Agilent E4402B ESA - LO Unlock - Looking for advice

[millerjs]

Hello all!

My father (a ham, like me) was recently given an Agilent E4402B by a fellow radio enthusiast. We're not sure where this guy got it, but he likely never turned it on - he's apparently a receiver guy and had no interest in a spectrum analyzer. My father and I, on the other hand, were incredibly excited to have such a piece of gear. All we knew about it is that "it might not work". Despite "only" being the 3GHz model, it seems to be well equipped - RF out, the 1D5 high-stability frequency reference, and a couple of cards I don't recognize. It has 16MB RAM and 32MB flash, too. Surprisingly few hours, too, 175 - at least 15-20 of us were from us fiddling around with it. It's pretty grody though, lots of what looks like mold under the display plexiglass. It has a sticker on the back suggesting the battery was replaced in Nov 2013, and it retains time & date, so that's still working. It even still has the original yellow floppy drive protector!

Upon powering on, it seems to pass all the software/digital self tests, but it fails to do an Align All because of the infamous LO Unlock error. It fails to display any signal at all, but we do know that the 10Mhz out on the back is good. I've done some research on this error, and I know that it could represent a lot of things, so I'm not really sure where to start. I have repaired and calibrated older analog scopes and frequency generators before, but nothing anywhere near this advanced. I watched Hugoneus' amazing 2 hour troubleshooting and repair video, and while I learned a lot, and could definitely do the repair he did, I unfortunately do not have the equipment needed to do the troubleshooting that he did (namely, a >3GHz scope, frequency counter, signal generator, etc.)

Additionally, I have read that if these ESA's get >0V put to their RF input, that'll completely blow out the front end, but I don't know if that's something that can be repaired - and I don't really know where to start as far as troubleshooting this. Like I said, we're both in over our heads here. While we probably could part it out or sell it as-is, we would like to at least take an honest stab at fixing it, if we can do so without making things significantly worse.

I guess what I came here looking for is some kind advice? He lives in the Raleigh, NC area, so right now my hope is to hook him up with a Makerspace, or someone with gear sufficient to begin properly troubleshooting and testing this thing, and see if we can get it ticking again. He's a mostly-homebrew, all QRP, radio operator, and a working spectrum analyzer would be something he could really benefit from.

I've uploaded a few pictures here if you want to check it out: https://photos.millerjs.org/?f=Agilent+E4402B

Thanks all,
John KC1QLN & Kevin KI4DEF

[bob91343]

It may need some sort of alignment.  Unlock can be a failed component or connection, or maladjustment of a PLL.  Do you have the manuals?

[Uunoctium]

I'm not very familiar with that instrument, but in general it is recommended to use "Preset" (green button) first to avoid that previous altered menues at automatic "recall last" at power up sequence are loaded. Saves a lot of time in searching why devices behavior is other than expexted.

Any entrys in Error log? (System ->Show Errors)

One of the screenshots shows something like "Rf skipped..." 50MHz detected. That seems to me wrong. 50Mcy are at the CAL out (BNC conn).
The yellow hint at bottom reads "connect Input to RF Out = TG Out! (left N conn).

Edit: if needed: PW = -2010

[Uunoctium]

In
System->more->more->service (Type in PW + enter)
under "Align" you'll find an item "LO", but i don't know, whats behind, cause i didn't activate (fear of).
Maybe that directs to the solution of your LO fault

[millerjs]

It may need some sort of alignment.  Unlock can be a failed component or connection, or maladjustment of a PLL.  Do you have the manuals?

We do have the manuals (PDFs, anyway) but haven't made much headway for this specific issue. Like I said, we don't have a whole lot of other gear that we can use to troubleshoot this yet.

I'm not very familiar with that instrument, but in general it is recommended to use "Preset" (green button) first to avoid that previous altered menues at automatic "recall last" at power up sequence are loaded. Saves a lot of time in searching why devices behavior is other than expexted.

Any entrys in Error log? (System ->Show Errors)

One of the screenshots shows something like "Rf skipped..." 50MHz detected. That seems to me wrong. 50Mcy are at the CAL out (BNC conn).
The yellow hint at bottom reads "connect Input to RF Out = TG Out! (left N conn).

Edit: if needed: PW = -2010

Thanks - we have tried some of this (including reseting the unit to factory defaults. I have tried getting into the service menu previously but -2010 didn't work - I'll try with it again. Unfortunately the error pages are completely empty. I think that error ("RF skipped 50MHz detected") is because of how we were poking around with the connectors at the front while it was attempting to align - it does not normally show.

