Mitsubishi A6M2 Model 21 and A6M3 Model 22

I

Ivan

Charter Member
Status Update on A6M2 Rei Shiki Sentoki Ni Ichi Gata.
(or a Tale of Two Zeros)

A week ago, I thought the A6M2 Zero was very close to done. It was missing a proper control panel, but the Hurricane panel wasn't too far off.

Since then....
I noticed that the engine output didn't drop even when the RPM was reduced by 150 or so.
I figured I had a pretty good idea what was causing this so went ahead to change it.
The change was to level out the Engine Efficiency to RPM curve which needed a corresponding change to the Torque Friction Loss per Cylinder to RPM curve. (Records 508 and 509.)

In playing with this, I found that although the power up to the critical altitude (12,500 feet) remained the same, the power at altitude was significantly reduced.
This was unacceptable because the absolute ceiling was already too low and less power would reduce the ceiling even further. The testing of the Aleutian Zero gave a ceiling of 38500 feet. I was getting more like 36,000 in testing.

The solution wasn't very difficult. It was basically to reduce the Engine efficiency and to also reduce the Friction loss. This needed a couple hours of testing to get an acceptable level of power at each altitude.

My original Zero was the A6M5 which I built and released about 10 years ago. Testing and research back then was much less than I do today. One of the things I have found since then was that the sea level speed of the Zero in most models was very close to 300 mph. The earlier flight model was limited to 270 mph at 500 feet.

One of the issues with the actual Japanese Zero was that its maximum speed in a dive was limited by structural strength. The maximum diving speed never exceeded 460 mph. We can't reproduce structural failures due to aerodynamic loads in CFS, so there needs to be something else to discourage excessive speed.
At the time, I found that I could control the directional and longitudinal oscillations of the aircraft quite easily and predictably and set the vibrations to start at around 285 mph TAS at 500 feet. This is only accurate to around 5-10 mph, so sometimes the vibration might start at 280 mph.

Since the new flight model was capable of around 290-295 mph at 500 feet, the vibrations needed to be tuned again. This needed a fair amount of looking around to find the notes I had used earlier. (Most recent use was around 2008.) To test this, I needed a way to push the aeroplane a bit faster. The A6M2 normally does not have WEP but for this test, gained about 10 inches extra boost and about 275 more horsepower to confirm that the new values I entered would not cause a vibration until 308 mph. This didn't seem to be enough so the next change put the speed limit up at 315 mph again at 500 feet. Under normal use, the virtual pilot should never see that this limit even exists.

The other speed limiting change which I now do for all new aircraft is to adjust the Mach Drag curve to put a "wall" just under Mach 1 which tapers back to adjust drag upwards at lower speeds to limit the maximum speed in a full power vertical dive. For this Zero, the speed limit is around 550 mph True Air Speed in a vertical dive from 30,000 feet. I didn't bother trying to convert this to Indicated Air Speed because I was lazy.

My protocol for testing level speeds is to run at full power with autopilot controlling altitude and heading.
As soon as 15 seconds pass without an increase, I note the current reading as maximum speed.
This may be 2-3 mph below where it finally stabilises, but is consistent and also compensates for an autopilot being able to hold much steadier than a human pilot.

Please respond with methods you all use to do this.

Speed @ 500 ft = 294 mph
Speed @ 12500 ft = 336 mph
Speed @ 15000 ft = 331 mph

I believe at this point, this makes a pretty good starting point for a flight model.

- Ivan.
 
Last edited:
The last post described recent changes and evolutions to the flight model.
This post describes the evolution of the visual model.

The first Japanese Zero I built was the A6M5 Model 52a which was first released around 2002. It was based on drawings from a book by William Green. The idea of dimensioned drawings did not occur to me at the time.

Up until a couple days ago, I was quite proud of myself for having "completed" the visual for the A6M2.
THEN I noticed that the flap animation wasn't very good. I also noticed that I had goofed a bit in the animation for the propeller as well. THIS I found out AFTER I had finished SCASMing the aircraft.

I figured as an educational experience to show my son what I was working on. HE commented that the closed landing gear doors could be seen even though the gear was down. (I never put in the wheel wells.)
My current task is creating the wheel wells, but I also noticed that the inner and outter gear doors don't match up very well and the animation doesn't line things up very well which means there is plenty more work....

