I've performed some of the microwave troubleshooting described in the HP
8971C service manual. The microwave circuitry in the 8971C test set
consists of a number of paths selected with relays. These paths consist
of a simple pass-through for low frequencies directly handled by the
8970B and filters and a downconversion mixer for higher frequencies (to
translate RF frequencies to an input frequency the 8970B can
handle). I've attached an image from the service manual showing these
paths.
The first part of the troubleshooting section involves injecting an RF
signal, selecting a path, measuring the signal gain, and comparing it
with the gain computed from the individual component gains. I've also
attached an image showing these individual component gains.
The gains I measure differ a fair amount from the predicted gains.
In SSB (single sideband) 1 I'm measuring a gain of approximately 23 dB
instead of the expected 15 dB. Moreover, this gain is pretty flat across
the bandwidth as tested by moving the RF signal in 100 MHz steps.
In SSB 2, I also measure a gain in excess of the predicted value. In
this case, 24 dB instead of 19 dB expected.
In SSB 3, I measure a gain of about 16 dB instead of the 20.5 dB
expected. If I move the RF signal by 10-20 MHz (without adjusting the LO
frequency) the gain moves up more toward the expected value.
Any thoughts on what might be going wrong here? SSBs 1 and 2 suggest to
me that maybe the attenuator (AT3) is passing the signal through
un-attenuated. However, the test for SSB 3 indicates that if this is
true, then the YIG filter + attenuator combo is attenuating the signal
by a margin even more than that predicted by the third test on it's own,
which is concerning. Unfortunately, I've only been able to test SSB 3
operation at the low end of its range (2.5 GHz). I'm using my 8340 as
the LO and the only other RF signal generator I have is an 8663A, which
tops out at 2560 MHz. I've got an 8341 too, but it needs repairs.
An obvious solution to these problems would be to open up the instrument
and test components individually. Unfortunately, I can't do that yet as
I bought this unit in a supposedly functional state and may lose my
right to return it if I open it up.
All thoughts appreciated.
Thanks
Matt
