Centaurus
8th Sep 2017, 14:00
January 1982; N62AF, 737-200 Air Florida; Washington, DC
The crew used reverse thrust to assist the push-back off stand and taxied close behind another aircraft for heat from the exhaust which may have contributed to the airframe show/ice accretion. They also did not use engine anti-ice while taxiing out during snow causing the PT2 (EPR) probes to misread. Although anomalous engine readings were called during the take-off, the captain neither aborted nor adjusted the thrust levers to allow sufficient thrust for take-off and the aircraft stalled and crashed into the frozen Potomac River.
Take-off was almost an hour after de-icing had been completed and the CVR shows that the crew were aware of 10-20mm of snow on the wings. Four of the five crew members and 70 of the 74 passengers were killed.
See description of the accident.
https://en.wikipedia.org/wiki/Air_Florida_Flight_90.
See description below of erroneous EPR indications that caused the crash:
https://books.google.com.au/books?id=oMkMBAAAQBAJ&pg=PA86&lpg=PA86&dq=erroneous+EPR+readings+due+blocked+PT2+probes&source=bl&ots=qUfgr-D17-&sig=TkQulRYYs5ghAirK7aupBL-Qp8c&hl=en&sa=X&ved=0ahUKEwjCtsSD2pXWAhWFH5QKHQH9C10Q6AEIKDAA#v=onepage&q=erroneous%20EPR%20readings%20due%20blocked%20PT2%20probes&f=false
:
Centaurus comment.
Following that accident Boeing issued a memorandum to all operators of the importance of checking the N1 during the take-off as the NTSB investigation showed that blocked PT2 probes could cause the EPR gauges to over-read. For example, if the take-off performance charts required a take-off setting of (say) 2.18 EPR with accompanying N1 of 101%, a blocked PT2 probe meant that an N1 of 101% would likely result in 2.25 EPR. A pilot would think that an engine was over-boosting and pull back the throttle setting to scheduled 2.18 EPR with resulting reduction of real power to 92% N1 or less.
Some months after the Air Florida accident In January 1982, this writer was dead heading in the jump seat of a 737-200 involved in an incident during a night take-off from Nauru atoll in the Central Pacific. The 5600 feet runway had no runway end safety areas. A road crossed five metres after the threshold and beyond were large boulders forming a sea wall. The take-off was runway length limited, with full bleeds-off thrust of 2.18 EPR required. The first officer was PF and the thrust levers were set for 2.18 EPR. The gauges had a needle and digital read-out.
Normally this amount of thrust is characterised by significant acceleration felt during the early part take-off roll especially as the aircraft weight was several tonnes less than maximum structural. Mid-way during the take-off run there was a vague feeling by the PM and the jump seat occupant that something wasn’t quite right; although a glance at the engines EPR indications confirmed that 2.18 EPR was correct. With about four runway lights to go, it suddenly became obvious that the aircraft was not going to become airborne before the end of the runway.
The captain took immediate control and “firewalled” both thrust levers against the stops. He simultaneously was forced to rotate some 15 knots below scheduled VR. There was an immediate feeling of acceleration with the over-boost and the aircraft became airborne. On setting scheduled climb power of 1.93EPR at flap retraction height, the acceleration was less than normal and the rate of climb at 280 knots clean was also less than expected. A subsequent FDR read-out after the incident revealed the 737 had flown at 20 feet over the water for half a mile before climbing.
During the climb, Operation Manual tabulated EPR v N1 revealed a marked discrepancy between the N1 and EPR relationship with the N1 approximately 10% low for the EPR gauge readings. In all probability the discrepancy occurred during the take-off run and was unknown to the three pilots in the cockpit. The N1 gauges are quite small and with flight deck illumination turned low for the dark night take off, the crew misread the N1 needles indication. In fact, the actual N1 readings during the take off run were probably about 92% and not 101% as planned; the difference being about three millimetres on the gauges.
There had been previous reports of erroneous EPR indications during take-off's at Nauru. On one occasion midway on take off roll, one gauge showed 2.25 EPR with both thrust levers set equally, while the other engine showed correctly at 2.15 EPR. . Without notifying the captain who was PF, the first officer attempted to pull back that thrust lever to set equal indicated EPR. This caused marked asymmetric thrust and the captain quickly directed him to set correct scheduled N1 and to disregard the erroneous EPR reading. The take-off was continued and the fault disappeared shortly afterwards.
In the case of the first named incident, a decision was made to return to Nauru and land. Engineers discovered that both engine PT2 probes were partially blocked by dust and insect debris – the dust coming from phosphate particles blown by wind from phosphate mines on the island. Crews knew of the Boeing warning of the possibility of erroneous EPR readings caused by ice blocking PT2 probes. But this was Nauru, where the OAT was 30 degrees Celsius and no possibility of icing. Also, on previous occasions only one engine was affected - but never both simultaneously. If only one engine was affected it would be easily noticed by a split in the throttles to maintain equal EPR. Later investigation revealed that the engine covers to the front of each engine designed to prevent foreign objects entering the PT2 probes, had not been in place on the night of the incident.
Only prompt action by the captain to firewall both engines to get airborne, saved the aircraft. At daylight it was noticed that jet blast from the 737 had lifted debris over the threshold. A late abort would have spelled disaster.
