Trying to fix Agilent n5181a - Error 508 frac-n loop unlocked

[DavidAlfa]

No obvious damage, the instrument always boots with this error, defaulting to 3GHz -110dBm.
Nothing comes out of it at any frequency or power.

There's no schematic and the service manual just says "Instrument may need servicing".
The lowest the service manual goes about this is : "Replace RF assembly".

Based in the available info from other models, this seem a frequency synthesizer error.
Found the 10 & 100MHz reference clocks, they seem healthy.

The board is huge, has thousands of parts and there's no silkscreen, would appreciate any info shedding some light at this.

I'm pretty sure circuits will be very similar to other models, but again zero schematics found.

Found this E4400B thread having the same issue:
https://www.eevblog.com/forum/testgear/a-look-inside-of-an-agilent-e4400b-signal-generator/

Linking to this board:
https://groups.io/g/HP-Agilent-Keysight-equipment/topic/synthesizer_module_in_esg/28079593

Tomorrow will check if this model has the problematic 1GC1-4210 ICs.

Thanks!

[Karnathe]

Sorry for the necro, wanted to provide / ask for some help. Did you ever get this fixed? I have two of these units, one works perfectly the other gets this error ~30 mins after running, having been off and unplugged for a while. It starts out perfect, lovely 3ghz, then after 30min throws this error and the frequency jumps around unusably.

If you want me to probe around / take photos of the good one / try to compare, let me know. Have not seen much info about fixing these things.

[DavidAlfa]

Hi, nope, I closed the unit and forgot about it, neither I will touch it again, it's from my work and I won't spend a single minute fixing their stuff ever again after they lied to me too many times (Give your best and you'll see... Blah blah).
Sorry! 

[Miek]

I've got a pair of N5182As with the same error that I've been working on for a little while. On mine, they boot up without an error and the output works, but if they are put into sweep mode then it starts throwing the error and it only locks intermittently on each frequency change - for mine I think I've narrowed it down to the 1GC1-4218 phase-frequency detector and I'm currently waiting for parts.

I've sketched out the rough signal path for the Frac-N loop which should hopefully help a bit, I'd suggest probing along that and see if the expected signal goes missing anywhere.

Also, if you haven't already, check out TSP #163 and #161 for an N5182A repair and also an MXA repair which uses a very similar VCO & PLL setup:

[DavidAlfa]

Have you checked the 1GC1-4210 freq dividers?
Seems to be a very common failure.

There's a replacement pcb you can build yourself:
https://groups.io/g/HP-Agilent-Keysight-equipment/message/112571

You can also buy them:
https://www.leobodnar.com/shop/index.php?main_page=product_info&cPath=125_127&products_id=324

[Miek]

I'm pretty sure this series doesn't have any 1GC1-4210 dividers, there definitely aren't any in the Frac-N loop. I've seen some similarity to the ESG design, but that part didn't carry over.

[DavidAlfa]

It's been a year, I barely remember anything...
Great work BTW!

[Karnathe]

Awesome, thank you. If that repair works, please let me know.

[Miek]

The part arrived. There were a few options on ebay and Aliexpress. I took a punt on one on Aliexpress that was supposed to be still mounted to a section of PCB from a salvaged board, as I figured that was the best chance of getting a genuine working part, but what turned up was a fresh-looking part in cut-tape.

I wasn't very confident in it being genuine but decided to try it anyway and, after a couple of tries, I got it soldered on and it seems to work!
I'm still a little suspicious of it being a clone part, but the phase noise doesn't seem to be wildly off - it's lower than I can measure on my spectrum analyzer at least.

In my first tests I still saw the error very rarely but then I did the reference oscillator adjustment, put it all back together, and left it sweeping overnight and have had no more errors since. So fingers crossed it's all good now.

However, I strongly recommend you do some more troubleshooting to narrow down the issue with yours before replacing this part. It wasn't a cheap part and it was a real pain to get the board heated up enough to solder it back on without doing more damage. There are a lot of other things that can go wrong in the PLL loop that will cause the same error.

I suggest a few things:

• Make sure the ref. oscillator is tuned correctly (or if you're feeding an external reference check that)
• Narrow down the symptom some more. When it warms up and throws the error, does it stay broken or will it lock again if you re-tune it? Is it more prominent at certain frequencies? Is it more prominent in sweep mode when it's constantly re-tuning? etc.
• When it's working, refer to the service manual block diagram and the sketch above and trace out the loop. Check that the VCO signal is at the right frequency, check that the signal still appears correctly after amps/doublers/switches, and check the outputs of the divider & phase-detector. Some stage may be low, right on the threshold of working, then fail as it warms up.