What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Residual Thrust Effect on Glide Performance

Vac

Well Known Member
Benefactor
We've been conducting some emergency turnback experiments to support an EAA/FAA effort to develop planning tools and techniques for dealing with power loss during takeoff.

Foremost, the point of this post isn't to advocate for any particular course of action if the engine quits on takeoff--it's just to share data and lessons learned :)

My RV-4 is equipped with a light-weight, two-blade Catto propeller and is quite the glider in IDLE power. Some of that "glide" performance is the result of residual thrust provided by the engine in IDLE. If the engine is OFF (mixture in cut-off), glide performance suffers (by about 20% at the same AOA/speed). This is shown in the first chart for glides at L/Dmax and ONSPEED with the engine in IDLE and OFF:

3c039a_9bcb3431183e4b9dbb436e166a5c9cbb~mv2.png


In the RV-4 with it's O-320/Catto prop combination, the propeller windmills at 875-900 RPM at L/Dmax with the engine OFF. Since the propeller is now driving a four-cylinder compressor, this windmilling prop adds drag and increases glide angle. Incidentally, with a fixed pitch prop and the engine OFF, lowest windmilling RPM is obtained with the throttle in IDLE (butterfly valve closed). Assuming oil pressure is available, the lowest RPM for a controllable prop occurs in coarse pitch (low RPM). If I slow the RV-4 to an ONSPEED condition (1.3 Vs), the prop stops, reducing drag at that AOA.

Since most of us aren't keen on pulling the mixture at low altitude for aircraft handling practice, it became necessary to develop a technique to replicate engine OFF performance in IDLE. Not too many options with the RV-4; but turns out deploying flaps 40 fairly well replicates flaps 0 engine OFF glide performance:

3c039a_74265cfea7cf4209b68b36b446f89355~mv2.png


The most important thing to compare is glide angle. You can see in the second chart, wing's level glide angle is a wash between IDLE/Flaps 40 and OFF/Flaps 0. Glide angle in a stabilized 45 degree banked gliding turn is a bit steeper IDLE/Flaps 40 than OFF/Flaps 0, but the difference isn't significant due the nature of the turns--with roll in, roll out and the alignment maneuver the airplane never truly stabilizes in the banked gliding condition. Also, for the test point at altitude (IDLE/Flaps 40), alpha was too high (in other words, I flew it too slow; so I'll have another chance to excel soon and will correct the table with new data if there is a change).

This "high drag" technique for dealing with residual thrust can be applied for practice or flight test: https://youtu.be/aIvBnva1At8

Incidentally, in the RV-4 the "high drag" configuration adds 20% to the minimum altitude required for turnback with a 3-5 second simulated startle delay. Winds were calm in the video. We need "no wind" data to build an accurate MatLab model. Eventually, we'll be able to use the model to "what if" various scenarios by adjusting conditions. In addition to flaps, it's also necessary to fly slower to replicate glide angle. In the video you'll hear me reference the "slow tone." That is a condition slower than ONSPEED (i.e., less than 1.3 Vs), but faster than stall warning (set to an AOA that provides 5 knots warning in my airplane).

The bottom line is if you practice and determine a "minimum turnback altitude" using IDLE power, things will be different when the engine actually packs it in. Assuming the data from the -4 extrapolates, using Flaps 30 (RV-9/10/14) or Flaps 40 (RV-3/4/6/7/8) will likely provide a better sight picture for practice.

And last thing to consider is that I'm able to "max perform" the airplane because I can listen to my alpha/energy and I don't have a requirement to look inside the cockpit when I maneuver close to the ground. I've never flown a maneuver like this looking at an airspeed indicator, because my brain isn't fast enough to adjust stall speed based on the square root of my G load and I'm not that brave as the two-seat RV's don't have good buffet cues at low G. One thing to consider is that a 45 degree banked gliding turn is still a 1.3-1.4 G maneuver and increases indicated stall speed by about 15%. I would be disinclined to conduct this type of practice without an accurate AOA system that allows me to know precisely what my aerodynamic margin (difference between my actual AOA and stall) is. Be careful out there.

Fly Safe,

Vac

P.S. Skylor, finally caught up with you :D
 
Last edited:
Good Stuff

Hey Vac,

Glad to finally see the no power data! Good on you.