The crew used reverse thrust to assist the push-back off stand and taxied close behind another aircraft for heat from the exhaust which may have contributed to the airframe show/ice accretion. They also did not use engine anti-ice while taxiing out during snow causing the PT2 (EPR) probes to misread. Although anomalous engine readings were called during the take-off, the captain neither aborted nor adjusted the thrust levers to allow sufficient thrust for take-off and the aircraft stalled and crashed into the frozen Potomac River.
Take-off was almost an hour after de-icing had been completed and the CVR shows that the crew were aware of 10-20mm of snow on the wings. Four of the five crew members and 70 of the 74 passengers were killed.
See description of the accident.
https://en.wikipedia.org/wiki/Air_Florida_Flight_90.
See description below of erroneous EPR indications that caused the crash:
https://books.google.com.au/books?id=oMkMBAAAQBAJ&pg=PA86&lpg=PA86&dq=erroneous+EPR+readings+due+blocked+PT2+probes&source=bl&ots=qUfgr-D17-&sig=TkQulRYYs5ghAirK7aupBL-Qp8c&hl=en&sa=X&ved=0ahUKEwjCtsSD2pXWAhWFH5QKHQH9C10Q6AEIKDAA#v=onepage&q=erroneous%20EPR%20readings%20due%20blocked%20PT2%20probes&f=false
:
Centaurus comment.
Following that accident Boeing issued a memorandum to all operators of the importance of checking the N1 during the take-off as the NTSB investigation showed that blocked PT2 probes could cause the EPR gauges to over-read. For example, if the take-off performance charts required a take-off setting of (say) 2.18 EPR with accompanying N1 of 101%, a blocked PT2 probe meant that an N1 of 101% would likely result in 2.25 EPR. A pilot would think that an engine was over-boosting and pull back the throttle setting to scheduled 2.18 EPR with resulting reduction of real power to 92% N1 or less.
Some months after the Air Florida accident In January 1982, this writer was dead heading in the jump seat of a 737-200 involved in an incident during a night take-off from Nauru atoll in the Central Pacific. The 5600 feet runway had no runway end safety areas. A road crossed five metres after the threshold and beyond were large boulders forming a sea wall. The take-off was runway length limited, with full bleeds-off thrust of 2.18 EPR required. The first officer was PF and the thrust levers were set for 2.18 EPR. The gauges had a needle and digital read-out.
Normally this amount of thrust is characterised by significant acceleration felt during the early part take-off roll especially as the aircraft weight was several tonnes less than maximum structural. Mid-way during the take-off run there was a vague feeling by the PM and the jump seat occupant that something wasn’t quite right; although a glance at the engines EPR indications confirmed that 2.18 EPR was correct. With about four runway lights to go, it suddenly became obvious that the aircraft was not going to become airborne before the end of the runway.
The captain took immediate control and “firewalled” both thrust levers against the stops. He simultaneously was forced to rotate some 15 knots below scheduled VR. There was an immediate feeling of acceleration with the over-boost and the aircraft became airborne. On setting scheduled climb power of 1.93EPR at flap retraction height, the acceleration was less than normal and the rate of climb at 280 knots clean was also less than expected. A subsequent FDR read-out after the incident revealed the 737 had flown at 20 feet over the water for half a mile before climbing.
During the climb, Operation Manual tabulated EPR v N1 revealed a marked discrepancy between the N1 and EPR relationship with the N1 approximately 10% low for the EPR gauge readings. In all probability the discrepancy occurred during the take-off run and was unknown to the three pilots in the cockpit. The N1 gauges are quite small and with flight deck illumination turned low for the dark night take off, the crew misread the N1 needles indication. In fact, the actual N1 readings during the take off run were probably about 92% and not 101% as planned; the difference being about three millimetres on the gauges.
There had been previous reports of erroneous EPR indications during take-off's at Nauru. On one occasion midway on take off roll, one gauge showed 2.25 EPR with both thrust levers set equally, while the other engine showed correctly at 2.15 EPR. . Without notifying the captain who was PF, the first officer attempted to pull back that thrust lever to set equal indicated EPR. This caused marked asymmetric thrust and the captain quickly directed him to set correct scheduled N1 and to disregard the erroneous EPR reading. The take-off was continued and the fault disappeared shortly afterwards.
In the case of the first named incident, a decision was made to return to Nauru and land. Engineers discovered that both engine PT2 probes were partially blocked by dust and insect debris – the dust coming from phosphate particles blown by wind from phosphate mines on the island. Crews knew of the Boeing warning of the possibility of erroneous EPR readings caused by ice blocking PT2 probes. But this was Nauru, where the OAT was 30 degrees Celsius and no possibility of icing. Also, on previous occasions only one engine was affected - but never both simultaneously. If only one engine was affected it would be easily noticed by a split in the throttles to maintain equal EPR. Later investigation revealed that the engine covers to the front of each engine designed to prevent foreign objects entering the PT2 probes, had not been in place on the night of the incident.
Only prompt action by the captain to firewall both engines to get airborne, saved the aircraft. At daylight it was noticed that jet blast from the 737 had lifted debris over the threshold. A late abort would have spelled disaster.