How easy is to repair LO_UNLOCK on an Agilent ESA-E series spectrum analyzer ?

[electronic_guy]

Hi,

I've seen the signal path video where the LO_UNLOCK is repaired. But this error can be caused due to a lot of reasons as I've heard. What do you guys think the chances of repairing this error on an ESA E series spectrum analyzer ? Note that I don't have a 1Ghz mixed domain oscilloscope as shown in the signal path video. But I do have a CMU200 and 100MHz oscilloscope. I don't want to waste my money on something that requires unobtanium and that would basically turn it to a boat anchor. I was wondering if I could try the final results of those repairs on my device with this error and may be have a better chance at repairing it. But I don't know how big is the gamble I'm thinking about. So seeking your advice on this.

Thanks in advance.

[DaJMasta]

It's going to be probability.  Unless you've got schematics and can verify you have a source for every part on there, it's a roll of the dice as to whether you can find it and whether it's available.  You can stack those odds in your favor with research and experience, but you're not going to be able to fix every last thing you get a hold of, and if your sample size is youtube videos - you're probably not seeing every piece of equipment that comes in, so the odds may seem better than they are.  That and Shahriar is about as expert as someone can get.

Look at whatever documentation is available and try to thing about where the fault could be, at least at a block-diagram level.  Try to guesstimate how well you are at probing and troubleshooting a board with the information you have.  Maybe deduct some points for custom ASICs/hard to find modules/stacked up boards that make probing particularly difficult in circuit, and then come up with a price where you think you can come out on top if you make the attempt, knowing well that it may just be a learning experience/parts unit.

It's going to be a gamble, and much of that equation comes down to your own skill in troubleshooting and actually making the repair.  Having a good understanding goes a long way, but you'll find stupid, hidden failures that could brick the whole project unless you happen to spot it.  Repair is a much surer thing with component level documentation, but with complex circuitry, it takes a huge amount of understanding to be able to spot any given fault even with full schematics, and we usually don't get that on modern equipment.  When you know what to look for, the work can be concise, when you don't, you'll spend hours of probing and troubleshooting and won't have much better of an idea than you started with.

In terms of equipment, I would want to be able to see the output of the LO on a spectrum plot as well as its IF or RF output for given mixing stages (I don't know what they are in the model you're considering, but I'd want enough bandwidth to cover all but the highest one, minimum).  I'd also want a synthesizer that can generate tones to put into the input to see what is displayed (for verification at the end as well as to learn more about the fault), and ideally, I'd like that synthesizer to be able to output the LO frequency, to be able to check if injecting the LO makes the thing come to life.  You can sometimes do it with less (looking at the current draw of an amplifier stage can tell you if its level is fluctuating without having to look at the full signal bandwidth), but I probably wouldn't try it unless I had to.

[coppercone2]

It's Christmas, electronic_guy. It's the time of miracles. So be of good cheer... and call me when you hit the last lock.

[jwrodgers]

I've repaired two of these with the LO Unlock error, and both turned out to be the divider chip in the RF section. You can take the bottom off and power it on, and you will see the same LED's as Sharir had if it is the same error. Anyway let me know how you get on, I have a spare chip if you want, and can give you some pointers in how to unsolder and resolder the new one as there are a few tricks to it. - I would only recommend you tackle that chip replacement if you are familiar with SMD work and have some skills. I am not great, but managed without wrecking the unit and got it working.

The chip is U60 here, you can see its an Agilent. I replaced it with HMC365S8GE IC FREQ DIVIDER DC-13GHZ 8SMD. Need to check the schematic in the CLIP, but from memory one of the leads was not connected. Again Sharir mentions this in his video too.

The trick I used to unsolder the offending item is to cut all the pins with it in place with a very sharp stanley blade, being careful not to damage the PCB pads or tracks! You can then heat up the large heatsink pad below - take your time, there is a lot of sinking capability thought the ground, and eventually the chip body will come away. You can then remove the chip legs from each of the pads and clean up around the pads and large heatsink pad.


Let me know if you get access to the RF section and check the LED lights in there.