Actual power out glide #s are perhaps one of the most important Safety considerations to the single engine pilot, and one area that some people don't feel comfortable exploring. Folks that have turn back altitudes based on idle power testing may have overly optimistic ideas.

Every aircraft/engine/prop/pilot combo will yield different results and the important thing is valid/usable numbers. If testing this area of your performance isn't your thing then conservative numbers are probably appropriate. For our RVs, conservative turn back altitudes may be in the 750-1000' range.

If you have a CS prop you may or may not be able to control it in a 'no power' glide, or may lose control in the final phase of an approach as you slow down suddenly affecting residual glide distance. Do CS prop folks need to be ready with two sets of numbers in case of an oil loss?

So how do we practice our 'turn backs' if that's what we want to do?

For my CS prop/gov/engine combo I know that idle power and 1450 RPM is a good simulation of actual 'no power' in term of alt loss/ROD. Only way to know that is to go out and find those actual 'no power' numbers and figure out how to simulate them at idle.

For those truly out to find performance consider what MAP does to your ROD in no power glides. MAP is still under your control in power out scenarios and depending on your gear may make a meaningful difference. In my aircraft getting the MAP up around 15-17" yields the best glide.
 
Last edited:
The most important thing to compare is glide angle. You can see in the second chart, wing's level glide angle is a wash between IDLE/Flaps 40 and IDLE/Flaps 0. D

Vac, thanks for the data. I do think there’s a serious typo, quoted above, which cannot possibly be correct. And that data isn’t even in the table. I think you meant ‘power off/flaps 0’, not ‘idle/flaps 0’.
 
Bob,

You are correct--did it twice, actually. Poor attention to detail on my part--no excuse. Corrected.

Also added a note about the IDLE/Flaps 40 spiral glide test. After chatting with one of the PhD's, suspect that had I flown it just a bit faster, glide angles would have married up better. I'll re-run that test next opportunity at a slightly lower (faster) AOA and make any corrections to the table. After the low-altitude drills, I still stand by my assessment that it's a wash do to the dynamic nature of the maneuver vs the stable, developed spiral at altitude.

Having done a bunch of tests previously at IDLE at low flap settings, definitely a different sight picture in "high drag" configuration!

v/r,

Vac
 
Thanks Mike !

I tested my RV7 glide with mixture out, thanks for this high quality data.
 
Last edited:
I am sure this is not what you meant . . .

. . . . In my aircraft getting the MAP up around 15-17" yields the best glide.

This just struck my funny bone, sorry . . . I thought; hmmmm, at 16" MAP I can guide at 125ktas for about 6 hours.

To your exact point though - - while a running engine may be rotated at some zero thrust point, if it is truly producing no power, are you adding throttle (thus MAP of 16") to result in less loss of altitude? If we think control volume and any energy used to rotate the engine is essentially drag. Why wouldn't the lowest friction be at the lowest RPM and MAP? I think I'll have to go test this.
 
Last edited:
This is very interesting and useful. I'm looking forward to the day when I get to apply it to my aircraft.

Just a suggestion, during the turning tests a yaw string might be useful, since for some aircraft the rate of descent increases a lot at non-optimum sideslip angles. Perhaps you could "calibrate" it against pedal input during steady heading sideslips and then see during the turns just how much sideslip inaccuracy is acceptable before it significantly affects power-off descent rates at onspeed.

Maybe you've already looked at this and found it wasn't significant? If not, it's not a big effort required and it might add some useful info.

Thanks for your work.
 
This just struck my funny bone, sorry . . . I thought; hmmmm, at 16" MAP I can guide at 125ktas for about 6 hours.

To your exact point though - - while a running engine may be rotated at some zero thrust point, if it is truly producing no power, are you adding throttle (thus MAP of 16") to result in less loss of altitude? If we think control volume and any energy used to rotate the engine is essentially drag. Why wouldn't the lowest friction be at the lowest RPM and MAP? I think I'll have to go test this.

The pumping losses are proportional to the volume but also to the pressure difference. Pulling that air past a closed throttle creates significant pressure losses, as anyone with a MP gauge knows. It makes sense that minimum energy loss will be somewhere between closed throttle (high pressure difference, minimum volume) and wide open throttle (minimum pressure difference, but large volume).
 
