May 26, 2012
OK so I am working on the VP-400, and artificially-intelligent avionics package that continuously considers power-off glides to every airport within gliding range and will take your airplane down to the best runway within gliding range for a power-off landing at the press of a single (red!) button.
One of the things I need to know to make this system work is how high you are, and how fast you are descending.
You can tell that from an altimeter and a vertical-speed indicator, right? Oh, you are SOOOOOOOOOOOOOOOOO wrong.
All altimeters and vertical speed indicators are based on pressure, and change of pressure with altitude! It is this change of pressure as you climb that winds up your altimeter, and how fast this pressure is changing that winds up your vertical speed indicator.
But here is an interesting thing about the atmosphere: While non-standard PRESSURE is always reported to us in every ATIS broadcast, any non-standard LAPSE-RATE OF PRESSURE WITH ALTITUDE is NEVER reported! I mean, listen to ATIS! Do they say: “Altimeter 29.85”, or do they say “Altimeter 29.85, with a lapse of 0.92 inches for the first 1,000 ft, then a laps of 0.90 for the next 1,000 ft, then a lapse of 0.88 for the next 1,000 ft…” and so on for the entire climb up to the flight levels?
Of course, they ONLY give the pressure, and NOT the change in pressure with altitude!
…And just as the pressure varies from standard on every flight (thus your need to enter a pressure into your altimeter) the pressure LAPSE RATE ALSO varies on every flight… but they never TELL you that, do they? And since they never correct for the non-standard pressure lapse-rate that we see every day, our altimeters get increasingly in-accurate the higher we climb, since we are climbing into air of increasingly-unknown density, since the lapse-rate WILL vary from day to day.
And, if the ALTIMETER is wrong, isn’t the VERTICAL-SPEED INDICATOR wrong as well, since it clearly will be affected by the pressure change with altitude, since that is exactly what the VSI measures? (The VSI fluctuates based on the rate of change of pressure as you climb or descend! If that rate of change of pressure is different than expected because the pressure change across altitude is different than expected, then the VSI will surely be wrong!)
So, writing my runway-seeker software, I needed to understand whether these pressure-based instruments found some way to be correct, or if they were wrong, and by how much. This would require a high-altitude test-flight in my (non-pressurized) airplane.
So, on a very hot South Carolina afternoon, it was off to 428X, where the tower controller was a bit surprised to see this little propeller airplane file IFR for FL250 (25,000 ft). He simply could not resist asking: “Just how high does that little thing GO?” This is actually a somewhat complex question that I will get to below.
So, before take-off, with minimum fuel to complete the mission on board (minimum fuel to maximize my climb abilities! Remember, the fuel in 428X can weigh a lot, with 108 gallons possible in the tanks!) my pressure-based altimeter (on my G-1000, and backed up by a static-port-based backup) correctly indicated 220 feet… within fifteen feet of the official airport elevation of 236 feet… perfect! As well, the GPS altitude on the MFD indicated 215 feet.. within five feet of the pressure-based altitude! PERFECT! With my IFR plan in place and a take-off clearance, it was full power down the runway in 428X, air conditioning ON to combat the hot, humid, sticky, 95-degree summer afternoon of South Carolina. Craft smartly airborne, I retracted the flaps and quickly set up for a lean-of-peak climb at 16 gallons per hour. Now this is interesting: At full-rich mixture at full-power, 428X burns a staggering FORTY gallons per hour!!! That is the full gas-tank in your car in maybe 20 minutes. For this reason, I always climb at lean of peak, max-lean-cruise power, and burn only 16 gallons per hour, getting maybe half the climb-rate, but the same forward speed and one third the fuel burn! Today, I set my lean-of-peak climb throttle, prop, and mixture, and eased up into the sky at maybe 700 feet per minute, setting up for the half-hour-climb to 25,000 ft to measure the altitudes and vertical speeds, both true and indicated, at every 1,000 ft from sea-level to 25,000 ft. At first, the data was sketchy because of the turbulence and winds of the low-altitude summer day, but I could still detect some alarming trends: Coming through 3,000 ft on the Garmin PFD, the GPS altitude was 3,200 feet!!! The error, at 3,000 ft, between pressure-based and GPS-based altitude was 200 feet! Lest any reader grasp at the imaginary straws of ‘instrument error’ or ‘GPS being in-accurate in altitude’ I can dis-abuse you of that notion right now! 428X is nearly new, it’s instruments carefully calibrated and tested at every annual by a shop that does not know the meaning of the word “skimp”, and the Garmin and backup altimeters were in PERFECT agreement with each other, and were in agreement with the GPS to within FIVE FEET when on the ground. There was no vague ‘instrument error’ or ‘GPS error’ here: Something ELSE was going on. As I continued the climb, the error between the pressure-based and GPS-based altimeters continued to steadily increase. The controllers had me change heading from time to time to avoid airspace as I meandered upwards, and I kept entering the latest and nearest altimeter settings from the nearest-available airport to keep the barometric-pressure setting in my altimeter as perfect as possible. As I broke though 8,000 ft, the air smoothed out to perfection, and my flight-test data became completely un-affected by bouncy air. As I climber, the error between my altimeter and true (GPS) altitude continued to widen. I was surprisingly busy working with the controllers on changing heading, constantly getting the latest altimeter setting from the nearest airport from the XM-weather in the Garmin MFD, putting on oxygen coming through 8,000 ft or so, and recording the indicated and true altitude every 1,000 ft during the climb. In fact, there were a number of times when controllers would have new frequencies or headings for me and I had to delay the readback to them to finish obtaining the latest altimeter setting or recording the latest climb data.
At 5,000 ft on the altimeter, my GPS altitude was 5,270… a 270-foot error.
