Scruffydog7347

I'm an aircraft technician for a major airline. I also have a private pilot's license, single engine, not instrument rated. I often read through accident reports in the hope of learning something. Recently I came across flight ifo-21, the flight that killed Commerce Secretary Ron Brown.
Aircraft: CT43A (737-200). Flying into Dubrovnik runway 12 (119 degrees). Time 3 pm, April 1996. No vor/ils available because the equipment for this had either been blown up or stolen during the recent war. The aircraft only had 1 ADF system. Airport elevation: 500'. Low lying mountains 1.5 miles to the left, the tallest of which was 2,300'. Overcast at 2,000', the top of which obscured the mountain tops. Winds on the surface right down the runway at 15 knots. Winds at 4,000' 160 degrees at 25 knots. light rain.

There are two NDBs in line with the runway. One is KLP, 12 miles from the runway. The second is CV, two miles from the runway. According to AWACs tracking images, this plane passed right over the first beacon at 4,100', pretty much on course. But according again to AWACs, they were going 30 knots too fast with flaps and gear up. With their single adf system, they had to tune in the next ADF, CV. Whether they could tune it in with a flip of a switch or had to 'tune' it, I could not find out.After a slight joggle in course during this tuning time I'm guessing, they then flew a straight 'course' of 109 degrees and hit that highest peak.
The accident investigation said that the ADF was still tuned to that first ADF (KLP) and not CV as it should have been. There were no black boxes on this aircraft.

There was a lot of controversy about this crash saying it shouldn't have been attempted with just one ADF system and the MDA should have been 2,350 instead of 2,150. Granted it shouldn't have been attempted, but I'd still like to ask a couple of questions. From what I've read about the pilots, Capt Ashley Davis and Capt Tim Shafer, they were good pilots. You don't assign second rate pilots to fly dignitaries around. This unit flew Hillary Clinton into Bosnia supposedly under sniper fire the week before.

So how likely is it that both pilots missed the passage of that first ADF beacon and didn't tune in the second beacon? Or is it possible that once they passed right over that first beacon, they trusted the INS system enough to keep them on line with the runway? The third question I have is if after they passed over the first beacon, if they tried tuning in the second beacon and couldn't pick it up, is it possible they might have used the previous beacon in a reverse sensing fashion? I know all these scenarios go completely against the book?

badflaps

From what I remember their ADF skills were lacking. (Not unusual, considering the amount available for practice.)

galaxy flyer

What you’re missing is approach design required DUAL ADFs, not one as installed. They only had one ADF and tried to fly it without the required second one. In the ADF era, this design was common—two beacons, one for the FAF and one at the MAP, providing positive identification of the MAP allowed for the minimums published. Jeppesen was brought in to review the design and found it complies with the PANS-OPS in effect at design.

There was another factor—mission accomplishment overrode AF Instructions. The crew and the supervisory team were dinged for not conducting a review of the approach and the equipment suitability. Including a General. Sounds like you have an agenda.

GF

Excargodog

And of course the two preceding aircraft, both appropriately equipped, had both gone missed approach. I think it is very likely that political pressure was placed on the aircrew to make the attempt, probably not by Brown himself but by one of the officials little aides who serve as dog robbers for the senior political class. And I think the aircrew caved to it.

I seriously doubt either of the pilots had actually done an NDB approach .... Perhaps ever. T-37s and T-38s didn't even have them, in fact very few USAF aircraft of that era still did. Their general non precision approach was a TACAN approach.

Adlerdriver

You make a few statements that I either disagree with or have some initial comments about.
Without a CVR or FDR, there's a lot missing from any detailed discussion of the crew actions. My general impression of what we do know shows a classic rushed, unstable approach with a late configuration. I believe one reason for this is a lack of distance information from the airfield and the navaids in use.

USAF pilots "grow up" flying from AF bases that are typically equipped with a TACAN co-located with the base. In case you're unfamiliar, TACAN operates very similar to a VOR/DME but both pieces of information (range and azimuth) are obtained by tuning a single control head to a single frequency. So, with the TACAN tuned on the base to which they're flying an approach, they can determine their exact position in relation to the field with a single glance. As simple as this sounds, knowing this is critical when calibrating one's mental clock and choosing when to slow/configure for landing while flying an approach in IMC conditions.

In this accident, these pilots didn't have that luxury. In fact, there appear to be no navaids available that would provide the crew with any sort of DME (distance) information related to the runway. The INS would have been a great back-up here. Programming a lat/long point at the airport or maybe approach end of the runway itself would have enhanced their situational awareness (SA) immensely. Using the INS to aid flying the course from KLP to CV with a wind-corrected heading would also have been a great use of that tool. However, this might not have been possible to due system limitations or may simply have not been something the crew regularly trained to.