[Digisol]

https://youtu.be/3R0txBEIvjQ
This guy has repaired a few with Lo lock error, both had the same chip replaced to fix it. If this helps please let me know. The name on YouTube is "The Signal Path" and I enjoy his post.

[millerjs]

Yeah, I think this is what I'm going to do next. I'd really like be able to actually test to know if the divide-by-four is broken or not, but at $20 a part I think I'm just going to order one and cross my fingers.

https://www.digikey.com/en/products/detail/analog-devices-inc/HMC365S8GE/5359926

[veedub565]

I've had that problem on an E4411B, reseating the PSU seemed to work at least temporarily. Also, in one of the firmware update revisions it mentions fixing an LO Unlock bug. I can't remember which one it was but have a read through the changelogs.

I updated the firmware on mine and the problem hasn't re-appeared.

[Uunoctium]

I've seen the two videos. How to remove and resolder that chip remains his secret, hmm...
Only one hint: not only pin 8 is GND - the whole backplane (Heat spreader) MUST be soldered! If not, you grill the chip, cause pin 8 is only a bond wire to die and not capable taking the current

[millerjs]

I've had that problem on an E4411B, reseating the PSU seemed to work at least temporarily. Also, in one of the firmware update revisions it mentions fixing an LO Unlock bug. I can't remember which one it was but have a read through the changelogs.

I updated the firmware on mine and the problem hasn't re-appeared.

I haven't re-seated the PSU yet but that's definitely something I can try. Something else I wan to try is to remove as many of the option cards as I can to isolate out all other possible sources of trouble. I haven't made the firmware floppies yet (I need to get the unit home where my USB floppy drive is, and then also FIND the firmware images!) but it's good to know that may resolve the issue, too.

I've seen the two videos. How to remove and resolder that chip remains his secret, hmm...
Only one hint: not only pin 8 is GND - the whole backplane (Heat spreader) MUST be soldered! If not, you grill the chip, cause pin 8 is only a bond wire to die and not capable taking the current

Thanks for the helpful tips. I am anticipating a challenge in removing this chip but I think I'll be able to manage, I have access to a hot-air rework station. I have also been informed to remove the N/C marked pins and not solder them.

[paul@yahrprobert.com]

There's a couple people in this forum who have been doing a lot of excellent work repairing the ESA series (E440x).  Search for smgvbest and analogRF.  Also, a great resource is on ebay you can find Artek manuals.  For about $25 you can get the CLIP (component level information package??) which has some valuable circuit diagrams.

[veedub565]

I've had that problem on an E4411B, reseating the PSU seemed to work at least temporarily. Also, in one of the firmware update revisions it mentions fixing an LO Unlock bug. I can't remember which one it was but have a read through the changelogs.

I updated the firmware on mine and the problem hasn't re-appeared.

I haven't re-seated the PSU yet but that's definitely something I can try. Something else I wan to try is to remove as many of the option cards as I can to isolate out all other possible sources of trouble. I haven't made the firmware floppies yet (I need to get the unit home where my USB floppy drive is, and then also FIND the firmware images!) but it's good to know that may resolve the issue, too.

You can download the firmware from here https://www.keysight.com/gb/en/lib/software-detail/instrument-firmware-software/esa-spectrum-analyzer-firmware-1000001085epsgsud.html
If you look at the revision history one of the early fixes was:

"Defects Repaired for ESA:
• Repair LO reclocking defect in 1.5 GHz instruments (E4401B, E4411B, and
future release of E7401A). This manifested itself as "Align Now All, required"
with noise floor near or at Top Of Screen. "

Which is the problem I was having. I'm not sure where it says "reclocking" if thats a typo and it should be unlocking. I've not had a problem with it since anyway.

[jwrodgers]

https://youtu.be/3R0txBEIvjQ
This guy has repaired a few with Lo lock error, both had the same chip replaced to fix it. If this helps please let me know. The name on YouTube is "The Signal Path" and I enjoy his post.

I used Sharair's videos to debug and fix and LO Error on my E4402B. I have 2 of these units, and both have the same LO Error. Fixed the first one, onto the second now. Unsoldering the wee beggar is the main issue, as it is soldered with a large pad underneath to a very large groundplane/heatsink! I first 'knipped' the leads with decent side cutters, which flattened the leads significantly, then carefully cut them with a very sharp stanely knife. With the leads disconnected from the body, I was able to get enough heat into the pad below the wee bugger to get it off, and then cleaned the pads for the other leads.