Teardown & Repair of an Agilent N9020A MXA 20Hz - 26.5GHz Spectrum Analyzer

[Hugoneus]

You can watch the video here: [40 Minutes]

youtu.be/0a6bpZZS23g

More videos at The Signal Path:

www.TheSignalPath.com
www.YouTube.com/TheSignalPath
www.Patreon.com/TheSignalPath

[bitseeker]

That was a fun one. It's a shame that repairing equipment becomes more and more difficult over the years, but it's always a treat learning about how these things work (and I follow along a little better after each video). Thanks, Shahriar!

[Hugoneus]

That was a fun one. It's a shame that repairing equipment becomes more and more difficult over the years, but it's always a treat learning about how these things work (and I follow along a little better after each video). Thanks, Shahriar!

Thanks. I will keep the module for potentially repairing it in the future.

[suj]

I think the unreliable doubler might be type 1GG7-4129. Datasheet available on the Keysight website. As for the amplifier, you missed the symbol on the chip under the microscope. It should also be possible to identify. On ebay I saw sellers of various strange chips. I saw a large selection of Agilent chips, maybe some remains of a fire that affected Keysight's headquarters or another sale. I think in the US it's possible to buy these chips directly from Keysight. If they send you a 110 GHz/256 Gsps oscilloscope to test, I think you should be able to buy these chips (if you can identify it). 
The question of assembly remains. I have wire bonder and have already made my first attempts but it is not at the level to be sure of the results. Again, the location in the US is helpful here in my opinion. I think a well-equipped laboratory at a technical university should have access to a wedge bonder. The chips are not very high power, so the attachment to the substrate (rather the substrate does not have a window to attach to the metal base) should be possible with a conductive epoxy resin.
I keep in mind that I have repaired my E4407B thanks to your tips - the prescaler was damaged. Thank you again.

Edit: Amplifier is HMMC-5617/HMMC-5618/1GG6-8002/TC905 type. I was sure that I had seen this chip somewhere before.

[Hugoneus]

I think the unreliable doubler might be type 1GG7-4129. Datasheet available on the Keysight website. As for the amplifier, you missed the symbol on the chip under the microscope. It should also be possible to identify. On ebay I saw sellers of various strange chips. I saw a large selection of Agilent chips, maybe some remains of a fire that affected Keysight's headquarters or another sale. I think in the US it's possible to buy these chips directly from Keysight. If they send you a 110 GHz/256 Gsps oscilloscope to test, I think you should be able to buy these chips (if you can identify it). 
The question of assembly remains. I have wire bonder and have already made my first attempts but it is not at the level to be sure of the results. Again, the location in the US is helpful here in my opinion. I think a well-equipped laboratory at a technical university should have access to a wedge bonder. The chips are not very high power, so the attachment to the substrate (rather the substrate does not have a window to attach to the metal base) should be possible with a conductive epoxy resin.
I keep in mind that I have repaired my E4407B thanks to your tips - the prescaler was damaged. Thank you again.

Edit: Amplifier is HMMC-5617/HMMC-5618/1GG6-8002/TC905 type. I was sure that I had seen this chip somewhere before.

Thank you. Very useful.

[vaualbus]

Still you haven't expalin how the instrument detect the errors....
I would suggest that it look at non fundamental mixer result?
 Maybe they have a loop co-me out before the filter after the mixer? Or they just digitize the spectrum and look for higher order harmonics?
Any way it so interesting looking at this chips, I would like to have photos on what are under the third shield!