Your VSI Lies Too
Your Vertical Velocity Indicator (VVI), or Vertical Speed Indicator (VSI) if you must call it such, lies to you too. As barometric altitude isn’t linear or tapeline or true altitude, its rate of change cannot be true either. And, as most errors in airspeed measures are related to atmospheric concerns not instrument concerns, any claim that a VVI works differently than an airspeed indicator does not apply. Nor do claims that the instruments are different as functionally internally they’re the same instrument merely connected to different things. This is to say VVI is susceptible to both airspeed differences concerns and also altitude differences concerns. The only thing mitigating VVI is that its lies are smaller due to its working in proximity to zero.
“High to Low, Look Out Below!” It plays. Though now we’re seeing low to high as an issue. Why have you not heard of such before? Why have you been misled to thinking VVI handles the errors? Because typically the errors are small relative to use and VVI is a secondary instrument given less concern. In almost no situation in which VVI lies will the lies impact how you fly. Almost no situation…
I’ve two stories for you:
I was instructing in the T-45C based in Meridian, Mississippi (KNMM). As a Navy jet, we don’t typically flare. At all. Constant AOA from not later than the abeam with increasing VVI around base turn but constant VVI through the groove aka final all the way in to smack the ground. Or ship. If on an approach or straight-in, constant AOA and constant VVI will be established further out.
We were taking several jets to El Centro, California (KNJK) for a training event. I had a student flying from the front while I had the back seat. We landed in Albuquerque (KABQ) on what we might consider a warm day. The student did the normal thing, onspeed AOA with a three degree path. Looked fine to me where I sat, nothing getting my attention. AOA control was solid, indicated speed looked normal for our weight while altitude clicked down as expected, VVI read normally and was properly controlled, far as we could tell path appeared fine. And WHAM!!! We hit hard. Why? Because tVVI is faster than iVVI just as TAS is faster than IAS. A bit of talking around, turns out we weren’t the only one to experience this. All the other jets in our flight experienced it. And others mentioned they’d seen similar things from previous trips.
As so many of us had done exactly this, the likelihood you might think of things like a sudden sinky, for which there was none perceived, or a mismanagement of path or energy on our part are rather reduced. We have a coherent explanation in tVVI while probabilities are against so many having hard hits but only at specific locations. The stats render any idea of poor airwork invalid. Poor understanding and headwork, however… ?
Perhaps we’ve become too sensitive and too precise in our perception of normal and our butts register hard landing with false positives… ?
Bit of both… ? Something other… ?
Granted, in non-Navy flying you may not typically watch VVI on final. Why would you? Glide path or glide slope give you indication for controlling descent, VVI isn’t needed. Even if you make for a constant descent final approach, you’re generally using altitudes against distances from landing giving a computational slope. If you use VVI, it isn’t for its absolute value, it is as a relational instrument. Start going low, what is the VVI, reduce its magnitude. Zero isn’t the VVI mark, its present value is for making changes from current value. You don’t particularly care about current value, just which way to change it with a guess as to how much to change by. It is a bit like making heading changes for crosswind on the localizer while you have no GPS track from which to cheat. But, if you know your groundspeed then VVI kinda becomes like the GPS track for cheating. Yet you do have an alternate cheat that alleviates need for VVI cheat in pitch vs AOA as descent angle matching such to glide slope angle. So, really, to what extent do we use VVI and how?
I think GA pilots either watch VVI more than they recognize or appreciate that they do, or should be watching it more than they do. Maybe you should watch VVI more while on instrument approach, set better initial tip-over conditions. 90 kts on the approach target 450 fpm (or better 480 fpm*). Adjust from there for winds. Use VVI as an initial target, then use range vs altitude to see how you want to adjust going back to VVI to make the adjustment. If you’re not in a sluggish to respond turbine**, use power not pitch for said adjustment (though power will result in minor pitch changes). Piston prop tip-over should mostly be power reduction, perhaps configuration change making drag in lieu of power reduction, and only little nose influence. Then check the VVI as pitch settles in response to the power or drag. Then check the path. It takes time to deviate from path, especially further out such as at tip-over, while VVI should give a sense of which way a deviation is going to go. In this, tVVI doesn’t matter as it is matching TAS while you’re looking at iVVI and IAS respectively. In this we see VVI is not compensated, rather the needs for compensation cancel each other during this particular use. A little faster on the descent matches a little faster on the translation.
