In the late 1970s I was based at Manassas, Virginia. I flew a Piper Turbo Arrow III in my travels for Flying magazine.
The Arrow was due for its 24-month altimeter check. As some shops do, an exchange altimeter was tested and certified on the bench. After that altimeter was properly tested and tagged it was installed in the Arrow and only a simple static system leak check was required to complete the recertification.
The altimeter was an encoding type, meaning the internal mechanism drove an optical device to encode pressure altitude for the transponder to report. With this type of altimeter the transponder would send the controllers the same information that I would see on the face of the instrument. And that was the rub.
The next flight after the altimeter exchange was, as usual for me in those days, out to Wichita. I needed to be there mid-morning the following day so I decided to fly to Evansville, Indiana, the evening before. Evansville was a little more than halfway to Wichita, and there was an excellent FBO on the field, and a hotel a short walk away. A great overnight stop I used often.
I departed near sundown. In those days Manassas was a very basic one-runway field with no control tower. The drill was to depart VFR when the weather permitted and immediately call nearby Washington Dulles approach control for your IFR clearance. Which I did.
The only unusual aspect of the departure was that when the Dulles controller gave me the altimeter setting it was way off what I had dialed in before takeoff. I was sure I had set field elevation into the altimeter as was checklist procedure. There were no automatic weather reporting stations then, and weather observers who could provide an actual altimeter setting were confined to larger airports like Dulles, or to some remote fields in the hinterlands where the weather service needed observation data.
I blamed the big change in altimeter setting I received from Dulles on myself for mistakenly setting the wrong field elevation before takeoff. It was my last and only chance to have prevented the near disaster that was ahead.
It was a great evening for flying IFR. I was on top of a solid overcast nearly all the way to Evansville. The air was smooth and frequencies quiet and perennial headwinds rather light.
The overcast at Evansville was just above visual approach minimums and the controllers cleared me to 3,000 feet and provided a vector directly to the airport. I popped out of the clouds before reaching 3,000 feet into clear night air. Jeez, the ground looked close. I brushed off the sensation as normal after spending a couple hours in the dark, out of sight of anything on the ground.
At the time there was a bunch of strip mining activity in the area and the enormous shovels that dug up the earth all had red obstruction lights. I flew past one of those lights and it looked really close.
A little panicked, I started looking around at other lights and towers and realized I wasn’t anywhere near the 2,500 feet above ground I should have been with the altimeter showing 3,000 feet. I spotted the airport and flew over to intercept the ILS and track the glideslope to the runway. When I landed the altimeter showed I was more than 1,800 feet above the pavement.
The next morning the altimeter showed the Arrow sitting more than 3,000 feet above the ramp elevation. If the weather conditions had been worse I could have flown into something with the altimeter showing me precisely at a safe altitude.
In some ways even more alarming, I had flown hundreds of miles in the dark at some unknown altitude under IFR. The separation the controller thought I had from other traffic was mythical. Because it was an encoding altimeter, the transponder sent down the same information it showed me on the altimeter dial. We had both been doing our jobs perfectly with unknown separation from other traffic.
TCAS collision avoidance systems didn’t exist then, but even if they had, they would have offered no protection because they rely on the Mode C altitude reported by nearby traffic to detect a collision threat. My transponder would have been reporting a non-threatening altitude when who knows what my real altitude was.
The altimeter has occupied a sacred place in basic instrument flying. It’s the only one in the “six pack” of instruments with no backup. No combination of the other five instruments can reveal your altitude in the way comparing the others can cross check on their function.
Larger airplanes have two altimeters so if they disagree you can be alert to a problem. Jets have three altimeters so you can compare and ignore the odd man out. But in light airplanes most of us flew blissfully along believing what those three hands on the altimeter dial showed us.
The situation is better now for several reasons. First, encoding altimeters are virtually all gone. The extra friction of moving the encoding system as well as the instrument hands on the dial doomed them to a relatively short life because of hysteresis so they couldn’t pass the 24-month check.
Development of low-cost digital electronic pressure sensors brought the price of blind encoders—so called because they don’t display their output on an instrument—down to a few hundred bucks. With a blind encoder, the controllers will see something different if your regular altimeter fails because they are not connected in any way. You may not know why there is a difference between your altimeter display and Mode C the controller sees, but you will be alerted rather than flying along fat dumb and happy as I did on that night.
The digital electronic altimeters that are part of the primary flight displays and other electronic instruments so many of us now have in light airplanes are not immune to failure, but they have a decent level of monitoring. You’re likely to get a flag warning if something goes wrong.
And most fundamentally, at its root ADS-B transmits a GPS measured height. In any system I’ve flown you can use menus to find that ADS-B altitude info, and equally important, the controllers can see that independent of your Mode C.
But, back in the bad old 1970s we had none of those stay alive advantages. So I bought a high quality, pocket-size altimeter built for the hiking/climbing market. For years when at cruise in a non-pressurized airplane I’d pull that thing out, set in the current baro, and compare its reading to the altimeter in the panel.
Never did a see a significant discrepancy between that little altimeter and the real one. But on that night I learned it only takes once. If I hadn’t been so lucky I wonder what the NTSB probable accident cause would have read. Pilot error, no doubt.
- Cirrus and Bonanzas dominated their eras - May 22, 2023
- Lead-free avgas STC—a historical preview? - November 21, 2022
- What Nexrad can—and can’t—tell pilots - March 14, 2022
Very good article, Mac. I have never heard of that occurrence of that type before but it was a good thing you were not flying out West somewhere and had to make an approach around some high terrain.
