American Airlines Flight 723

File under: Why we check the tower lights every day (or have an automated tower light reporting system):

56 years ago, on September 16, 1953 American Airlines flight 723 flew between the center and northeast towers of the WPTR antenna system while attempting to land at Albany County Airport.  The plane crashed about 3/10 mile away near NY route 5 (Central Avenue) killing all 28 persons on board.  To date, this is the worst aviation accident in the Albany, NY area.

Several years ago while cleaning out different AM transmitter site, I found a bunch of files about this accident in the trash bin.  It seems that some engineer had moved a file cabinet during the great consolidation of the 1990’s to the wrong transmitter site.  In any case, I rescued the file and for your reading pleasure, have scanned the following documents:

Original telegram to FCC in Washington DC regarding tower/aircraft colision
Original telegram to FCC in Washington DC regarding tower/aircraft colision

Retel WPTR north and center towers struck three hundred foot level 0930 morning 16 September by American Airlines Convair.  Damage inspected afternoon 16 September by representative Zane Construction and afternoon 17 September by engineer Ideco tower used in array.  Both report slight damage to center tower requiring straightening above 300 foot level.  North tower has two legs bent result of wing passing through tower 18 inches above top guys at bolt intersection.  Tower above bent and twisted but all right unless subject to high east wind.  Impact sheered north beacon clevises and shattered glass.  Replacement ordered and beacon restoration expected early next week.  Measurements of directional pattern afternoon 16 September show pattern unchanged and mulls, base currents and loop currents within tolerances except center loop current which reads ten percent low.  Indications are center pickup loop has been jarred and pattern unaffected.  FCC representative Turnbull who arrived noon Wednesday concurred.  No time lost.   Operating at full power

George Wetmore, Assistant Genl Mgr, Radio Station WPTR

This is the statement of the transmitter engineer on duty at the time of the crash.

Statement from the Engineer on duty at the transmitter site
Statement from the Engineer on duty at the transmitter site

I, Robert S. Henry engineer for Patroon Broadcasting Co. would like to make this statement concerning the collision of an American Airlines Plane with our transmitting towers.

On Wednesday morning September 16, 1953 I was on duty alone at the WPTR transmitter when at approximately 9:30 A.M. a low flying plane was heard overhead. The carrier trip circuit which protects the transmitting apparatus from sudden overloads at the antenna almost instantly actuated and a low sound of explosion followed about 3 seconds later. I ran to the back door of the transmitter in time to see what proved later to be aluminum sheets fluttering down through the fog. I estimate the ceiling at the time to have been approximately 75 feet and not sharply defined. The centre (sic) tower of our three tower array swayed violently for approximately for approximately (sic) 1 minute. At the time I knew a plane had hit the tower but it was above the ceiling and invisable to me. I called W. R. David and George Wetmore and the incident was reported to the Albany airport.

Back in these days, the studios were located at the Hendrick Hudson Hotel in Troy, the transmitter site was manned by a licensed transmitter engineer when ever the station was on the air.

What is amazing is that the IDECO towers remained standing after being struck by Convair 240, a pretty good sized aircraft.

Convair CV-240
Convair CV-240

Official report by the International Civil Aviation Organization:

