A320 pitch rate - Autopilot ON
Thread Starter
Join Date: Jun 2006
Location: CASEY STATION
Posts: 191
Likes: 0
Received 0 Likes
on
0 Posts
A320 pitch rate - Autopilot ON
Can anyone direct me to information regarding the pitch rate that can be achieved with AP ON for the A320.
Specifically a good description as to why Altitude or Speed targets can be overshot under some circumstances ( usually pilot induced ).
For example using a combination of Speed Brakes / Open Desc / High Speed from high altitude can create very high descent rates / low pitch attitudes. When correcting this situation the FD command and actual pitch can be slow to align leading to an over speed. Similar for low altitude level off, if the aircraft is not handled correctly. The FG will capture but overshoot the intended level off altitude. Is this a pitch rate limit or G loading limit?
The only reference I can find in the FCOM is a statement that the ELAC and SEC computers limit what the AP can order.
Does anyone have any more detail mainly to settle a debate among colleagues!
Specifically a good description as to why Altitude or Speed targets can be overshot under some circumstances ( usually pilot induced ).
For example using a combination of Speed Brakes / Open Desc / High Speed from high altitude can create very high descent rates / low pitch attitudes. When correcting this situation the FD command and actual pitch can be slow to align leading to an over speed. Similar for low altitude level off, if the aircraft is not handled correctly. The FG will capture but overshoot the intended level off altitude. Is this a pitch rate limit or G loading limit?
The only reference I can find in the FCOM is a statement that the ELAC and SEC computers limit what the AP can order.
Does anyone have any more detail mainly to settle a debate among colleagues!
Join Date: Sep 2016
Location: USA
Posts: 803
Likes: 0
Received 0 Likes
on
0 Posts
I can't say anything Airbus-specific, but the designers of closed loop controllers (i.e., the parts of the AP brain that control the pitching to seek/hold a speed in speed mode, or an altitude in altitude mode, etc.) always have a balancing act to play: If they set it to be too sensitive, it'll correct with too much intensity and overshoot the target value, and even worse overshoot the correction to the overshoot, etc. and have oscillations. If they set it to not be sensitive enough, then it'll be too slow to act and not be able to stop a developing deviation in due time.
Finding a sweet spot of sensitivity between these 2 extremes can be elusive. Especially since the airplane has different behavior in different regimes. So it can be very difficult to make the AP interface with that behavior throughout the entire range. It'll be OK in one range, but be too slow or fast in another range. For example, the CRJ in speed mode in a climb, is very good at not chasing the speed in normal turbulence. It holds a nice steady attitude. But at high altitudes (I mean mid-20's, "high altitudes" right?
) you'll get these long-period, and very slight in intensity, up and downdrafts... even on a smooth day. But the rate at which they come, happens to match the rate at which the AP pitches for speed, in a way that results in the AP chasing the speed back and forth, slowly bouncing in vertical speed from +2000fpm to level flight. It's a resonance, like an extremely slower version of the opera singer shattering the crystal glass.
Finding a sweet spot of sensitivity between these 2 extremes can be elusive. Especially since the airplane has different behavior in different regimes. So it can be very difficult to make the AP interface with that behavior throughout the entire range. It'll be OK in one range, but be too slow or fast in another range. For example, the CRJ in speed mode in a climb, is very good at not chasing the speed in normal turbulence. It holds a nice steady attitude. But at high altitudes (I mean mid-20's, "high altitudes" right?
) you'll get these long-period, and very slight in intensity, up and downdrafts... even on a smooth day. But the rate at which they come, happens to match the rate at which the AP pitches for speed, in a way that results in the AP chasing the speed back and forth, slowly bouncing in vertical speed from +2000fpm to level flight. It's a resonance, like an extremely slower version of the opera singer shattering the crystal glass.
Last edited by Vessbot; 28th Feb 2018 at 03:58.
Join Date: Apr 2007
Location: The Woodlands
Age: 64
Posts: 175
Likes: 0
Received 0 Likes
on
0 Posts
Not sure if this will answer the question to design sensitivity although focusing particularly on the A320 family, the questions are... the ELACs and SECs merely actuator controllers for the control surfaces? or do they also carry the autopilot flight control algorithms? Or is the "autopilot" a separate entity, which feeds desired control surface positions to the ELACs and SECs or are these latter simply "slaves" to the autopilot?
Then if the autopilot is indeed a separate "master" entity which feeds desired control surface position signals to the ELACs and SECs, then where does the autopilot get its inputs from?
Short answer I believe, the autopilot is a feature of the Flight Management And Guidance System (FMGS), and in particular each of the two Flight Management And Guidance computers (FMGC) manage its autopilot sensitivity, more to the case of yaw and roll order. As to pitch order I would be curious if it is the same.
I do find it frustrating in an open descent or climb to have the A/C slow pitch oscillations to speed.