* As pilots we like to think we care about precision in our actions. This is false. We have lots of approximations and rules of thumbs and we mistakenly put effort to being precise to these. The reality is we seek good enough so as to preserve mental capacity for whatever may be next. Our effort is or should be at preserving buffers and maintaining capacity. Target VVI is an example. In GA a range to typically shoot is 450 to 500 fpm. No more precision than within these. Why 450? Our rule of thumb is that on a three degree glide path we descend three hundred feet per nautical mile. But our actual descent is closer to 318 ft per nm. 300/1 is 450/1.5. Use 318 and you get 477. We also tend to go into headwind, which will slow our target, so have a range about 480. Now we do train to be precise, but this is actually exercising the brain generating future mental capacity as we develop skills and shift them to “system one.” Striving to be precise makes us work harder which makes our future selves easier at work. Precision is often a training tool not a performance target though it has an often unacknowledged purpose in effect. There are places to be precise but they’re fewer than we admit.
** Turboprops may be sluggish too though typically are not as so. Why? On the one hand, the propeller is inertia, but usually it is set to an rpm hence not a factor. The compressor is still inertia, but typically these are smaller for turboprops. Large turbofans with mass in the fan and mass in the compressor that both need to change rates take time. Early turbojets needed fuel to be changed slowly to avoid compressor stall hence took time.
Admittedly, this next story though the VVI still plays, it is less about the VVI lying and more about airspeed lying. IAS isn’t TAS but turn radius and turn rate depend on TAS not IAS. The two go hand in hand in setting you up as together they make things look normal. Your iVVI vs tVVI still play as because of tVVI being larger you have more force required to arrest your descent in the round out. Yet more-so it is the increased turn radius requirement with increased density altitude that will catch you as you’ll want to start your round out at a point similar to the sight picture you’re used to which is a little late for the increased true speeds.
Now as it happens sometime down the road, I had to pass through Rocky Mountain Jefferson County (KBJC). But I remembered the non-flared landing with high density altitude hitting harder as actual tVVI is bigger than indicated iVVI. So I elected to flare. Well, not really me, I told my student to give me a practice precautionary approach from which we’d flare as we always flare out of those. The student did and flared at a normal sight picture yet we still “fell out,” well, really the jet started to rotate in pitch but the VVI was slow to follow. Before touchdown, likely in his peripheral seeing he was “falling out,” though there wasn’t the typical feel of a fall out, he correctly gave it a shot of power. But the engine being spooled back was slow to kick. We hit hard relative to a flared landing though inside normal non-flare touchdown forces.
Visually from the sides of our eyes, it looked like falling out while it didn’t feel so. Because it was really a bigger turn in the vertical. What do I mean by feel of fall out? The burble be it from the tower of the ship or the tree line messing with quartering headwinds. Losing your headwind or having a gust of tail wind. Leaving a lifty or hitting a sucky. That “stomach drop” feel which is really your body drop then your stomach catching up. Typically you’ll also see such in the periphery as unexpected vertical movement (or lack thereof if you’re pulling the stick seeking your mini-zoom as is a round out). If you flare high and bleed your speed but have just a touch too far to settle for the runway, before the Bam, you feel and possibly see the fall out sensation. It is a feel that tells you to give it a shot of power to cushion.
After these, I incorporated tVVI as yet another concern for high density altitude fields whenever we made trips across. Though I don’t recall and am unaware of anything written about these particular two concerns.
Now a Navy jet landing at a high density altitude field is a corner case. For the rest of us, we’re less concerned with the instrument itself rather picking up the concept of differing tVVIs for a given iVVI and niche impacts as density altitude changes.
What is common in both stories is that density altitude changes change what you need in landing in terms of when and where you round out and flare. High density altitude makes iVVI read less than actual tVVI creating need to start a round out earlier while radius will be larger than expected for indicated airspeed in the round out and force required to arrest tVVI is more.
Alternately, fly a shallower slope. If you fly a shallower slope, tVVI won’t be increased relative to what you’re used to and normal round out should do. Flying shallower also means a higher neutral power point and more responsive engines. Especially with turbines or even turbo-charged pistons. Fly three red on the PAPI. But the longer your fuselage or lower your wheel base the less you can do this.
Some high field elevation fields, with surrounding terrain and obstacles permitting, could incorporate a shallower path by design. And they’d do it for these very tVVI and TAS turn radius reasons. But they don’t as they’re capped by regulation to a shallowest of three degrees. Is 2.5 really a problem?
That’s it. These whole thoughts are ground school type stuff that you wouldn’t generally consider while flying so as to get to the point useful in flying you might want to start a round out a little earlier and maybe give a shot of power to cushion if not used to landing at high density fields. Or, terrain and obstacle permitting, come in a little flat carrying a little more power with an engine more ready to kick in more when sought.
