Narrative:

An instrument-rated private pilot and I decided to conduct a training flight in actual instrument conditions. The student had planned and filed a roundtrip IFR flight plan. Weather in the area was reported and forecast to be 800-1;000 feet overcast with visibility 2-6 miles. We planned to do an ILS approach at [another airport] followed by a missed approach. The return was planned to be with a simulated vacuum system failure; partial-panel; VOR-DME approach; circle-to-land. As we taxied to the run-up area; we performed the standard instrument checks for the attitude indicator; heading indicator; turn coordinator; compass; etc. We noted a 20-25 degree precession on the heading indicator which is more than it had been during previous flights; but we didn't think too much about it at that time. The vacuum suction gauge did read a little on the low side of the green range during the run-up; but that indication was consistent with 3-4 previous flights in this airplane and neither the student; nor I; nor any other previous pilot had noticed or reported any anomalies with the instruments during flights in the previous several weeks. Additionally; the airplane was equipped with two vacuum pumps and there were no warnings on the annunciator panel. The vacuum system had been inspected by the maintenance shop the previous day and no discrepancies were noted.shortly after departure; as we entered IMC; we started experiencing occasional moderate turbulence. We proceeded to fly the ILS approach without incident and executed the climbout instructions as per ATC. TRACON cleared us to proceed direct to the fix from which to begin the VOR-DME approach. My student set up the instruments for the approach and turned the airplane towards the fix using GPS navigation. Over the next 5-10 minutes; my student seemed to be having difficulty maintaining accurate control of the aircraft and we drifted significantly off course. I assumed this was because of the strong crosswind and moderate turbulence we had been experiencing. As we turned back to the course; TRACON queried us and I responded that we were correcting. Several minutes later; I noticed that we were significantly deviating again. Now I suggested he use the autopilot in heading mode to reduce his workload until he could get established back on course and get his mind caught up with what was going on. When he engaged the autopilot; it seemed to me that something was not quite right. We had commanded a turn; but I didn't feel anything in the seat of my pants. Not wanting to trust my kinesthetic senses; I scanned the instruments. It seemed that we were turning; albeit at a slow rate. I attributed that to the turbulence; but I was growing suspicious; thinking about the excessive dg precession during our taxi prior to departure. After a few minutes; we had not corrected back to course and we were tracking about 40-50 degrees off. I took control of the aircraft and initiated a turn back towards course. I watched the ai carefully as I applied inputs to turn right. The ai indicated no turn; or perhaps only a very shallow bank. I scanned the turn coordinator and it did show a standard rate turn. I rolled level using the turn coordinator as my reference; but the ai did not show much movement. TRACON queried us again about our position and course and I declared an emergency; stating that we had an inoperative vacuum system. We requested and were given vectors to the ILS approach at our destination. The return flight was difficult; but once we stopped looking at the ai; we felt more confident with control of the aircraft. Though we used the compass as our primary heading reference; the GPS track information was very useful; especially once inside the marker on the ILS approach. My student called out the GPS track every 2-3 seconds while I hand flew that portion. I believe this cooperation was a key to us being able to break out of IMC in a reasonable position to make a landing. In retrospect; I'mnot sure if we could reasonably have done anything to prevent this emergency from happening. Though we had some indication prior to departure that the vacuum system was a little off from previous flights; every piece of supporting information indicated that there were no problems. I do think however; that we should have accepted the no-gyro vectors offered by ATC. I think that due to the workload and stress of the situation; once we had a plan that allowed us better control (use turn coordinator and GPS information); I effectively blocked out any other suggestions that were a significant departure from that plan. Another area where I should have done something different is with how I interacted with the student.... I usually tend to rely singularly on my experience and role as instructor and not usually ask the student for opinions and ideas. However; in this case; the student was an experienced private pilot and may have been able to provide more input or at least help me not to get too focused on one course of action. 5-6 days after this emergency flight; the airplane was flown under VFR and no instrument anomalies were noticed. The next day; the maintenance shop discovered a cracked plastic elbow at the vacuum line connection to the attitude indicator. I believe the crack had been there for some time. I think perhaps the moderate turbulence we experienced during our flight may have stressed the elbow such that the crack opened up more. After the airplane sat on the ground for a few days; the elbow probably returned to its normal shape and the crack became small again.

Google
 

Original NASA ASRS Text

Title: C172 instructor pilot and pilot trainee experienced a partial vacuum system failure in IMC; returned to departure airport flying partial panel with GPS directional information and landed in VMC without further incident. A cracked plastic elbow in the vacuum line was subsequently found.