This just struck my funny bone, sorry . . . I thought; hmmmm, at 16" MAP I can guide at 125ktas for about 6 hours.

To your exact point though - - while a running engine may be rotated at some zero thrust point, if it is truly producing no power, are you adding throttle (thus MAP of 16") to result in less loss of altitude? If we think control volume and any energy used to rotate the engine is essentially drag. Why wouldn't the lowest friction be at the lowest RPM and MAP? I think I'll have to go test this.

That's really the point, go see what your gear does.

With zero fuel flow (Mixture to zero fuel) and at best glide configuration I moved the throttle position to see if MAP was a measurable variable on glide performance. In my case I saw an immediate RPM rise with increasing MAP. It's a bit counterintuitive, but after crunching the numbers I determined I was observing the engine robbing less power and was able to observe better glide range.

I wouldn't be so bold as to offer that will be the case for everyone, but it is a control available to all of us so perhaps it's worth knowing its impact.
 
Last edited:
Really interesting, and thanks.
I'd be interested in engine off (mixture idle cut off) with prop windmilling, compared to engine off prop stopped for your fixed pitch prop.
 
Mike - great stuff, thanks.

Given that it takes more power to idle a fixed pitch vs a constant speed prop in static conditions (i.e., those used to set the idle), would not the difference between idle power glide and no power glide for a C/S prop plane (with the prop remaining set at high rpm) be notably less? That has been my experience when I've pulled the mixture during simulated engine out glides.
 
The pumping losses are proportional to the volume but also to the pressure difference. Pulling that air past a closed throttle creates significant pressure losses, as anyone with a MP gauge knows. It makes sense that minimum energy loss will be somewhere between closed throttle (high pressure difference, minimum volume) and wide open throttle (minimum pressure difference, but large volume).

Perfectly stated Bob, but still not obvious that the decrease in pressure drop losses at low mass flow are higher than high mass flow pumping losses at a lower pressure drop.

Engine performance analysis was my day job for a decade, but all my work was determining pumping losses in running engines with restrictions for intake/exhaust and turbo systems with after coolers.
 
That's really the point, go see what your gear does.

VERY well said!!

I appreciate the additional data Vac.

The thing I think people often forget is that you cannot know what 'state' your gear is going to be in when you are "forced" to make that turn back decision. Your prop may be stopped, it may be windmilling, it may have slightly more than idle power, it might not even be attached the airplane any longer. Shucks... you might have the compression of only two or three cylinders. You just don't know. IMO, finding out what your gear will do in SOME scenario is the only approach that seems reasonable. You can then make adjustments and pick a number that is very likely to be adequate.
 
JD,

Yup :D

Dan,

We actually record yaw on both the reference gyro and the "YAPS" (yaw, alpha, pressure system) boom fitted to the wing-tip of the RV-4:

3c039a_2ae427b9c1d84c7eaa8da7a135665ad3~mv2.jpg


In the pic, you can see the alpha (AOA) and beta (yaw) vanes on the boom. The boom also has a Keil probe that measures dynamic and static pressure independent of the of the aircraft pitot/static system for reference purposes. We also measure yaw with the VN-300 reference gyro in the airplane that is much more accurate than an EFIS. All of this data is recorded on the 32GB SD card in the ONSPEED system.

The reason we added data recording capability to the system is to allow folks to easily record data during experimental test flight (Phase 1, for example). Data is recorded in a .csv file that you can download post-flight via wifi. The system integrates data from your EFIS as well (if you have one).

John,

We didn't make an intentional effort to stop the prop, since below 3000' AGL CAFE data show that the pilot shouldn't try to intentionally stop the prop to improve drag characteristics. If you have a lightweight, fixed pitch prop it will simply stop as you approach ONSPEED/Vref. Metal fixed pitch propellers have a bit more inertia; but will also stop as you approach stall speed. Stopping a constant speed prop (especially one with metal blades) can require sustained high alpha (i.e., stall), thus the 3000' AGL rule. Our thinking is the prop will simply do what it does in each airplane, and other than trying to minimize the drag on a constant speed prop (by setting low RPM assuming it’s controllable and no restart is practical), the pilot should simply focus on maintaining aircraft control and landing the airplane, wherever that may occur.

Interestingly, in analysis, we know the drag effect of the prop stopped and the prop windmilling at various RPM and can adjust for this in the MatLab model.