American Airlines’ Flight 723 was a scheduled flight between Boston, and Chicago, with intermediate stops among which were Hartford (BDL), and Albany (ALB). The CV-240 arrived at Bradley Field at 06:57. Weather at the next stop, Albany, at this time was below the company’s landing minimums, but was forecast to improve to within limits by the time the flight arrived there. Departure from Bradley Field was made at 07:14. Because of poor visibility at Albany, several aircraft were in a holding pattern. The special Albany weather report issued at 07:50 indicated thin obscurement, ceiling estimated 4,000, overcast, fog, visibility 3/4 miles. Two aircraft left the holding pattern, attempted to land, but both executed a missed approach procedure. A third airplane landed at 08:16 following an instrument approach to runway 19. Immediately following this landing, Flight 723 was cleared to make an instrument approach to runway 19. Three minutes later the flight advised the tower that its approach was being abandoned because the aircraft’s flaps could not be lowered.
At 08:30 Albany Tower reported:”All aircraft holding Albany. It now appears to be pretty good for a contact approach from the west. It looks much better than to the north.” Flight 723 was then cleared for a contact approach to runway 10. On finals for runway 10, the Convair descended too low. The right wing of the aircraft struck the center tower of three radio towers at a point 308 feet above the ground. The left wing then struck the east tower. Seven feet of the outer panel of the right wing including the right aileron and control mechanism from the center hinge outboard together with 15 feet of the left outer wing panel and aileron separated from the aircraft at this time. Following the collision with the towers, ground impact occurred a distance of 1,590 feet beyond the tower last struck. First ground contact was made simultaneously by the nose and the left wing with the aircraft partially inverted.
The weather reported at the time of the accident was thin scattered clouds at, 500 feet, ceiling estimated 4500 feet, broken clouds, visibility 1-1/2 miles, fog.

PROBABLE CAUSE: “During the execution of a contact approach, and while manoeuvring for alignment with the runway to be used, descent was made to an altitude below obstructions partially obscured by fog in a local area of restricted visibility.”

The above reports notes that the aircraft traveled 1590 feet and struck the ground partially inverted.  I do not know what the flaps up landing speed of a Convair CV-240 is but the cruising speed is 280 MPH.  It would be safe to say the aircraft was traveling in the 120 to 130 MPH range or about 220 feet per second.  At that speed, it was likely airborne for about 7 seconds after it hit the tower.  Enough time to look out the window, realize what was happening and say “Oh, Shit!”

IDECO towers WDCD antenna system
IDECO towers WDCD antenna system, northeast tower is farthest

IDECO stood for the Internation Derrick Company, they build cranes, derricks and bridges as well as radio towers. Apparently they made pretty good stuff because those same towers are still standing today.

Not that checking the tower lights would have averted disaster in this case, it appears to be pilot error compounded by bad weather that caused this incident.  But there have been more recent aircraft/radio tower accidents, some of which have involved possible faulty tower lights.  I wouldn’t want that on my conscience.

Module swap guys

Nautel V-10 Power Supply
Nautel V-10 Power Supply

Gone is the day when the radio station engineer had to trouble shoot down to the component level, often crawling in and out of transmitters to get at the suspected part.  I for one, spent many a long night at a transmitter site chasing some weird combination of symptoms down to the $0.34 1N914 diode in the directional coupler (see previous post about the MW-50).

It is a skill set now mostly confined to manufactures’ repair departments, for which they charge a pretty penny.  Nowadays, the technician simply slides out one module or circuit card and slides in another.  If that doesn’t fix it, panic ensues.  I know of several class C FM radio stations that are now relying on the computer guy to fix transmitters, because, you know, it’s cheaper.

To be fair, most engineers are contractors and many of those simply do not have the time to trouble shoot to the component level.  So, they ship everything back to the factory then pass the cost on to their client.

Then of course, most circuit boards these days are surface mount systems, which are hard to work on if you don’t have the right tools.  Normally an expensive temperature controlled soldering station is required, as well as a magnifying glass.

All of these things combine make circuit board work something to be outsourced.  Unfortunately, a night spent trouble shooting was often a great learning experience.  I have done some of my best work when my back was up against a wall and I was out of options.

I make the attempt to fix things locally, unless the transmitter or other item is under warranty or not having a spare/attempting to trouble shoot will take the station off the air.  I think it is important to keep abreast of technology and keep my trouble shooting skills up to par.  Besides, I find it gratifying that at least I can still fix things.

My friend, the MW-50B

I began fooling around with radios when I was 10 years old or so. First, I built one of those shortwave radio kits from Radio Shack, which was back when they still sold radios.

Then I bought a small tube type AM transmitter at a garage sale.  The woman there said her son built it several years ago from a kit and it had the instruction manual.  I don’t even know who made the kit.  After some experimentation and changing out some tubes, I got the thing to transmit on about 1600 kHz, although it was a little hard to nail down as it drifted quite a bit until everything heated up.  I don’t know what power that thing put out, but it was certainly less than a watt.