- Ivan.
 
Last edited:
Adding the wheel wells for my son gave the result in this screenshot.
I also found that each of the stabilisers can lose 4 polygons without hurting the shape and that was sufficient to cure the AF99 error.

Note that the colour is a simple gray because of the trip through AF99.

- Ivan.
 
Finished adding wheel wells. Spent a couple hours getting the animation of gear doors to match the wheel wells. Aircraft Animator doesn't offer much flexibility in sequencing, so adjustments should be made via SCASM.

- Ivan.
 
Pain in the SCASM

When going back to fix the texturing on the inner gear doors, I found that the change in shape had pushed the rear edge past the edge of the texture. Fixing this kind of thing in AF99 would take just a few minutes. Using SCASM, this took me a couple hours find the new values and modify the source code.

Here is the result with the textures relocated (the hard part) and the updated BMP textures. I'm not sure the visual improvement was worth the effort, but the problem bothered me while it was there.

The other screenshots show a very low pass that resulted in scraping some paint off the wing tip but not crashing.

- Ivan.
 
Sorta Flight Testing

This was pretty much a test of Scrape Points.

Note that on the belly landing screenshot the plane is sitting a few inches lower than expected.
There is nothing wrong with the location of at least the front scrape point as can be seen in the other screenshots.
Apparently if the Fore and Aft scrape points are too far apart, the aircraft sits noticeably Lower.
I am not quite sure what causes it, but I have seen it before.
 
Aircraft Handling Qualities

In addition to the straight line performance, there is also the general handling, stability, and control response than can be tuned.
I believe this makes more of an impression on the 1G Virtual pilot than the straight line performance.
I don't really know how to affect things here, but after lots of tries, I figure I had better quit before I really mess things up.
The end result is that the Zero points very well as can be seen in these screenshots but unfortunately does not have as much longitudinal stability as it should. It is not easy to trim for hands-off flying and tends to have a long period oscillation (Phugoid).

Although the screenshots can't show this, on a low level run at about 250-270 mph, it will fly the length of my test runway hands off without gaining more than a couple feet of altitude.

- Ivan.
 
A6M Control Panel

Here is a first cut at a Control Panel for the A6M2. Screenshot was taken on its first "Test Flight".

It uses all stock CFS gauges. While it is functional, there are a whole bunch of problems.
Many of the gauge ranges (such as for Manifold Pressure) are a bit off for this aircraft.
Perhaps I can find a better substitute gauge.

The Japanese measured manifold pressure in mm Hg above 760 (ambient pressure).
A typical throttle setting would be +200 mm or so. Obviously there isn't an equivalent in US, German or British aircraft.

The background was created from a heavily edited screenshot of a virtual cockpit model for this aircraft.
As expected, it doesn't quite line up with the rest of the aircraft, but I believe it is a fair compromise between what is needed for CFS controls and what the real A6M2 cockpit looks like.

While I believe the virtual cockpit model looks very cool, it cannot appear in the final model because of the heavy use of resources.

- Ivan.
 
This is an updated version of the control panel for the A6M2.

The biggest change is the manifold pressure gauge. A German gauge reading in Ata is now used.
There are also minor adjustments in sizing and location and a small placard under the Compass.
Most of the gauges were also adjusted so that they are fairly close to round.

The control panel for the A6M is not easy to represent for CFS:
Most of the gauges are below the cockpit sill because the space above them is taken by a gunsight and the breeches of the cowl machine guns. This would put them below the area of a CFS control panel.

The choices are to zoom out so that the control panel is smaller on the screen, but that would also mean that the gauges are much smaller and probably unreadable or to distort the panel as was done for the CFS 2 Zeros. I chose to do both but hopefully in a reasonable manner.

The actual controls and gauges on the A6M sometimes have no equivalent in CFS. There are also a fair number of controls on the consoles on each side of the cockpit.

The space occupied by the fuel quantity gauge (FW 190A) on this panel should actually have a exhaust gas temperature gauge. It was used to fine tune the mixture for increased range. The actual A6M fuel gauge does not read all of the fuel in the tanks. It only gives useful readings when there is about 20 gallons remaining.