Narrative: An instrument-rated Private Pilot and I decided to conduct a training flight in actual instrument conditions. The student had planned and filed a roundtrip IFR flight plan. Weather in the area was reported and forecast to be 800-1;000 feet overcast with visibility 2-6 miles. We planned to do an ILS approach at [another airport] followed by a missed approach. The return was planned to be with a simulated vacuum system failure; partial-panel; VOR-DME approach; circle-to-land. As we taxied to the run-up area; we performed the standard instrument checks for the attitude indicator; heading indicator; turn coordinator; compass; etc. We noted a 20-25 degree precession on the heading indicator which is more than it had been during previous flights; but we didn't think too much about it at that time. The vacuum suction gauge did read a little on the low side of the green range during the run-up; but that indication was consistent with 3-4 previous flights in this airplane and neither the student; nor I; nor any other previous pilot had noticed or reported any anomalies with the instruments during flights in the previous several weeks. Additionally; the airplane was equipped with two vacuum pumps and there were no warnings on the annunciator panel. The vacuum system had been inspected by the maintenance shop the previous day and no discrepancies were noted.Shortly after departure; as we entered IMC; we started experiencing occasional moderate turbulence. We proceeded to fly the ILS approach without incident and executed the climbout instructions as per ATC. TRACON cleared us to proceed direct to the fix from which to begin the VOR-DME approach. My student set up the instruments for the approach and turned the airplane towards the fix using GPS navigation. Over the next 5-10 minutes; my student seemed to be having difficulty maintaining accurate control of the aircraft and we drifted significantly off course. I assumed this was because of the strong crosswind and moderate turbulence we had been experiencing. As we turned back to the course; TRACON queried us and I responded that we were correcting. Several minutes later; I noticed that we were significantly deviating again. Now I suggested he use the autopilot in heading mode to reduce his workload until he could get established back on course and get his mind caught up with what was going on. When he engaged the autopilot; it seemed to me that something was not quite right. We had commanded a turn; but I didn't feel anything in the seat of my pants. Not wanting to trust my kinesthetic senses; I scanned the instruments. It seemed that we were turning; albeit at a slow rate. I attributed that to the turbulence; but I was growing suspicious; thinking about the excessive DG precession during our taxi prior to departure. After a few minutes; we had not corrected back to course and we were tracking about 40-50 degrees off. I took control of the aircraft and initiated a turn back towards course. I watched the AI carefully as I applied inputs to turn right. The AI indicated no turn; or perhaps only a very shallow bank. I scanned the turn coordinator and it did show a standard rate turn. I rolled level using the turn coordinator as my reference; but the AI did not show much movement. TRACON queried us again about our position and course and I declared an emergency; stating that we had an inoperative vacuum system. We requested and were given vectors to the ILS approach at our destination. The return flight was difficult; but once we stopped looking at the AI; we felt more confident with control of the aircraft. Though we used the compass as our primary heading reference; the GPS track information was very useful; especially once inside the marker on the ILS approach. My student called out the GPS track every 2-3 seconds while I hand flew that portion. I believe this cooperation was a key to us being able to break out of IMC in a reasonable position to make a landing. In retrospect; I'mnot sure if we could reasonably have done anything to prevent this emergency from happening. Though we had some indication prior to departure that the vacuum system was a little off from previous flights; every piece of supporting information indicated that there were no problems. I do think however; that we should have accepted the no-gyro vectors offered by ATC. I think that due to the workload and stress of the situation; once we had a plan that allowed us better control (use turn coordinator and GPS information); I effectively blocked out any other suggestions that were a significant departure from that plan. Another area where I should have done something different is with how I interacted with the student.... I usually tend to rely singularly on my experience and role as instructor and not usually ask the student for opinions and ideas. However; in this case; the student was an experienced Private Pilot and may have been able to provide more input or at least help me not to get too focused on one course of action. 5-6 days after this emergency flight; the airplane was flown under VFR and no instrument anomalies were noticed. The next day; the maintenance shop discovered a cracked plastic elbow at the vacuum line connection to the attitude indicator. I believe the crack had been there for some time. I think perhaps the moderate turbulence we experienced during our flight may have stressed the elbow such that the crack opened up more. After the airplane sat on the ground for a few days; the elbow probably returned to its normal shape and the crack became small again.

Data retrieved from NASA's ASRS site as of July 2013 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.