— — — — — — — — — —
The nature of the airspeed instrument matters going from IAS to CAS. But CAS to EAS and EAS to TAS have little to do with the instrument and a lot to do with the atmosphere. This means issues CAS to EAS to TAS apply in the same way to VVI as they do to airspeed. You’ll have iVVI, cVVI, eVVI, tVVI. Granted, as iVVI tends toward zero for most flight, and as it works around zero, the differences will be small. The closer to zero, the smaller the differences. At zero, all are zero. There is reason no one else writes about or discusses VVI errors. Having said this… Do you notice a difference in the instruments?
Just for funsies, let’s do the same with the PHAK.
As an aside, these diagrams are functional. But, despite one coming from the PHAK, I’ve been challenged to the accuracy of the instrument design. The challenger pointed out the restricted orifice in the VVI happens inside the instrument not at the case. For most VVIs, this is true as most have only one static line connection, but it is functionally irrelevant. Where you branch off the static line in parallel to provide the restricted orifice and the free flow bellows has no effect on how the air flows nor on how the instrument works.
Emory Riddle shows it at the case with the instrument having one connection:
If they’re essentially the same design, they are, and the one being airspeed fails to compensate for pressure, temperature, density why would we expect the other being VVI to somehow achieve such?
While mechanized to different magnitudes, these are the same instruments. Other than source line, there is no difference airspeed vs VVI. Any faults the airspeed has the VVI will too.
As VVI has the restricted leak, it actually has other concerns. The restricted orifice creates lag. Lag must either be accepted as is or can be compensated but if compensated, then the compensation will cause turn initiation to read as vertical displacement.
Does the metered leak alleviate VVI of airspeed’s concerns? No because even a perfect instrument will still have them as the concerns arrive from the nature of the air not those of the instrument.
As VVI is a measure in change of barometric altitude, not only do we have speed error concerns, we also have altitude error concerns.
“Imagine doing climb testing on a hot day from 5000 to 6000 feet pressure altitude. You think you are measuring the time to climb 1000 feet, but on this day the 5000 foot pressure altitude contour and the 6000 foot pressure altitude contour are 1075 geopotential feet apart. Because climbing is an exercise in gaining potential energy, the amount of potential energy to be gained is dependent on geopotential feet, not on pressure altitude feet. Therefore, the climb will take longer because more geopotential altitude must be gained.” — Russ Erb, Pitot-Statics and the Standard Atmosphere
Get James McClaran Allen’s stories in your inbox
Join Medium for free to get updates from this writer.
Not only will the altimeter have shown only a thousand feet change, the VVI will have acted accordingly. If the climb took you a minute, you would have seen 1000 fpm not 1075 fpm.
Think about your cold weather modifiers for approaches. If you fail to do them, on the non-precision you will see normal altitudes and normal VVIs despite being low and shallow. You might pick up on it for having a further forward power setting for being shallow, but as both cold and dense helps engines, maybe not. On the precision, you may see higher VVI than expected while on glide slope and you’ll be tight to an unadjusted DH due to being lower further along glide slope when seeing DH.
But the cold case is one of low density altitude. What’s the flip side in high density altitude? Our barometric feet didn’t just shift up, they expanded up. So a foot of real travel may only be a half a foot of barometric meaning the VVI will read slow while altitude will read low on the glide slope. Translate this to a non-precision, if we were to use altimeter and VVI, we’d show low hence go high and steep to obtain normal numbers. We’re really descending faster than normal. A steep puts our neutral power point further aft closer to idle, which for a turbine, or even a turbo-charged piston, has an implication towards sluggish power response. You’re going faster which necessitates quicker corrections while your ability to correct has slowed. You have a larger specific energy loss rate while your means to timely add power and inhibit Es loss are reduced.
There is a counter aspect to the neutral power point as with increased density altitude, we get less power from the engines hence such tends to drive the point forward. In the cold case, we have a shallower path and less tVVI but increased engine capacity may cancel out any forward shift of neutral power point. In the hot case, lack of engine means less reduction for reduced neutral power point for the steeper path increased tVVI, but decreased power output may not fully cancel reduced need.