If their INS was anything like the one I used in the USAF, they probably couldn't do both of these options at the same time and may have had to choose one or the other. From what I can see online, the INS was found in a position to provide wind data to aid them flying the final approach course. That may have ended up being a moot point if they lost SA on their position from the airport.

So, it's very likely that the crew got rushed or may not have been 100% certain exactly how far from the field they were when they received approach clearance. If they weren't using the INS to provide distance information from the field, their first reliable indication of position would have been station passage at KLP (the final approach fix). They arrived there at 220 knots, most likely in a clean configuration. Normal instrument approach guidelines for USAF pilots would have been to arrive there fully configured at final approach speed. In fact, typical guidelines are to plan to be fully configured, on speed, 3 to 5 miles prior to the FAF. Since this FAF is almost 12 miles from the field (most are more commonly 5-7 miles out), the crew may have chosen to delay configuration slightly.

However, I seriously doubt the crew planned to arrive at the FAF clean at 220 knots. So, I think it's a fair to assume they had lost SA on their position in relation to the airport and were probably scrambling to salvage the approach when they passed KLP. It definitely would have been a scramble. In order to land out of this approach, they would have needed to be down at the MDA (2150' MSL) NO LATER than 5 miles from the airport. This is known as the VDP (visual descent point) and is based on AGL altitude to lose from the MDA to the runway threshold on a typical 3-degree glide path. They probably would have wanted to get down to the MDA prior to 5 miles and level off (the "dive and drive" approach). This would have allowed them a chance to get down to minimums, try to get a visual on the runway and set up a normal 3-degree descent to the runway. Unfortunately, unless they were using INS for distance (or manual ground speed timing) from the runway, they would have no way of knowing their exact distance from the runway during this point in the approach. I can't tell if there was a timing block associated with this approach.

Starting at 220 knots, clean and trying to descend from 4000 to 2150 in less than 6 miles while slowing and configuring is almost impossible. So, to say this approach was rushed is an understatement. One of the first thing any pilot trying to salvage such a descent would do is drop his landing gear. That's going to provide much needed drag and can typically be extended at higher speeds than any high lift devices like flat/slats. Once they're at the MDA (2150) and at the VDP, every second that passes, every fraction of a mile closer to the runway they get while not descending on a normal glide path to the runway means it's very likely they would be unable to land out of the approach.

I mentioned earlier the winds were a factor. On any NDB approach, correcting for winds can be a challenge. It takes some trial and error starting from a known, stable reference point to determine how the winds are affecting you, input a correction, determine if that's enough and so on. When you add in additional challenges like trying to lose 80 knots of excess speed, fully configure your aircraft while descending to mins and only using one of two required navaids....... I seriously doubt they had a lot of excess time to try to figure out the winds, even if they were using their INS to help. So, since the winds were from 40-degrees right (160⁰), they got blown to the north of course during the approach and missed approach.

Some of your statements about the ADF use on this approach are incorrect. As has been stated, two ADFs were required to fly the approach. So first, saying there is a proper tuning hierarchy - i.e. when a pilot should tune one navaid instead of the other, is incorrect. Two NDBs should have been in use throughout the approach. Second, the final approach guidance from KLP to the runway (or missed approach point) would have been provided by KLP, not CV. NDB signals are generally more stable and easier to track the further from the station you get. Using the CV NDB would become more difficult the closer they got to it. The display of its signal would become more and more sensitive to changes just when it's guidance was needed most. So, the only use of CV on this approach was to identify the missed approach point by signaling station passage. Finally, the missed approach procedure called for a right turn back to KLP. So, having that navaid tuned during the missed approach was, in fact, the correct setting for their primary (and only) ADF in this case.

I'm sure they were good pilots. No matter how good someone is, there's a point where enough negative factors on a situation can overwhelm anyone. Improper navaids, lack of "typical" distance guidance, pressure to get the job done, unfamiliar non-precision approach, rushed, unstable approach, terrain issues and some weather. Any one of those things would be a threat. All of them combined was obviously a recipe for disaster.

Adlerdriver

This is not correct. The T-37 was equipped with an RMI which is essentially the same thing as an ADF. While the main focus in training was use of TACAN, LOC or ASR for non-precision approaches, we did fly and train with "RMI only" approaches. I honestly can't remember if we could tune NDBs with it or simply used the azimuth only signal from a VOR.

My point is that both of these pilots received training in UPT that would have allowed them to fly an NDB. I also expect that since their aircraft was so equipped (obviously not enough for this particular approach), they received regular training and check rides involving NDB approaches.