Alex,

The engine/prop combo is the key. Unfortunately, it's the ultimate "it depends" answer--hence Duck's admonishment that you simply must test every airplane to determine performance. Your glide performance "is what it is" at high and low RPM (assuming you can control blade angle). In the event of a catastrophic failure and oil loss, it may also simply stop turning. It is possible to test all these various conditions at altitude. The last case requires a prop stopped glide. Alternatively, you could also simply assume "worst case" scenario: prop windmilling at high RPM for test and training.

v/r,

Vac
 
Thoughts on the conduct of the testr

Thank you for the the thought provoking study and data you have collected. You have pushed me personally to expand my thoughts on engine failures after take off. I have never thought about the drag created by the prop itself and have always trained with the idle thrust. The advantage idle thrust gives is huge and can be the difference between making it to a safe location or not.

I do have one question and am interested in your thoughts. after I watched your YouTube video where you conducted the turns back to the airport I noticed something interesting in your sequence of actions. What I saw was
-power idle
-3 second pucker moment
-full flaps (to simulate added drags from prop not in idle)

I’ve been thinking about the test more and think maybe a better version of sequence to simulate a real failure would be
-power idle
-full flaps
-3 second pucker factor

I only propose this to change because I would think the added drags from the power not at idle would more accurately simulate an engine failure.

Now I understand that this may raise the alt for your decision to turn back. I only propose this because I always want to plan for the worst. I was blown away by the difference in power idle and mixture cut off. What are your thoughts on the impact of changing the flaps movement for training and data collection?

Respectfully,
Evan
 
Evan,

Great thought!

The reason for the sequence is simple: Armstrong flaps. I guess I need to spend more time in the gym :eek:

I'm climbing at V4 (L/Dmax), so air loads for flaps 40 are pretty high and decrease during the "startle time."

If I had electric flaps, you bet: IDLE, flaps, delay, turn would be an optimum sequence.

I suspect overall, minimal effect based on the dynamic nature of the event, time required for flap deployment and overall variability; but that's just an informed hypothesis. At least that's my story and I'm sticking to it ;)

v/r,

Vac
 
Last edited:
The final configuration which nobody would like to test, is when a crankshaft fails and you have a truly windmilling propellor (that has no backpressure from pushing the cylinders around). That will be the highest drag configuration and the worst case scenario (for many reasons).
 
Mike,

I believe your hypothesis is correct that it has no anal effect. I haven’t built my RV yet so the difference in manual and electric flaps didn’t even occur to me. Thank you again for the data you’ve collected. I’ve been a CFII for a couple years and your research has provoked me to think deeper about how emergencies at accomplished. Thank you again!

Evan
 
This effort is VERY worthwhile. So my comments are only intended to supplement what is above.

1. Many years ago a friend and I explored the optimum bank angle for a turn back to the runway. We started with the math and then verified it in the real world. I no longer remember all the details but I do remember that the best ratio of sink to turn was nearer 60 degrees of bank, not 45. It makes a big difference.

2. The G forces for 45 and 60 degrees of bank in a LEVEL turn are 1.4 and 2.0. However, the case we are discussing is not a level turn and if you try it you can see that the G forces if you allow it to sink are much, much lower without diminishing the turn rate. This can be very important to a correct understanding of the issue.

3. If you did have a broken crankshaft then the pumping work would be less than with a windmilling prop on an intact engine. I believe this contradicts one post above.

4. While it is not relevant here, the correct test for zero thrust was demonstrated by Jack Norris many years ago. If anyone is interested, his 2 books are available free at www.propellersexplained.com. You can PM me if this link does not work for you. There is a lot more there, of course. There is also an old Sport Aviation (EAA) pair of articles about his experiments in the CAFE group. Page 84, August 1995. Jack found a way to re-do the experiment without removing the prop, later. The EAA article was a C-152 but Jack used his own Luscombe. His article was March of 1995, also in EAA's Sport Aviation.

5. The dead engine will produce drag (AKA negative thrust) whether the prop is windmilling or stopped although the amount will differ. A CS prop will offer additional complexity so simply experimenting (at a safe altitude) is the right approach.

6. A turn back to the runway is more complicated because of many factors. It's difficult and it's risky. I do not mean to say otherwise. Let's be careful out there.
 
Last edited:
Back
Top