The space occupied by the fuel selector gauge (Hurricane Mk.I) on this panel should be a signal strength readout for the DF loop that is typically installed behind the pylon behind the cockpit. It wouldn't work on this aeroplane anyway since there isn't a radio direction finder loop installed in the model. ;-)

The typical Japanese practice with symmetrical wing tanks was to select them in pairs as a single fuel tank. I don't know if this applied to the A6M but am treating them as separate tanks.

My initial choice of engine controls was to use the set from the stock Hurricane Mk.I. While the appearance was not bad, I found that the mixture on that gauge set could not be set to idle-cutoff. Keyboard controls were needed for that, so these controls are from the Spitfire Mk.I.

I believe this is a fair though unsophisticated representation of a fairly functional panel for the A6M series with gauges as correctly located as can be done for CFS.

Comments?

- Ivan.
 
By the way,

The Oil Temperature is within the normal operating range as stated in the manual.
The Cylinder Head Temperature is typically a bit under the normal operating range according to the manual, but above the minimum. Perhaps this changes a bit under sustained WOT operation at low air speeds such as in a climb.

The fuel and oil pressure readings are given in the manual but the unit of measure is not stated so I don't know if they are correct. They stay in the middle range of the gauge being used, so I will not be adjusting unless I can confirm they are incorrect. Any additional information would be appreciated.

- Ivan.
 
A6M2 Model 21 is Completed

I always try an installation of the final ZIP file on a home computer to make sure things work as I expect before uploading.

In this case, my son was showing an interest, so I did some of the testing in front of him.
I was showing him the displays for the oil temperature and cylinder head temperature and how they were rising as I sat on the runway with brakes on and engine turing at 1000 RPM.
He asked where the cylinders were and in order to show him the cowl opening, I had to change to an external view.
We changed the view back into the cockpit, but apparently the keyboard activity reset the brakes.
As the plane started moving forward, I applied the brakes quickly.
The plane tipped forward and sat on its nose which unintentionally satisfied yet one more test.

After that, we took it up for a short hop and landed without incident though I bounced it a bit.

In any event, the project is finally done and released. If you all download and find a bug, please let me know.

- Ivan.
 
at the risk of sounding like a braggart,
i have had the pre-release version
of this aircraft for a few weeks now.

granted, this is only my opinion,
but, i've got to tell you all,
this little sweetheart is a peach.
...and there is no pit included.

as far as i'm concerned, this aircraft
ranks at the top of the cfs aircraft list.
but, don't take my word for it.
get off your duff and go get your own.
 
Thanks Smilo.

Here is a quick poll:

The A6M2 could officially carry a pair of 66 kg bombs. That is a total of about 265 pounds.
It could also carry a 300 or 330 liter drop tank.
The weight of the fuel in the drop tank would weigh around 523 pounds.
Later in the war, the A6M2 would typically carry a single 550 pound bomb.
Should the DP file allow a pair of 132 pound bombs (current setting) or a single 550 pound bomb?

Opinions?

- Ivan.
 
Ivan said:
Thanks Smilo.

Here is a quick poll:

The A6M2 could officially carry a pair of 66 kg bombs. That is a total of about 265 pounds.
It could also carry a 300 or 330 liter drop tank.
The weight of the fuel in the drop tank would weigh around 523 pounds.
Later in the war, the A6M2 would typically carry a single 550 pound bomb.
Should the DP file allow a pair of 132 pound bombs (current setting) or a single 550 pound bomb?

Opinions?

- Ivan.
Click to expand...

One single bomb for me, and that for these reasons;

1- Any bomb in CFS1 has the same explosive power, whatever you did in the DP file.
2- The general purpose bomb was ±500 lbs on Axis and Allied side. The 250kg bomb and the 500 lbs were "caliber" measurements and did vary in weigh, particularly in explosive weigh. Therefore, the 250kg and 500lbs were practically of the same destructive power, give and take a few pounds.
3- CFS1 only releases from the center line.
4- It makes for proportionality and, hence, fairness in MP games.
5- Historically, the 500lbs/250kg bomb was the most versatile bomb, used by heavies and fighter alike. Virtuality should mimic reality here.