I also wonder where you got the information about the preceding two aircraft going missed. The only thing I could find for certain was a pilot who had landed one hour earlier radioed the crew while they were on approach and told them weather was at mins for the approach. The final report concluded weather was not a major factor in the approach. 500 BKN, 2000 OVC and 5 miles vis isn't great but it's hardly a show stopper.

galaxy flyer

Ausgezeichnet, Adlerdriver!.

The seeming failure to use the INS to keep up the SA; plan the slowdown at KLP; and use the wind output to help with tracking was a mystery to me. It probably had the Carousel “whirling steel” INS, but plenty good enough for those tasks. We used it that way in the C-5 all the time. If they had just done a turn in a hold at KLP, it all might have worked, illegally, but they wouldn’t have contended with the high energy problems.

We did plenty of “RMI Only” procedures in UPT in both planes. Then again, I was going to the Hun, which had a Tweet panel, essentially. . The Tweet didn’t have an ADF, but the Hun did and we were checked on ADF in RTU.

GF

Adlerdriver

Was ist los Herr Flyer?

We've all been there and that willingness to admit temporary defeat and request a turn in holding can be tough to do. I agree the outcome probably would have been far different had they chosen to "light up a lucky" and wind the clock.

I don't remember an RMI in the -38 but I'll take your word on it. I'm stretching my memory on a lot of that stuff as it is. The only reason I remember it being in the T-37 was due to the NDB-A with a circle to land opposite direction to actual mins I did on my ATP Seminole check ride. After I landed I just remember thinking "how in the holy #@$% did I pull that out of my a$$?". I just chalked it up (very thankfully) to T-37 muscle memory and a vague recollection of that damn RMI (not to mention it's almost impossible to get behind in an airplane that slow).

Tweetdrvr

GF, auch eine gute Erklarung!!

The T-37 was VOR/DME and ILS only. The T-37 syllabus I went through in the early 90s did not emphasize RMI only. I learned most of my RMI/BDHI only stuff in the T-43 at MHR in SUNT in the late 80s. As a student, I lived off the RMI because I never really liked the crappy CI and the toilet bowl indicator in the T-37. It was really only useful (IMHO) for ILS/GS info.

As a T-37 IP in the late 90s and early 2000s, I found it useful to emphasize RMI only stuff to keep students from chasing the CI and over correcting based on CI displacement. Understanding of the VOR needle would have shown them only 2-3 radials off course, and required only enough change to drop the head or raise the tail of the needle to the desired course. The guys with prior GA time who had flown NDBs or the Nav rated students going through were able to grasp this better than the other students. It also seemed like the non fighter MWS IPs also put a bit more emphasis on the RMI than the FAIPs and Fighter MWS IPs. I would attribute this to the backgrounds of flying NDBs out in the AMC world of Africa, Eastern Europe, and South America.

The only formal RMI only training I received as a pilot was at Little Rock going through initial Herc qual in 1993, because you did NDBs in the C-130. I have heard that the T-1s are trying to drop or have dropped NDBs from the syllabus because they are getting harder to find in the CONUS as we transition to the NextGen NAS. If they do or did, I'd hope they would keep doing RMI only training using the VOR/TACAN needle because there are still places in the undeveloped hinterlands where AMC and Special Ops go that have NDBs as an only option.

I went into Dubrovnik after the Ron Brown accident and the wreckage was still being pulled from the side of the mountain. The weather was VMC that day and the high terrain off to the east was clearly visible. We flew the approach in an H model 130 with no difficulty. But we had a Nav onboard, SCNS/INS Nav displays on the HSI, and most importantly 2 NDBs

The big thing I remember from the accident report was not only did they have only 1 NDB, it was the old school coffee grinder NDB like the C-130 E models had. Not the more modern, set the frequency numbers individually and forget it. Tuning those things was an art form, and you could get the needle to point reasonably well, but not close enough unless you hit the null sound in the audio tone dead on. This was easy to do at LRF because the Toneyville NDB was directly off the approach end of Runway 25. Airplanes were parked parallel to the runway, so you had it correct when the needle pointed at 12 or 6 o'clock exactly. I believe the report said a contributing factor was the NDB possibly not being tuned correctly.

galaxy flyer

I can’t positively say on the -38s, but I did them in Tweets. The CBM 37 Commander (Lt Col Terry) has done several tours in Huns, USAFE and RVN, so he took us three ANG guys under his wing and beat us silly. Character building, but a life lesson. 77-05, btw, memory maybe fading.

The C-5 A-models had two coffee grinder ADFs, the Bs only one digital one. Went into Yerevan, Armenia in an A-model, 2 beacons to the LOC intercept, lots of mountains. Very carefully flown, with all three pilots talking thru each leg. Ridges obscured, but decent ceiling and viz.



Tweetdrvr

The Hun had the identical panel—MM-3, ID-249 and ID-250 at 166 plus fuel.

GF