Figured it out.
?
So the LT1084 arrived from digikey. I swapped U18 and U19 on the A3; no changes in LED state or boot up error.
?
At this point, I had also removed PSU and removed PSU daughterboard, cleaned all the connectors and put more deoxit on the pins, and put it all back together.
So I'm sure it wasn't a power supply issue.
?
There are 5 right angle male SMB connectors on the front of the A3 module. Two on the left are CH1 IF and two on the right are CH2 IF. The middle one is a 100 MHz clock.
?
Looking at the layout of the A3, this is a clock input, as there are multiple clock distribution ICs which spread out to all the FPGAs and ADCs.
?
So I put a low frequency directional coupler in the path of this clock and checked the coupled signal on a scope. No signal at all.
?
I then used my signal generator to inject a 100 MHz, -20 dBm clock in manually. After that, the instrument got past the original error and booted to the measurement screen!
However, it failed nearly all of the power-on tests. Strange, since the results were mostly gibberish (no actual voltage values, lots of fails on power rails that I confirmed manually).
?
?
I decided to investigate the lack of 100 MHz clock. This SMB cable leads back to one of two identical local synthesize modules (A4/A5). Removed both as they are identical.?
?
?
Doing some side-by-side measurements, I noticed that there was a blown surface mount fuze on the module which was responsible for the clock.
?
I traced the circuit before and after, it looks like this fuze separates a switching and linear regulator locally from the power rail coming in through the backplane. There are lots of tantalum and aluminum caps, so I did some in circuit DC measurements. No sign of a blown or failed cap, and all test points measured identically between the two.
?
Of course, I don't have a 1 amp SMT fuze in this size, so I had to macgyver a 5 x 20, 1A glass fuze with heatshrink and some wire to put it in circuit.?
?
?
Popped the modules back in and viola, its working.
?
?
No clue why the original fuze opened, as the unit seems to run normally otherwise. All power on test points are normal and measurements also look reasonable.
Example measurement of recently OEM calibrated sig gen (E8257D UNX), seems to align with specs.
?
?
?
Now have to order the proper fuze and make the repair permanent.
?
Cheers,
Kevin
Kevin A.