EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: bistromath on April 11, 2018, 06:17:22 am
-
Hey everyone,
I might need an expert for this one. Bear with me, there's a lot to go through here.
I bought an HP 8593E a few years ago for cheap money, and overall it seems to be working pretty well. There's only one problem, really: the close-in noise is bad. Awful, really. Pictures are worth a thousand words here:
(https://i.imgur.com/uT9QuGX.jpg)
So other than that it seems fine. ;)
Digging into it, I went ahead and followed the assembly-level service guide and calibrated it, both the self-cals and bandpass filter adjustments. Freq and amplitude calibrations went A-OK, no problems. I verified that both the comb generator and cal outputs are clean and have correct amplitudes. I verified that the 10MHz reference oscillator was clean, but the 321.4MHz first IF and 21.4MHz final IF both showed the problem. This pointed me toward the first LO, so I removed the A3A10 directional coupler's output and sampled the 4.223GHz LO and lo and behold (see what I did there), it was nasty close-in. But why?
At this point I checked all the power supplies. The 12V rail has some fuzz on it, but 12V isn't used for analog circuitry so it's not a big deal. +15/-15 look great. So.
Following up, I went back and noted that the FM coil diagnostic test had shown a flat line where the assembly-level service guide says there should be a nice ramp:
(https://i.imgur.com/YIZ1Csz.jpg)
In addition, the occasional "FREQ UNCAL" appears during this test (but never otherwise). At this point, I figured I had something to dig into, and bought a set of schematics (Thanks Artek!). Looking into what we're actually measuring, the FM coil diagnostic is muxing in the actual FM coil voltage -- not the output of the coil driver amplifier, but the voltage on the coil itself. The thing is, the FM coil is only around 1 ohm, so the voltage it sees is around 50mV or so. So it should in fact look like a flat line on the diagnostic screen. What gives?
I put a scope on the FM coil amplifier output (before the resistors that make it approximate a current source) and it shows a perfectly fine ramp. In fact, so does the voltage at the FM coil itself -- it's just that it's only around 50mV. I'm chalking this one up to a service manual that doesn't actually reflect the device under test, although I'd be happy to be told differently.
Oh, I also followed the service guide's recommendation to try unlocking the LO by removing the counterlock board connections. No change in the output (besides going all drifty on me).
So, leaving out about twenty hours of dead ends and false leads, this is where I stand now: the LO is noisy, the FM coil amplifier seems to be fine, and I don't know where to look next. Do I just have a bum YIG? Help!
Thanks for reading this far,
Nick