开云体育

I think the IF filter architecture of the 70902A may be roughly the same as in the 8566. If so, those crystal filters are all 3.0 MHz center frequency, with the bandwidth set by controlling the load impedance, hence Q, by virtue of the RF resistance of PIN diodes. It should be a cascade of similar stages, all with varying PIN currents according to the selected IFBW. So, all of the sections should be active, but the BW is controlled by logical steps supplying PIN currents set by precision series resistors. The higher the PIN currents, the narrower the BW. There should be other PINs to route the IF around the crystal section at the wider bandwidths, and also some to provide gain compensation to match the gain for all the BW settings.

In the 8566, the 21.4 MHz IF is mixed down to 3 MHz, processed through the filters, then converted back up and sent on its way. I don't recall if the LC filters also process at 3 MHz, or at 21.4 MHz - either way is possible.

Those indicator lights may be the key to solving the problem, but not for the reason you suggested. It's possible that those are overload or other status indicators - not necessarily logical on/off. Maybe some of the logic states or levels are wrong, causing some filter BW settings to do the wrong thing. You would have to find the individual BW selection logic lines and see what they're doing. They may be high voltage logic - directly driving the PIN current setting resistors with up to +/- 20 V or so. Just remember, when the crystal section is needed, all the stages should be on, while the logically-controlled PIN currents set the BW.

Since these PIN signal routing and "tuning" lines actually can be considered to be analog control, a loose line out of control can allow for various wrong currents to flow between stages, causing interesting results - that is what I think may have happened. A wrongly tuned or stuck filter stage, or a partially turned on switch that should be on or off, could make quite a mess.

You may want to look at the down/up conversion process too, to make sure it's actually on the right frequency - there should be a crystal oscillator nearby running at 18.4 or 24.4 MHz. The filter crystals should say 3 MHz, or whatever they run at. If they are 21.4 MHz, then no down/up conversion is needed. If you can ID these you'll have a pretty good idea of what's going on.

It sounds like you don't have a schematic or manual, but maybe similar systems like the 8566 would at least provide some info you can use, presuming it's actually similar. I don't recall there being indicators on the filters, but I've never needed to look closely at the IF/display section yet.

The breakpoints around 10 kHz or 3 kHz make sense - the crystal filters are only used at the narrowest bandwidths, while the LC ones do the upper, and may also be in-line when the crystal section is used. I believe the LC filters are fixed BW, set only by the L and C adjustments - not by PIN loading - although there would be associated PIN switches for signal routing.

Ed