P-38 Rollrate data (with and without aileron boosters)

[DJBscout]

I'm going to draw data from two major sources here. The first is a graph that shows rollrate for a P-38J-25, both with and without boosters.

Spoiler

The second is another chart, that shows the rollrate of various aircraft, including the P-38E (with and without use of rudder)

Spoiler

 

I recently made a suggestion regarding the implementation of multiple blocks of P-38 as modifications, in order to provide more options and create a P-38 variant more suitable for the Battle of Normandy. As I am interested in having as accurate a P-38 as possible, I decided to dig into the data and see what I could find.

 

I transcribed the values of the various tests in excel. Since the test of the J model measures seconds to roll 90°, and the test involving the 38E measures rollrate for both 0-60 (60) and 60-60 (360) degrees (this appears to have been done to give roll measurements that both include and exclude the time required to deflect the ailerons). Thus, some conversion was necessary. I converted the P-38E test to match the P-38J test. (I used the data that did not include the rudder, as this seemed to be the standard testing condition, and when charted against the P-38J had a much closer trend. The Excel file has all data converted and included.)

 

For each speed, I divided 60 by the rollrate for 0-60° to get the time to roll the first 60 degrees, then divided 30 by the rollrate for 60-60° to get the time for the last 30 degrees. I then added these two numbers together to get the overall time required for the P-38E test to roll 90 degrees. I then charted the results (also using best-fit trendlines, which turned out to be polynomial functions of varying degree) on the same graph for comparison. These are the results:

 

So, why is the P-38E seemingly so much better at low speeds than the P-38J?

 

First, let's note that the P-38E is incredibly close to the boost-off P-38J from 250mph on, indicating that when at higher speeds, (where the limiting factor is the high aerodynamic forces preventing full deflection) the two unboosted aircraft have very similar results. This indicates that aside from the boosters, the two aircraft are very similar in their aerodynamics.

 

Now, the E and J were different variants, with some major differences between them. It's important to note the physical differences between the aircraft, and figure out what effects they may have had.

 

The P-38J as tested had outer-wing fuel tanks (twin 55 gallon tanks in the leading edge) and hydraulic systems for the ailerons. Additionally, the intercooler type was changed, from a skin-cooled system in leading edge of the outer wings to a radiator-style pair in the chins of the twin booms. While the new intercoolers may have been slightly bigger and/or heavier (they certainly were draggier), they were also closer to the center of mass, reducing their influence on roll inertia. Between these changes, if tested in standard configuration (~full internal fuel, as most US tests of the time were done), the P-38J would have significantly more weight in the outer wings than the P-38E. (At 110 gallons of total capaciity and avgas' density of 6.01 lb/gal @15°C, that's over 660 lbs of weight in the outer wings!) This would result in the P-38E having significantly less roll inertia than the P-38J, which would be particularly impactful in roll changes at low speed, though little effect at higher speeds.

 

Thus, for any given deflection of the ailerons, the P-38J would be slower to change roll (beginning, ending, or reversing), as the force exerted by the ailerons would be the same, but facing a higher roll inertia. As speed decreases, aileron effectiveness decreases, but roll inertia remains the same. Thus, even if one plane could deflect its ailerons much more easily/quickly, if it had a higher roll inertia, as speed decreased, the plane with the lower roll inertia would still be quicker to roll a given amount. As speed increased, the opposite effect would be observed, with the faster-deflecting surfaces gaining an advantage.

 

In addition, the non-boosted test of the P-38J is denoted as "boost off", not "unboosted". This likely indicates that both these tests were done in a P-38J-25, and the difference between the tests was simply the pilot flicking the aileron boost switch to the "on" position for the "boost on" set of tests. This is important to note, as it means the weight of the hydraulic system was still present in the "boost off" tests, and furthermore that the test pilot would have to not only actuate the ailerons, but also the disengaged hydraulic system. This would increase stick forces.

 

So, what trends are visible in the test data, and what might explain them?

 

Let's begin with the 2 P-38J tests. Even at low speeds (where the ailerons may be fully deflected without mechanical assistance), there is a significant gap between the boost-on and boost-off tests. This can be explained in two main ways: Firstly, when the P-38 pilot did not have to fight a powered-off hydraulic system, and was rather aided by it, the lower stick forces would allow them to (for any given exertion of force) deflect the ailerons more quickly. As speed increased and the ailerons naturally became harder to deflect, a powered-on hydraulic system would help more and more. (you can notice this in the slightly steeper angle of the boost-on line). A non-boosted P-38 in which the pilot did not have to fight a powered-off hydraulic system would be worse than the boosted system, but better than the boost-off 38J.

 

Past 200mph, the gap between the two P-38J tests widens much more rapidly. This is easily explained. Past 200mph, the pilot cannot fully actuate the P-38J's ailerons (and powered-off hydraulics) without assistance. With the hydraulic boost on, there is no (effective) limit to the maximum speed at which the pilot can fully deflect the ailerons. While the boost-off P-38J rollrate does increase for a brief time after the max deflection speed is reached, this is because the force exerted by the ailerons on the passing air is still greater than at lower speeds, even though they cannot deflect fully. However, past 250mph the airflow pressing on the ailerons increases stick forces even further, to the point that the pilot cannot deflect the ailerons enough to continue increasing rollrate.

