Gregory's ServerWith the hemisphere shells removed from the ball, you can see the internal mechanism. The pitch motor rotates the assembly around the pitch axis. The hemispheres are rotated by the azimuth (yaw) motor (not visible). 3/N
|
24 comments
 The attitude indicator uses 1960s-era electronics, with each angle transmitted over three wires from a synchro. The indicator uses a servo loop for each axis, using a feedback loop to rotate the motor until the shaft angle matches the input angle. A "control transformer" compares its shaft angle to the electrical input and generates an error signal. The amplifier turns the motor as needed to eliminate the error. The three amplifiers are in a unit that attaches to the back of the attitude indicator box. Inside are three amplifier boards, a small power supply board, and an AC transformer. Here's a closeup of one of the amplifier boards. It's a bit of a mess with components stacked on other components to save room. The spider-like component in the middle is a pulse transformer. It drives the transistors on the right, one for each motor direction.  And an admirable sparsity of 10-turn pots! Says person formerly responsible for a thing called a "neutron induction computer," a design gone bad as it was a 2U rack unit containing -93- 10-turn potentiometers. Nightmare. This photo shows the attitude indicator in the F-4 cockpit. It is in the center of the control panel, below the purple radar screen. There are two sources for attitude information, selected by the primary/standby switch in the upper left. The mounting of the switch is questionable: screwed down to the console with visible wiring. While researching the use of the attitude indicator in the F-4, I came across the "nuclear store consent switch". If the plane has a nuclear bomb attached, you flip this switch from the SAFE position to the upper position (REL/ARM) to arm the bomb for release. Somehow, I was expecting something more elaborate. For more on the F-4's attitude indicator (including reverse-engineered schematic), see my blog post at https://www.righto.com/2024/09/f4-attitude-indicator.html Credits: I worked on this with @tubetime and CuriousMarc. Thanks to the collectors who provided attitude indicators and amplifiers. Aircraft photos from National Museum of the USAF: https://www.nationalmuseum.af.mil/Visit/Museum-Exhibits/Fact-Sheets/Display/Article/196051/mcdonnell-douglas-f-4c-phantom-ii/  @kenshirriff I had trouble wrapping my head around the mechanism. There's a diagram about halfway down the Wikipedia page that helps: @kenshirriff I could have used this years ago. I was in responsible for testing the standby flight instruments for the B-2, and the "turnball" (not sure anyone else calls it that) would work once, then fail when QA came to witness. I eventually found out why from the spec (which the USAF made hard to find), and rewrote that test. As usual, a great view of interesting tech.  @kenshirriff @tubetime @yourfutureex @tubetime That's the reason that the attitude indicator is vitally important: in an aircraft, gravity can feel like a completely different direction from down. For example, see the famous video of pouring iced tea in a plane while doing a barrel roll.   @kenshirriff https://en.wikipedia.org/wiki/Permissive_action_link sound like a fine idea until you find out they were all set to 00000000 for a time. @kenshirriff @0tracas  @kenshirriff I've wanted to know how these things work for ages, I'll check this out when I have time!  @FritzAdalis The reddish-purple circle is the radar scope for the AN/APQ-120 radar system.  @kenshirriff wow, that's some workmanship. Late change in requirements leading to iffy solution? @tehstu I've seen the switch in other aircraft photos and in the manual, so it's not a one-off hack. I think the attitude indicator originally took a single input source but later planes had two inputs.  |
With all these rotating assemblies in the attitude indicator, the wires would get tangled up. To avoid this, the indicator uses two sets of slip rings: one on the roll axis and one on the pitch axis. 4/N