One very big problem with the old three needle altimeters was the 10,000 foot marker which a lot of people did not pay much attention to. The problem was that if you did not notice it you were 10,000 feet off and that would definitely kill you if you made that mistake! There were several fatalities due to that and it is the reason that all the airlines switched to the digital drum counter with only two needles. It was much safer and eliminated the 10,000 foot error which I have seen occur several times with pilots leveling off or miss reading their altimeters by 10,000 feet.
You’re exactly right about the 10,000 foot error potential of the three-pointer altimeter.
Another pointer worry some safety “experts” express is the altimeter clock some pilots hang in their home or office. The concern is pilots will become accustomed to reading straight up as 12 instead of 10 that is straight up on a real altimeter.
An over-blown concern, probably. But who knows?
Altimeters can present a few problems just by being what they are, and how we interact with them, yes. I climbed into an Airbus 320 one night in Baltimore just after an extreme low-pressure storm, a “Nor’ Easter”, front had moved out of the area. While passing in the jet bridge before I stepped aboard the airplane, the inbound crew said the weather cleared to beautiful VFR conditions just as they began their descent to Baltimore and the ride was good. Consequently, they never did have to level off anywhere; they simply descended all the way to the runway. After settling in, I received the ATIS information with an altimeter setting of an astounding 28.97 inches!! I had never seen it that low before… I then proceeded to set all three altimeters in the cockpit. To my astonishment, the inbound crew had set all three altimeters to 29.97 inches. All three were one-thousand feet off! From that day forward I promised myself that when responding to “altimeters” during a checklist I would verbalize all four digits. And, so far, for over twenty years I have done just that.
Thank you for sharing with us! That is scary but luckily you trusted your gut instincts. I am a student pilot and whenever there are hundreds of extremely obscure instrument questions that I want to skip over, I will say to myself, “because you need to know when and why you have instrument errors and how to correct them!! You can’t rely on other people to notice in sketchy situations!”
Hi Mac, thank you for bringing attention to the lowly altimeter. I work on these systems for a living and have seen my share of issues. Even with a blind encoder, there are single points of failure that can cause an altimeter discrepancy plus provide an incorrect altitude to ATC (static system blockages come to mind). We have seen a return to “encoding” altimeters in the form of air data computers in EFIS systems – many of these computers feed both the altitude on the Primary Flight Display and the transponder / ADS-B system (virtually all of these systems have backup altimeters, but they likely share the same static system). I had a CFI buddy tell me to pay close attention to the glide slope intercept point on precision approaches. If there is a significant discrepancy in altitude when you cross the intercept point on the approach something is not right – it could be your altimeter or could be a navigation failure – in any case you have a real problem! Also, be aware of emergency procedures for your particular aircraft (does your aircraft have alternate static, what are restrictions on it’s use, have you ever tested it in flight in vfr conditions, etc). Again, a great article! Thank-you for posting it!
Great points, Josh. It’s crucial to understand the altimetry system in the specific airplane you fly.
For example, the King Air 350 I now fly as my semi-retirement job has three altimeters, or more exactly, three air data computers. But it has only two static systems. The standby instrument shares a static system, so there is a point of failure that could be challenging to identify.
However, airplanes certified in the transport category have at least three totally independent altimetry systems from static port to instrument display.
So, what is the rest of the story? Did you fly to Wichita with a bad altimeter. When and where did you get it fixed? Did the original shop take it back and give you a good one?
The shop on Evansville airport replaced the altimeter and I flew to Wichita. Got there late.
The Arrow was leased from a Piper dealer and the deal included maintenance. Since the dealer also owned the shop that did the original work, I don’t know how the costs were handled.
I do remember the Evansville shop told me the altimeter was junk because the aneroid bellows that is the sensor in mechanical altimeters had failed, probably a leak.
About 1970 I landed at an airport and was on the ground for about 15 minutes. When I came back to my aircraft the altimeter showed about 500 Ft, higher than when I had landed. When I made arrangements to send the altimeter to a rebuilder, he said that it was probably a temperature compensating spring that had failed. Thinking of this was one of the reasons I never got an instrument rating.
Your story is a very good one stuffed with several lessons to live or die by.
As you point out, it’s kinda difficult to “trust, but crosscheck” (also known as ‘Trust, but VERIFY’) with just one of those darn altimeters at night.
Didn’t the avionics shop you got your lender altimeter from bench check it before it was installed in your aircraft before your night flight? It’d be interesting to read a transcript of your conversation with the lead tech at that shop when he or she was informed of the very near miss.
Your story is clearly another really good reminder that ANY maintenance, ANY new avionics or swap out of avionics with a borrowed shop instrument must to be treated the same: A very thorough pre-flight (what could go wrong?) with a new (to you) altimeter? Then a post takeoff checkout in the pattern under VFR day conditions… hopefully when the shop is still open for business.
BTW, a friend of several years told me about a flight he took on an undisclosed airline about a decade and a half ago. He was wearing a wristwatch equipped with an altimeter that he used for work. He noticed during the climb after takeoff that his watch had just passed through 11,000′ msl. He rang the cabin crew call button and told the responding stewardess that (by then) they were passing through 14,000′ feet. The aircrew leveled off soon after. It was a pilot induced error of omission or commission ’cause climbing resumed after a few minutes. The lesson from that one was… Errors are sometimes made by pro crews too, just like with maintenance techs and mechanics.
Yes, the altimeter had been certified on the bench and “yellow tagged.” It also performed normally when it was installed and the static system checked a few days before I took off.
As I said, I did have one chance to find the problem when I received a very different baro setting from Dulles approach then what I had set before takeoff. But I blamed myself for an error in setting the altimeter on the ground, not a failure of the altimeter.
Flying IFR at night in basic single engine airplanes requires many leaps of faith in items that there are only one of. It is that need for essentially total faith in something as basic and essential as an altimeter that I blame for me doubting myself instead of the instrument.