Narrative:

During the final portion of flight; the approach controller issued a final vector to intercept the localizer to [land] and instructed us to maintain 2000 ft until established. The first officer was the pilot flying. I was the monitoring pilot in command (pm). The altitude pre-selector window was showing 2000 ft and I believe we were already leveled at that altitude. We captured the localizer and a few seconds later I noticed we were at 1800ft descending in fpa mode; below the glideslope; still outside the final approach fix. I told the first officer that we were to maintain 2000 until established; as I was speaking the first officer re-selected 2000 ft but the airplane remained descending in fpa; almost immediately the first officer disengaged the autopilot and manually raised the nose to go back to our assigned altitude of 2000 ft. The altitude alert also sounded as the first officer was correcting the deviation. The lowest altitude that I saw on the tape was 1700 ft. While the autopilot was off I selected flch to restore proper flight director guidance; soon afterward the first officer requested the autopilot back on; and I engaged it. The airplane leveled off at 2000 ft and shortly after captured the glideslope. The remainder of the approach was uneventful. At the gate the first officer and I discussed what had happened and he told me that he wanted to dial back the speed; but mistakenly twisted the altitude selector knob instead of the speed selector knob. I am not sure how we ended up in fpa mode since; like I said before; I believe we were leveled at 2000ft in altitude mode. But after discussing the situation with the first officer; I'm second guessing myself and I think it is possible we were in leveling at 2000 ft in asel mode; when he mistakenly changed the altitude. This would have resulted in fpa becoming the active vertical mode.I identified the following contributing factors to this scenario:-IMC conditions and light rain at 2000 ft. Convective activity in the area.-report time was [morning] and PIC (myself) had been flying all [evening] shifts; several past midnight; in the two weeks prior; and had not had a report time for a flight before [morning] in three weeks. I believe the above factors played a role in slowing down my scan; which resulted in a deviation of more than 100 ft.[suggestion] there should not be pairings that result in a sudden shift from [evening] operations to [morning] operations and vice versa; since it is very difficult to readjust while in the middle of a trip or a reserve stretch. Put more emphasis on CRM and crew communications during training events. Had the first officer expressed his intentions of slowing down I would have payed immediate attention to his actions; and could have caught the error instantly.

Related reports:

Google
 

Original NASA ASRS Text

Title: EMB-175 Captain reported the FO mistakenly twisted the altitude selector knob instead of the speed selector knob causing the aircraft to drop below assigned altitude.

Narrative: During the final portion of flight; the Approach Controller issued a final vector to intercept the localizer to [land] and instructed us to maintain 2000 ft until established. The First Officer was the pilot flying. I was the Monitoring Pilot in Command (PM). The altitude pre-selector window was showing 2000 ft and I believe we were already leveled at that altitude. We captured the localizer and a few seconds later I noticed we were at 1800ft descending in FPA mode; below the glideslope; still outside the final approach fix. I told the First Officer that we were to maintain 2000 until established; as I was speaking the First Officer re-selected 2000 ft but the airplane remained descending in FPA; almost immediately the First Officer disengaged the autopilot and manually raised the nose to go back to our assigned altitude of 2000 ft. The altitude alert also sounded as the First Officer was correcting the deviation. The lowest altitude that I saw on the tape was 1700 ft. While the autopilot was off I selected FLCH to restore proper flight director guidance; soon afterward the First Officer requested the autopilot back on; and I engaged it. The airplane leveled off at 2000 ft and shortly after captured the glideslope. The remainder of the approach was uneventful. At the gate the First Officer and I discussed what had happened and he told me that he wanted to dial back the speed; but mistakenly twisted the altitude selector knob instead of the speed selector knob. I am not sure how we ended up in FPA mode since; like I said before; I believe we were leveled at 2000ft in ALT mode. But after discussing the situation with the First Officer; I'm second guessing myself and I think it is possible we were in leveling at 2000 ft in ASEL mode; when he mistakenly changed the altitude. This would have resulted in FPA becoming the active vertical mode.I identified the following contributing factors to this scenario:-IMC conditions and light rain at 2000 ft. Convective activity in the area.-Report time was [morning] and PIC (myself) had been flying all [evening] shifts; several past midnight; in the two weeks prior; and had not had a report time for a flight before [morning] in three weeks. I believe the above factors played a role in slowing down my scan; which resulted in a deviation of more than 100 ft.[Suggestion] There should not be pairings that result in a sudden shift from [evening] operations to [morning] operations and vice versa; since it is very difficult to readjust while in the middle of a trip or a reserve stretch. Put more emphasis on CRM and crew communications during training events. Had the First Officer expressed his intentions of slowing down I would have payed immediate attention to his actions; and could have caught the error instantly.

Data retrieved from NASA's ASRS site and automatically converted to unabbreviated mixed upper/lowercase text. This report is for informational purposes with no guarantee of accuracy. See NASA's ASRS site for official report.