 

Having compared the two P-38J tests, let us move on to include the P-38E.

 

At very low speeds, the P-38E is the fastest-rolling aircraft, even when compared to the boost-on P-38J. This would be most readily explained by the 38E having lower roll inertia (recall that the P-38J has the extra weight of hydraulics and filled outer wing tanks), meaning that even if its control surfaces were actuated slightly more slowly, the P-38E would still roll more quickly, until the quicker-actuating ailerons of the boosted P-38J take over. The boost-off P-38J is in even worse shape, with not only the higher roll inertia but also the additional stick forces from the powered-off hydraulics potentially causing them to actuate the ailerons even more slowly than the P-38E. So far, it has been assumed that both pilots were exerting equal amounts of forces, but it is also possible that the test pilot in the P-38J was not exerting as much force as the P-38E pilot.

 

At low to medium speeds (between 130 and 250mph), the P-38E roll time gradually moves away from the boost-on 38J and towards the boost-off 38J. The most reasonable explanation seems to be that as the speed increases, roll inertia becomes less important, and the speed/ease with which the ailerons can be fully deflected (which is determined by the amount of stick force required) becomes more important. Thus, while its lower roll inertia gives the P-38E an advantage at low speed, as speed increases, the light controls of the boost-on P-38J begin to work in its favor, giving it the lead at a mere ~130mph. As speed increases further, the lack of aileron boost on the 38E and boost-off 38J becomes more and more the dominant factor, causing the gap between the two aircraft to close as they approach max speed of full deflection.

 

It is at these speeds (250mph+) where the lines for the P-38E and boost-off P-38J rollrates closely align.  However, while close, the P-38J with boosters turned off is slightly slower-rolling than the P-38E, even at these high speeds. This has several possible contributing factors. It may indicate that for any given speed, the boost-off P-38J is deflecting its ailerons slightly less. This would align with slightly higher stick forces caused by the powered-off hydraulic system. Additionally, while higher speeds do minimize the effect of roll inertia (as the faster airflow over the ailerons exerts more force and overcomes the inertia more quickly), it cannot be entirely discounted in any test where the pilot must begin a roll, rather than continue one. (This is exactly why the P-38E tests include both a 0-60 and a 60-60 test, allowing both roll from a standstill and maximum rollrate to be measured). This will give the boost-off 38J a slight disadvantage at any speed, though empty wing tanks would lessen this.

 

So, why have I discussed all of this?

 

Firstly, it should be noted that any time the outer wing tanks are full, there's an extra ~660lbs of weight in the outer wings (55gals*2 tanks*6.01lbs/gal [avgas density]=661.1lbs). If you think about how much removing a couple MGs (~120lbs for a pair of AN/M2s) can help the rollrate of other fighters, imagine how much 660 lbs even further out the wings would make. Thus, if those tanks are left empty, roll inertia in the P-38J will drop precipitously, and and thus low speed rollrate/roll change will greatly improve. In such a case, the P-38J's low-speed rollrate would look more like the P-38E in the graph (nearly identical, I suspect.) This would apply to both the boost-on and boost-off 38s, with both essentially receiving a rollrate buff (I think, I haven't carefully measured 38 rollrate, but it doesn't feel like it's quite as fast as >200deg/s at 400mph). EDIT: Currently, max rollrate at 400mph is (roughly) 140 °/s for a 90 degree roll, a far cry from 205!!

 

Now the main reason for me looking into all this is my suggestion to add the P-38J-15. (or more specifically, make it the default, with the J-25 becoming a mod). The P-38J-15 did not have the aileron boosters at all, and thus simply using the "boost off" data would be incorrect, as the rollrate would be unrealistically poor at all speeds, but especially when slow. Rather, a value closer to the 38E's line would be more appropriate (assuming that the wing tanks are left empty, it might even be slightly better with no intercoolers in the outer wings). Otherwise, the resistance of the hydraulics would be unrealistically modeled on the J-15, and hurt the P-38's low-speed roll.

 

It's also worth noting that at all speeds, some use of rudder was quite effective in aiding both instantaneous and sustained rollrate at any speed above 190mph (see the 38E chart, the one below, or the Excel spreadsheet), but hurt sustained roll above 230. However, a rudder kick was still a huge help to get a roll started at high speed. I am unsure if or how well this is modeled.

Spoiler

 

Hope all this information helps!

 

EDIT: Excel file (in a zipped folder) of the data I transcribed and processed attached. You can double-check my numbers and see everything I did. Particularly, the P-38J data might be off by a few mph/.1s, as I rounded off to the nearest 25 and eyeballed the numbers. (the rollrate graph sadly doesn't work so well, as no trendlines excel makes will fit)

RollCalculations.zip

Edited August 28, 2020 by DJBscout
Excel file

[Gavrick]

On 8/28/2020 at 4:06 AM, DJBscout said:

EDIT: Currently, max rollrate at 400mph is (roughly) 140 °/s for a 90 degree roll, a far cry from 205!!