Then if the autopilot is indeed a separate "master" entity which feeds desired control surface position signals to the ELACs and SECs, then where does the autopilot get its inputs from?
Short answer I believe, the autopilot is a feature of the Flight Management And Guidance System (FMGS), and in particular each of the two Flight Management And Guidance computers (FMGC) manage its autopilot sensitivity, more to the case of yaw and roll order. As to pitch order I would be curious if it is the same.
I do find it frustrating in an open descent or climb to have the A/C slow pitch oscillations to speed.
Last edited by jimmyg; 1st Mar 2018 at 01:50.
Join Date: Sep 2011
Location: FL390
Posts: 238
Likes: 0
Received 0 Likes
on
0 Posts
I think it’s probably FMGC limited. If you disconnect the autopilot when it’s very slow to respond to a mode change (ie OP CLB to VS), the flight directors move immediately to show a more aggressive response if followed.
Join Date: Mar 2006
Location: USA
Posts: 2,515
Likes: 0
Received 0 Likes
on
0 Posts
No empirical information, but anecdotally, I've noticed that the bus seems to be programmed for pax comfort. For example, in NAV mode it seems to initiate turns en-route a bit earlier than other planes, presumably to minimise bank angle.
Likewise, I've seen it react pretty lazily to small (1000-2000ft) altitude changes in OP CLB/DES mode.
Likewise, I've seen it react pretty lazily to small (1000-2000ft) altitude changes in OP CLB/DES mode.
Join Date: Jun 2007
Location: Wanderlust
Posts: 3,406
Likes: 0
Received 0 Likes
on
0 Posts
Likewise, I've seen it react pretty lazily to small (1000-2000ft) altitude changes in OP CLB/DES mode.
Note: If the change is less than 1 200 ft in OPEN CLB mode, the aircraft responds with a rate of climb of 1 000 ft/min.
Join Date: Apr 2004
Location: United States of Europe
Age: 40
Posts: 502
Likes: 0
Received 0 Likes
on
0 Posts
This is fine for the enroute phase, but could create ATC trouble on terminal procedures. Mitigation would be to put a PD waypoint at -5.
Join Date: Mar 2006
Location: USA
Posts: 2,515
Likes: 0
Received 0 Likes
on
0 Posts
Enroute turn anticipation criteria applies above FL200 (or whenever the predicted altitude at the waypoint is above FL200). This will schedule for min 5 degrees bank but a turn no earlier than 20 nm or so, if my memory serves me right.
This is fine for the enroute phase, but could create ATC trouble on terminal procedures. Mitigation would be to put a PD waypoint at -5.
This is fine for the enroute phase, but could create ATC trouble on terminal procedures. Mitigation would be to put a PD waypoint at -5.
20nm seems a bit excessive for an upper limit for a subsonic aircraft 🙃
Join Date: Apr 2004
Location: United States of Europe
Age: 40
Posts: 502
Likes: 0
Received 0 Likes
on
0 Posts
Valid for fixed leg flyby transitions. I.e. normal turns with no overfly:
Terminal Area (below FL200):
Enroute (FL200 and above):
The enroute vs terminal logic is based on the predicted altitude at the turn bi-sector: i.e. predicted altitude at the waypoint. That's why you could see a very shallow bank angle at much lower altitude when the predicted altitude at the to-waypoint is above FL200. Normally this would be more noticeable with larger course changes, i.e. more acute angles. This is undesirable when still in the Terminal area.
When operating in this condition, you could avoid this phenomenon by inserting a Place-Distance waypoint before the TO-waypoint. e.g. 'ZZZ/-5'.
The FMGS does not start turn-anticipation construction till the previous waypoint is sequenced. i.e. can only think 1 waypoint ahead.
Hope this helps.
Terminal Area (below FL200):
- Philosophy is a short turn sticking to the path
- Turn-anticipation not more than 10NM
- Roll-angle (phi) is equivalent to the Course Change required
- Minimum 10 deg bank
- Max 25 for pre-computation, Max 30 once in the turn.
Enroute (FL200 and above):
- Philosophy is pax comfort
- Turn-anticipation not more than 20NM
- Roll-angle (phi) is the one required for a turn tangent to the 5NM nominal airway limit.
- Minimum 5 deg bank
- Max 25 for pre-computation, Max 30 once in the turn.
The enroute vs terminal logic is based on the predicted altitude at the turn bi-sector: i.e. predicted altitude at the waypoint. That's why you could see a very shallow bank angle at much lower altitude when the predicted altitude at the to-waypoint is above FL200. Normally this would be more noticeable with larger course changes, i.e. more acute angles. This is undesirable when still in the Terminal area.
When operating in this condition, you could avoid this phenomenon by inserting a Place-Distance waypoint before the TO-waypoint. e.g. 'ZZZ/-5'.
The FMGS does not start turn-anticipation construction till the previous waypoint is sequenced. i.e. can only think 1 waypoint ahead.
Hope this helps.