Take this example from John Denker:
“As an example: Suppose you are flying an instrument approach into Saranac Lake, NY, according to the FAA-approved “Localizer Runway 23” procedure. The airport elevation is 1663 feet. You obtain an altimeter setting from the airport by radio, since you want your altimeter to be as accurate as possible when you reach the runway. You also learn that the surface temperature is −32 Centigrade, which is rather cold but not unheard-of at this location. That means the atmosphere is about 45 C colder than the standard atmosphere. That in turn means the air has shrunk by about 16%. Throughout the approach, you will be too low by an amount that is 16% of your height above the airport.The procedure calls for crossing the outer marker at 3600 MSL and then descending to 2820 MSL, which is the Minimum Descent Altitude. That means that on final approach, you are supposed to be 1157 feet above the airport. If you blindly trust your altimeter, you will be 1157 ‘shrunken feet’ above the airport, which is only about 980 real feet. You will be 180 real feet (210 shrunken feet) lower than you think. To put that number in perspective, remember that localizer approaches are designed to provide only 250 feet of obstacle clearance.”
https://www.av8n.com/how/htm/atmo.html
Now, if you’re sixteen percent lower at all altitudes yet showing normal, your descent is accordingly slower despite what will show as normal VVI. Again, flip this. Consider the hot case. Look to Albuquerque in August. Your percentage is going to be higher, well higher. But we don’t do altitude compensations for approaches because why compensate for being inadvertently high? Though you will still be a consistent percentage higher through your approach while your VVI will read normal meaning your descent will be faster. And this isn’t just a temperature thing, height matters too as sixteen percent of one hundred is only sixteen while sixteen percent of five thousand is eight hundred. (Granted, with cold weather modifiers, they seek to make the field the zero point so errors are smaller as you get closer to field elevation. But we don’t make any sort of corrections in the hot case so we aren’t getting significantly closer to a zero point.)
— — — — — — — — — —
“The instrument is compensated for by pressure and temperature errors so that the correct vertical speed is indicated at any density altitude you may encounter.”
https://www.experimentalaircraft.info/articles/vertical-speed-indicators.php
This a common misperception, perhaps we hold it because generally errors don’t seem significant. Working at and near zero is the biggest factor here. There is no compensation but the closer to zero the less the errors. iVVI,cVVI, eVVI, tVVI are all coincident at zero.
I have found a quiz that must correlate to something justifying how a VVI compensates; it is the only thing I’ve found trying to explain a reason as to how such a compensation might occur:
Hypothetically you’re indicating 130 KIAS and you’re flying along at 17,500 feet. You descend to 16,500 and it takes you two minutes to do so. What was your VVI? Later, you’re down at four thousand though still indicating 130 KIAS. You descend to three thousand in two minutes. What was your VVI? But what were your respective TASs? TAS has nothing to do with it. VVI doesn’t match TAS, it matches changes in barometric feet. But barometric feet aren’t consistent feet. To make this more extreme, imagine the 17.5K case over Yuma in August while the 4K case is over Fairbanks in January.
This seems to mistake the restricted orifice as capillary tube. What is a capillary tube? A very narrow tube that enables a fluid to climb higher than its normal level via friction and tension to the internal surface of the tube. This is what trees do. Where do you typically see a capillary tube? The connection between your engine and oil pressure gauge uses one. What does it do? Allows the transfer of pressure with very little to no motion of fluid. Regarding indicated vs true we are less concerned with the changing of temperature and pressure as we change altitude; these changes are small as we don’t actually change altitude very fast, rather we are concerned with differences in the pressure and temperature in the altitude at which we are at as these differ from true feet.
Let’s consider a moment how such a capillary tube would work in a VVI. If connected to the case side, small diameter would mean slow flow which means slow pressure changes. That is exactly what we want from our restricted orifice. Capillary tube here makes for no changes to our understanding of the instrument. What if such tubing went to the bellows side? Then that too would be slow to change in pressure, you would have no differential pressure, and your instrument would be useless. Further, assume we had these tubes, over time the pressure still changes. The pressures match the environment’s. Nothing is thickening a warm air or thinning a cold air. Further, what kind of capillary action would you expect at forty thousand feet where the atmosphere is a quarter its sea level density? This notion that capillary tubing compensates for temperature and pressure is bullshit. Except Q13, the answers to the quiz are not cases of “most correct” or “least incorrect.” They’re all just plain wrong. Q13 is at times correct and when incorrect is much closer to being correct yet what it suggests is just wrong.
The only coherent explanation for compensation is proximity to zero which makes absolute value of errors small even as percentages remain the same. This means there is no compensation merely acceptance of small errors.
— — — — — — — — — —
“… Taking Aspen for example, my back of the envelope calculation is that, since TAS increases 2% per thousand feet of altitude, a 100 MPH Vref gives a 112 TAS at 6000 MSL. If the usual rate of descent is 500 FPM, then 2% more is 560 FPM. Is he so good he can tell the difference? I know, ASE is 7800' but I wanted to use a mile up. A bigger concern would be the 2% up slope of the runway, which he ignores.”