At 10,000 ft on the altimeter, my GPS altitude was 10,530… a 530-foot error.
At 15,000 ft on the altimeter, my GPS altitude was 15,750… a 750-foot error.
At 20,000 ft on the altimeter, my GPS altitude was 21,150… a 1,150-foot error!
At this point, both myself and 428X were starting to struggle. Even with my oxygen on, I felt dizzy and light-headed. Was my O2 system malfunctioning? The cylinder head temperatures were departing the green and heading into the yellow, the tremendous heat from the twin turbochargers working overtime in the thin air simply too much for the engine to dissipate in the incredibly hot summer afternoon. Suddenly, the engine stuttered out to near-silence for just a moment and then came back… were the magnetos not remaining pressurized? Was the fuel beginning to vaporize? Suddenly a flashing alarm on the G-1000 caught my attention! Why hadn’t I noticed it earlier? Was I truly suffering from hypoxia? A new frequency-request was coming in from the controllers, the alert on the G-1000 indicated that I need to turn on the vapor-suppression, a pump that keeps the fuel in 428X from vaporizing in thin, hot, air before it reaches the engine, I felt dizzy… but my O2 was surely spurting in oxygen… I could HEAR it! The new controller had a new heading for me to keep me away from aircraft departing Charlotte.. I needed to record the next altitude and error… the engine was getting close to over-heat… I was already lowering the nose to try to speed up and cool the engine more, but it was not enough! The CHT’s continued to increase, and the climb-rate was starting to bleed off to near-nothing… how could I complete my test? I needed to cool the engine to get it’s temperatures back into the green, and get more power out of it to leap to 25,000 ft to finish my test, and there was only way to do that: FULL RICH mixture and FULL power. Vapor suppression ON to give a steady flow of non-vaporized, cool, liquid fuel to the engine, and easing the mixture control to full-rich to BATHE the engine in more fuel than it could actually burn, thus leaving the rest to evaporate in the cylinders and cool the engine, I advanced the RPM and throttle to REDLINE, and the engine surged with new vigor and the climb-rate surged smoothly to 1,000 feet per minute as the cylinder temperatures fell back to the green. I was still dizzy and light-headed and confused, but the airplane was surging strongly upwards, and I was still able to record data every 1,000 feet.
At 21,000 ft on the altimeter, my GPS altitude was 22,200… a 1,200-foot error!
At 22,000 ft on the altimeter, my GPS altitude was 23,200… a 1,200-foot error!
At 23,000 ft on the altimeter, my GPS altitude was 24,270… a 1,270-foot error!
At 24,000 ft on the altimeter, my GPS altitude was 25,290… a 1,290-foot error!
Fuel was now dumping overboard at an alarming FORTY gallons per hour, and my remaining 30 gallons was suddenly starting to look sort of thin. I was now using everything the airplane had, and curtailing my future options as the fuel dumped through the engine to give me the power and cooling I needed to power up through the hot, thin, air. I continued to use the best altimeter setting I could find, violating all rules about using 29.92 in the flight levels, because I needed to use the best altimeter setting to get accurate results from the experiment! My clearance to FL250, though, provided the buffer that I needed to avoid an altitude violation. So how high could 428X GO? At lean of peak, it can barely limp to 25,000 ft, but with full-rich mixture and full power it could easily surge well above 25,000 ft, and surely up to 30,000 ft, but the fuel would be exhausting at an alarming rate, and the pilot may be passed out inside. At 24,000 ft, I decided that was as high as was willing to go, since my clearance was to FL250, and since I was using the most accurate altimeter setting I could find, not 29.92, so there would be some error in my altitude compared to what I was assigned! Thus my need for a 1,000 ft buffer in my clearance to avoid an altitude excursion. Temperatures high, fuel low, pulse-rate elevated, O2 saturation low, I requested a descent back to Columbia. Now it was time for the second half of the experiment: Is the VERTICAL SPEED INDICATOR accurate, or is it just as bad as the altitude? The answer soon became obvious: The vertical speed was PERFECTLY tied to the altimeter. I dialed in a 1,000 foot-per-minute descent, and every 1,000 feet on the pressure-based altimeter took EXACTLY one minute to elapse. I am not exagerating when I say that going from 24,000 ft to 4,000 ft on the pressure-based altimeter took EXACTLY 20 minutes… to within about ONE SECOND. Back at 4,000 ft again, I could breathe, the engine could cool, the recorded data could sit on my pad, and it was with some relief that I landed and taxied in.
So, what were the lessons?
The lessons are:
=>Your altimeter is perfectly accurate at the airport if you enter your barometric pressure correctly.
=>At the cruise altitude of about 8,500 feet or so, your altimeter is off by about FIVE-HUNDRED FEET.
=>At that altitude, it does not matter much that your altimeter is off, because everyone else’s is off by the same amount, and you are far from the airport, and thus landing.
=>The error always seems to be off in the safe direction: You are always HIGHER than you think. Clearly, when they designed the system in 50’s, they designed the altimeters to follow the most conservative lapse-rate we could encounter, so in real-world flying, we are almost-always much HIGHER than we thought, which must be the safest error we can make.
=>The vertical speed indicator perfectly tracks the PRESSURE-based altimeter, not the true altitude, so the vertical speed indicator can be perfectly used to see how much the altimeter will change over time, the ACTUAL climb or descent rate of the plane is actually considerably DIFFERENT… at high altitudes, by 1,290 parts in 24,000, or just OVER FIVE PERCENT.
I don’t know how much this matters to you, but as I design the software for the VP-400 A-I runway seeker, all of the altitudes and descent rates will be computed by GPS, and if the altimeter disagrees with what the GPS says, I will know exactly why.