 

Two more sources:

-Report (P-38J, 1944)

Spoiler

 

-From book "Americas hundred thousand":

Spoiler

 

 

 

 

[DJBscout]

I'm gonna be entirely honest here, I'm a bit confused. That data doesn't match what I've got charted out at all. 

 

I added the chart from "America's hundred thousand" to my Excel doc, and the data is just all over the place in comparison to the other docs I've got. It's similar (but better than) the boost-off P-38J chart (though seems to have similar levels of roll inertia at low speeds), but is much later to reverse, and stronger until it does.

 

In short, I'm confused, and in all honesty your two docs do seem to agree fairly nicely with one another. That being said, America's Hundred Thousand fails to specify whether it's measuring instantaneous or max rollrate, and the Air Force document also specifies the boosted rollrate to be estimated, and completely untested.

Edited August 31, 2020 by DJBscout

[69th_chuter]

The unboosted ailerons traveled 25° up and 20° down inducing little adverse yaw.  The boosted ailerons traveled 25° up and 25° down inducing some adverse yaw (rolling right swings nose left) but increasing overall roll performance across the envelope.

[DJBscout]

I saw this post was referenced elsewhere, so I wanted to add some information and clear some things up.

First off, I tried to analyze just about everything I could, breaking down all the information I had, and then extrapolating what might account for differences. (For example, a peak on an unboosted aircraft shows approximately where aileron deflection ends, or can show a difference in stick force when testing the same airframe.)  I thought that, like a jigsaw puzzle, if I spent enough time looking at differences, I could figure out what the deal was. This was....false, to say the least. When trying to do calculations like this, you need to take into account:

• model of aircraft
• wing tank conditions (full or empty?)
• stick force (also how quickly the yoke was thrown to full deflection)
• rudder vs no rudder
• level of aileron deflection
• altitude
• TAS vs IAS (fun fact, roll is a function of TAS and not IAS)
• boosted vs unboosted ailerons
• type of roll measured (0-60 will be slower than 0-90, which will itself be slower than 60-60, which might be slower than a max rollrate even if only very slightly)

Now, a lot of the data I looked at would show a fair number of these, but lack several others. Given the differences and gaps in the data, I was never able to come to any conclusive results, try as I might. Even one of these factors can introduce massive differences into what the graph looks like.

Here's what roll plotted against TAS and IAS looks like for a P-38L (Bonus altitude comparison too):

Spoiler

(This chart started in TAS, so I had to convert to IAS. Assuming standard atmospheric conditions, the 38's IAS at the lower end would be stalling (or on the verge of a stall) in a clean configuration. Even some of the best data can have you scratching your head.)

Here's the difference between 0-60, 0-90, and 60-60 rollrate will be for a P-38E with no boosters

Spoiler

 

(0-90 was not included in the E test data, so it was calculated by taking the 0-60 rate and dividing by seconds to get the time for a 0-60 roll, then multiplying the 60-60 rate by 30 degrees, and adding the two figures to get the 0-90 time, which was then divided by 90 to get the 0-90°/s)

Here's the difference between the E's instantaneous and maximum rollrates with and without rudder:

Spoiler

Notice how there's crossover where the rudder eventually becomes beneficial. (The lack of a benefit to using rudder at low speeds is honestly confusing, as that's where it should in theory be especially helpful in overcoming roll inertia for the 0-60 tests. If I had to guess, I'd say there was no rudder kick below ~190mph.) 

In some cases, a rudder kick will be better than either no rudder or constant rudder:

Now, the difference between boosters and no boosters is....complicated. As the shrewd of you may have figured out, that's the whole point of this thing, right? Sadly, there are only two examples of comparisons between boosted and unboosted sources. The first is the original doc I produced, and the second is the data Gavrick produced. These documents don't really agree with each other, to say the least.

Spoiler

This is the kind of gap we're looking at. If I had to GUESS, I would say the yellow line is an absolute best-case scenario, with the orange being a worst-case. Reminder that this is the source document:

Now, if the pilot was kicking the rudder for the boosted example and had empty wing tanks, that could make a huge difference.

That being said, the high-speed range of the boost-off test is almost equal to the 38E 90° numbers, but the low-speed is abysmal by comparison. Meanwhile, the boosted numbers outclass literally everything else. It's simply bewildering.

The real problem here is that there's too much data, and not enough. The data is all over the place, but we don't have enough information about testing conditions to figure out why the numbers varied so much, and where the in-game model should fall. (How all over the place is the data? Look at the chart below if you dare.)

Spoiler

 

Meanwhile, other aircraft with scant sourcing pretty much just get whatever the single available source says. That can result in aircraft that are far too strong, or too weak. Either way, this is about all I could find on the 38. If anyone else thinks they can decipher everything, be my guest. Attached below is a copy of the Excel file I used to crunch all the numbers.

RollCalculations.zip

[TP_Sparky]

Isn't the 60-60 roll rate based on 60 degrees bank to vertical to 60 degrees in the other direction for 120 degrees of roll and not a complete 360 degree roll as in the message above?