- “dtuuri” https://www.pilotsofamerica.com/community/threads/angle-or-vx-vy-on-takeoff.154749/page-6#post-3753766 (comment #235)
This will seem to hurt my argument but two percent of five hundred is ten not sixty. Not hurting my argument, looks to me like “2%” is a typo and we’re really looking at 12%. That gives us the sixty difference and matches 100 vs 112. Two percent may not seem that significant, though your credit card company disagrees as it charges merchants that two percent for every swipe. Twelve percent, however, actually is significant and you know it if you don’t pay that card off monthly.
While an extra sixty fpm may not seem consequential and smoothed over as normally most are not actually watching VVI on final rather controlling glide path visually, such is only little larger than half a knot after all, but imagine how your gear interprets hitting at an extra sixty fpm. Flying a tail wheel? An additional sixty fpm descent will bounce you.
Go back to that T-45 which normally lands around 650 indicated fpm with no flare for the normal landing. Twelve percent of 650 is 78. Is this going to feel harder? That puts you at 728 true fpm descent. The normal range for landing VVI is to be within 500 to 700 fpm. If you hit 900 fpm, such is a hard landing that will require maintenance inspection to include jacking the plane and swinging the gear. And you still want margin for being high and steep and for wind gusts. Now imagine you’re heavier, or doing a half flap or worse no flap landing. Still looking for 17 units AOA correlating to faster speeds hence faster VVIs as still on a planned three degree glide slope. Imagine adding a little sucky or a tail wind. Now you’re finding yourself in situations when every little bit counts. Here are the Goshawk Vrefs correlating to onspeed angle of attack.
Let’s take an exploratory detour. Thirteen thousand pounds at 32.2 feet per squared seconds give us 403.7 generic units of mass. As we land, we’re going to stop the extra 78 fpm. Such isn’t instantaneous though it is abrupt. Let’s assume it takes a full second from wheels making first contact through full compression of the shocks and struts. 78 fpm goes to 1.3 fps and as we’re using one second, our acceleration is 1.3 feet per squared seconds. This doesn’t sound like that much more deceleration, but 403.7 times 1.3 gives us 525 pounds. We’re hitting with an extra 525 pounds. That’s not nothing.
— — — — — — — — — —
While the VVI is called a secondary instrument, VVI isn’t really a secondary instrument. It isn’t even tertiary. It really is a bastard red headed step child. And it is treated as such. No one is going to put effort into making it better as such complicates it making it both more expensive and more susceptible to maintenance needs while there is no apparent return on investment for any improvement. Due to that proximity to zero, there really is a diminishing returns situation.
“There are no regulations that require a vertical speed indicator by federal aviation regulations. The vertical speed indicator is not required to meet a calibration standard or even be functional for flight to be legal.” — https://www.cfinotebook.net/notebook/avionics-and-instruments/vertical-speed-indicator
Searching part 23 airworthiness standards, part 43 maintenance, and part 91 general operating and flight rules for the terms ‘VVI’ ‘VSI’ and ‘vertical’ confirms this. Though such could exist in a more obscure location.
“Lag times on the order of nine seconds make it unsuitable as a flight test instrument for anything but trend information. Use altimeter and time hacks to measure rate of climb.” — Russ Erb, USAF TPS pitot-statics instructor
It is not hard to see why no one cares that VVI lies. There is no reason to try to fix it. Any such effort would be better at both speed and altitude first. Yet VVI’s lying isn’t as innocent as we may take it. We have a case where understanding that it lies actually gives us a consideration for adapting our actions. When landing at high density altitude, especially if not used to doing such, start the flare just a little early and consider giving a little shot of power to help cushion. Maybe use the “chip and hold” technique for the round out and flare as opposed to a normal smooth round out and flare. Or, as your fuselage length and wheel base allow, fly a little low shallow approach, three red one white on the PAPI.
“… he’d approach his flare in discreet pieces keeping a ratchet in mind. ‘Chip and Hold, Chip and Hold, Chip and Hold Hold Hold.’ Picture your descent and take a chip out of it shallowing but still descending. Hold then make the next chip. Like a ratchet, this is one-way. Yes, you’ll bleed speed but it will be a slow bleed and you won’t stall. Hold your stick or yoke open handed such that you cannot wrongly push over forward… If you flare too high, extend your hold. If you flare too low, shorten your hold even eliminating it going straight into your second chip. If you get to your third chip and do the holds while still not getting down, you can add some power to cushion so as not to drop, or if short field, Go Around. The Holds of the Chip and Hold give you opportunity to fix flare errors.” — https://medium.com/@jamesmcclaranallen/improve-your-landings-with-aoa-power-techniques-04601584fb3a
