should help to stabilize the supply more.

I also installed a 10dB attenuator between the CAL signal and the input of the LIMITER. To my shock and amazement, I can now run RECALL 8 and I see a 100MHz signal there. Granted the front end attenuator is out of circuit, but this is more than I've seen in the past. The shape of the signal is odd. See attached pic.

Here is what I'm getting.

When I run RECALL 8, I get the image of the kinked 100MHz signal, and when I open up my scan to around 3MHz, I get the correct-looking signal.

Do you have any thoughts on what could be causing this?

The final image is what I'm seeing when I run the RECALL 9 routine. I'm not able to adjust the line into view.

I was also able to get MUCH further into the SHIFT+W calibration routine this time so I feel like I'm making progress!

I'm looking forward to learning more about this incredibly built piece of equipment!

Thanks for any council you can provide!

[squadchannel]

Definitely, the C7 that was replaced will have an impact. Why did you even change the capacity?
C7 is the capacitor that hangs in the tuning circuit of the RF section YTO coil; installing a capacitor 20 times larger would naturally have an effect.



I am not familiar with RF, so I don't know.

[MarshallPlexi]

Holy crap. You’re right. I’ll change it back tomorrow morning and report back!

[MarshallPlexi]

I couldn't sleep thinking about it so I had to get up and swap out C7 for a proper 10uF cap. It's sorta better.

IMG_5869 shows the scope when RECALL 8 is pressed. The kink is clearly gone but no longer visible because it's out of frame.

IMG_5870 shows the scope when the RES BW is set to 300kHz

IMG_5871 and IMG_5872 shows the scope when the center freq is selected between 99.94 and 99.96MHz. It's a jump across and not a smooth transition.

It seems like progress, though.

[MarshallPlexi]

New attenuators in use and unless I totally misunderstand how the 8566B is running things, that's not a constant step in attenuation.

Thoughts?

[MarshallPlexi]

Here is a video of the 100MHz calibrated signal going into the input of the attenuator. Does this look correct?

https://youtu.be/Irq7qqfU4XA

[MarshallPlexi]

Any ideas on if this is right guys?

[Kean]

No, the first video viewing the attenuation steps with the MSO5000 doesn't look right.

I suspect a problem with the attenuator, as these are known to often have problems - even when bought "new".
You'll find lots of info on inspection of the contacts and o-ring replacement if you search.
e.g. https://groups.io/g/HP-Agilent-Keysight-equipment/topic/8566b_attenuator_fix/97498780
and the linked video - but read the comments in the above thread.

I suggest sticking to using the sig gen and scope until the attenuation steps make sense before trying again with the SA.

[MarshallPlexi]

Great stuff Kean! I watched all of this and will tear into the attenuators tonight and see if I can find anything obvious.

[MarshallPlexi]

I got the attenuator open tonight and was able to confirm that solenoids 1 and 3 refused to extend. They would retract but that's all.

For your viewing pleasure:

[Kean]

OK, that is kind of what I expected.  But I'm sure they can be fixed.
Be careful as you disassemble further - unsurprisingly it can be easy to damage the parts.
I've still got a couple here in the pile of things to work on (for my HP 8657B).

[MarshallPlexi]

I found this video and have already purchased some O rings from Amazon that will be delivered tomorrow. Assuming that the O rings are bad, would that cause the solenoid to fire only in the retract position? Or could the driver chips have failed as well?

[MarshallPlexi]

Another good video for people like myself who are working through this.

[Kean]

I found this video and have already purchased some O rings from Amazon that will be delivered tomorrow. Assuming that the O rings are bad, would that cause the solenoid to fire only in the retract position? Or could the driver chips have failed as well?

Ah yes, I recall seeing one of his other videos on this topic



No idea if electronics are a common failure point on the newer models, but certainly possible.  The mechanical operation issues are quite a common issue with age.
From what I've seen other people reporting, yes they may stick in either state due to degraded o-rings.  My guess is that it might be due to a short actuation time not able to overcome extra friction.
Keep an eye out for any crap left behind clogging the holes or around the solenoid moving parts.

Another good video for people like myself who are working through this.

Oh, an hour long... I'll have to watch that another time.  I assume from a quick scrub through the video he has an electronics fault?

[MarshallPlexi]

Have you ever had one of those days where you fixed something and have no idea what you did that fixed it? That was my day.

[Kean]

Happy days!