All of this lead to a brief stint in the military as a radioman.  That was an interesting field, albeit different from what I thought it would be when I signed up.  It was during this time that I did some part time work at an AM/FM/TV station assisting the Chief Engineer.  Once it was established that I actually knew something, my responsibilities grew until I was assigned the AM/FM part of the deal.

After a year of that, I moved to a different city for family reasons and took the Chief Engineer job at a local AM/FM station.  The AM station was a 50,000 watt directional in the high end of the band which had a Harris MW-50B transmitter.  My previous station had a Bauer 10,000 D AM transmitter.  What could be so different? Plenty I learned, on my second day.

Harris MW50B transmitter with 50 KW air cooled power supply
Harris MW50B transmitter with 50 KW air cooled power supply

We were subjected to a wicked lightning storm, which, Murphy being present, took out the main transmitter.  The backup was a GE BTA25 which was running at half power because of the age of the 5891  final tubes.

The symptoms of the MW-50 where as follows:  It would run along fine then there would be a big blue flash and a cannon shot boom, followed by the step start relays cycling and it would come back on the air.  There were no overload lights nor any other symptoms leading up to the overload or subsequent to it.

I began by killing the power and shorting out all the high voltage parts with a shorting stick.  I noticed that things inside this transmitter where a little unusual, so I got the manual out and started reading.  The most unusual aspect of this transmitter is the 25 KV isolated box that the PA stage occupies.  25,000 volts DC is a great big potential and what I found over the years is that this transmitter needs to be kept very clean.  Of course, this unit had not been, and that was a part of the problem.

The other unique aspect of this transmitter is the damper diode, which is required by PDM transmitters to conduct voltage during the negative modulation peaks.  If the damper diode breaks down for any reason, the PA supply voltage tries to go to infinity, which is a good deal larger than 25KV and all sorts of problems begin.

To make a long story somewhat shorter, this is the problem I had.  The solid state damper diode had one bad section, which was causing all sorts of corona problems during heavy negative modulation peaks.  It took a call the Harris factory to determine this.  The entire diode assembly needed to be replaced because every section is matched.  That cost a couple of thousand dollars as I recall.

While I was working on the MW-50B transmitter, I was not impressed.  It seemed a little cheap and flimsy.  Later, when I voiced my concerns with the station management, the Harris transmitter salesman stopped by and said I needed to get with the program if I wanted to work in that market.  This was a Harris town you see, if you start bad mouthing our products, you’ll be the one to suffer.  Well, he retired, I kept looking around for other AM transmitters.  Three years later I went to work for the competitor across town.  Today that station has a Nautel ND-50.

The MW50 went off the air once every 6 months for the entire time I worked at this station.  It was always something different, power supply rectifier, bad PDM board, bad directional coupler, arcing insulator on the isolated box, etc.  I began to feel it didn’t like me, and I know I didn’t like it.  In fact, you could say I have never really liked Harris transmitter products ever since.

Update: Okay, I left a few things out of the narrative:

The 50 KW air cooled power supply was the light weight version.  Most MW-50 transmitters had 100 KW oil cooled supplies.  The problem with the 50 KW power supply was it was designed with a zero safety factor.  All of the rectifier were running at or near maximum current and voltage.  It only took one of 144 diodes to go bad, either short or open, and the whole transmitter would crash.  Again, no overload lights or other indications of problems.  We later installed air flushing fans in the power supply cabinet to keep things cool and that helped out quite a bit.

The other thing was a DC feedback sample to the PDM card.  It seems that if the filaments were turned off before the bleeder resistors took the 25 KV supply to zero, the remaining voltage would be routed to the PDM card via the DC feedback sample, blowing the foil off of the circuit card.  We fixed this by installing a gas discharge tube with a series resistor at the connection point for the DC feedback sample.

Then there are the infamous 1N914 diodes in the directional coupler that Dave points out below.

I am sure I am forgetting something else, but you get the idea.