HP 8563E W16 and A10 schematic?

[Oldtest2]

Would anybody have a schematic or signal  pin-out of the W16 cable and the A10 RYTHM module in a 8563E spectrum analyzer? I have one here I would like to fix. It powers on with no errors, CRT appears ok. The response in the low band range is very low, 70 dB or more, in the microwave bands, 30-40 dB low, slightly less as the freq is increased towards 10 gHz.   The YTO freq and power appears to be ok. There is a lot of measured loss through the A10 module in low band mode.  Was able to find a service manual, but no component level schematic. I wonder if the A10 is removed if its possible to verify its bad.

[kipp]

Hi,

The A10 "RYTHM" assembly is a combination diode based RF switch, YIG filter, and YIG-based mixer (a YIG sphere near saturation provides the non-linearity for the mixer).  The 8563's documentation treats it as a black box, but others have tracked down what looks like a good description of its guts in European patent 0 280 389 A2 (application # 88300393.1).  I don't know the copyright status of the document, so I don't want to post it here, but the filename you can find online is EP0280389A2.pdf.

I'd focus on the low-band issues first because that path is simple and mostly fool-proof.  What the schematic describes as the low-band switch input to the A10 module is the bias for the diode switch.  When the diode is conducting, the low-band output port is a short to ground (at least, frequency components above about 2.9 GHz think it s), and RF energy is coupled into the first of the YIG spheres, and from there into the high-band path.  But when the diode is *not* conducting, looks like an open circuit, incident RF energy passes directly through from the input port to the low-band output port.  The A10 module is basically a straight-through connection at low frequencies (well, I'm sure there are capacitors in there, too, to prevent the diode bias voltage from appearing on any of the ports).  That's why the low-band path should be reliable:  the low-band signal path is what the A10 module does when it's "off".  Pretty much the only way the low band signal path can fail is if the diode has become a short circuit and (partially?) conducts all the time, regardless of the bias voltage across it.  That's possible, but I feel like that's an unlikely scenario.

The diode is driven by a totem pole circuit:  a pair of transistors switch its bias voltage between +ve and -ve 15 V with respect to ground.  -15 V = low band (biased off), +15 V = high band (biased on).  The output of those transistors goes through R155 before going onto the cable to the A10 module.  You can use the voltage drop across that resistor as an ammeter to check the health of the diode.  In my unit, I see 0 V drop across R155 in the low band (diode is off), and 1.8 V across that resistor in the high band (diode is conducting about 35 mA).  My unit's high-band signal path is not working so I don't know if that voltage is what you would see in a working unit, I think it is, I think my problem is something else, but if you see 0 V across that resistor in the low band *and* -15 V to ground on both sides of that resistor then that diode is not conducting, and I'm pretty sure your low-band path through your A10 is working correctly.

There are typos in the service manual in the section related to the A10 drive circuitry.  The U416 chip has 4 sections in it labelled A, B, C, D, and I believe whoever typed up the diagnostic procedure in the manual mistook "B" for "D" or vice versa and swapped them in the table listing the voltages that should be on that part's pins.  Some of the voltages for pin 6 on the J15 diagnostic connector differ from what I measure in mine, but my unit would match the table if you believe, similarly, that some 8's got turned into 3's when being typed up for the manual.  That sort of thing.

Another thing, though, that has nothing to do with the A10 module, but that seems to be a common age-related problem in these instruments, *and* that causes the instrument to seem to have no sensitivity to input signals anywhere across the entire spectrum, possibly in combination with a variety of nonsensical error messages (but not always) is the loss of gain in the amplifiers in the circuitry used to synthesize the internal 100 MHz and 300 MHz references from the 10 MHz reference.  The amps just get old, and work ever more poorly, causing the 100 MHz and 300 MHz tones to fade in amplitude until stuff stops working right.  It happens slowly, and at first the instrument just gets annoying to use, then looses its ability to see input signals, then eventually it stops working altogether and spews dozens of error messages about ADC failures when you turn it on (why?  I have no idea).  The amplitude of the 100 MHz cal. output signal is *not* affected by this failure, because it has its own levelling circuit that fixes its amplitude, so checking doesn't rule out this problem.  The main culprits reported by others seem to be the MSA-0505 amplifiers, but in mine I ended up replacing every amp in that block of circuitry:  1x MSA-0505, 1x MSA-0386, and 2x MSA-0486.  Between Amazon and Aliexpress, I got a 2-pack, 7-pack, and 10-pack, respectively, of those parts for a total of about $30, delivered, so it's a remarkably economical repair as these things go.  My unit was instantly cured of (nearly) everything that ailed it.