BTW- I too flew the beta version and this one is a keeper!:applause:
 
my question was answered by hubba.
even though i knew the answer,
i still wasn't sure and hoped i was wrong.

"3- CFS1 only releases from the center line."
therefore, "a pair of 132 pound bombs"
would not release from each wing..visually.

a center tank would be cool,
but i can't really see anyone
flying a mission long enough to warrant one.
but i could be wrong there.

then again...thinking about it a little,
a drop tank version would be nice to have.
especially, if that tank could be dropped.
 
Hello Hubbabubba,

Thanks for the compliment. I am somewhat pleased that everything from an exterior view was done entirely in AF99 and Aircraft Animator with the exception of fine tuning the timing for the Flaps and Gear Doors.

I sent out the initial release with just 66 kg bombs because the early Zero didn't really carry ANY bombs. I was thinking though that because of the similarity of weights between the drop tank and a standard late-war bomb, I should have done it that way so that virtual pilots could experience what a take-off with a belly tank would feel like. This paint scheme was only good through about October 1942 anyway. After that, they started wearing IFF stripes on the wing leading edge.


Hello Smilo,

We really can't do drop tanks in CFS 1. Besides, one of the things I did not check was the fuel consumption rates. There is also no way that I know of to simulate the drag of a drop tank. As for a visual on the drop tank, I was running out of resources pretty quickly at the end, so that wasn't an option except via SCASM and even there, I don't know how I would signal releasing the tank.



Just noticed the pop-up label on the attitude gauge on my screenshot. Oops.
Also, has anyone else here noticed some weirdness with some of the stock gauges?
I might have already pointed out that the engine controls for the Hurricane Mk.I don't allow using mixture control to cut the engine.
The Cylinder Head Temperature gauge from the P-47D also seems to start at 0 degrees Centigrade while most other gauges show outside conditions.

- Ivan.
 
I found my first post release bugs in the A6M2 Zero today.

The textures on the underside don't line up with the wings.
Its only off by a couple pixels across one of the spars, but I corrected this already in other projects.
It also is missing the fittings for the drop tank.

Also, the inner landing gear wells don't quite line up with the underside of the aircraft.

- Ivan.
 
Yesterday was our game...

MM plans was for us to first test the aircraft earmarked for PTO duties and see what was their stall speed for deck landing. Your A6M2 was the first used and we went to Duxford. Most pilots, me included, found stall speed to be around 65 knots (we, aviators, disdained miles and kilometers per hour...).

I must admit that the dozen or so participants impressed me; they kept to the plan for more than fifteen minutes!:applause:

But everyone had really one thing in mind; could she fight?

So it turned - surprise!!! - into a giant FFA fur ball. At first, it was mostly Zeke vs Zeke, but some pilots - not aviators - wanted to compare with their own favorite stock and non-stock a/c.

The general verdict was mixed; some liked her, some didn't. This is generally a good sign; an a/c that everyone's like is usually a "mod" that flies itself.

I personally liked it. It's a "fighter aircraft". Nimble, but not to the extreme, I was able to keep my end of the bargain against stock P-51D. Against Fw90, it couldn't keep up in a straight line but, in a turning fight, the Hun doesn't stand a chance.

As a target, it burns and explode with regularity and the smoke and fires effects were well noted by others. Most critics were complaining about her quick loss of "controllability" once pepered.

But it does punish as well. The canon are short lived and should only be used close and personal for closing arguments while the guns have plenty of ammo. Deflection shots did wonders in my case.

I liked the birdcage as it was a good "situational awareness" marker.

All in all, a worthy opponent or mount.
 
Hi Hubbabubba,

Sounds like you all had a good event. Thanks for the after action report. Let me know if there are any issues that need corrected. I already have two or three minor issues that will get addressed in time. I will check about the quick loss of controllability with damage but I believe this is true of most of the CFS aircraft. Generally it is loss of lateral control.

Cannons only have 60 rounds each which I figure is good for about 3 good squirts. Machine guns have 680 rounds per gun which is good for around a minute of continuous fire. The ammunition load for the cannon was increased in later models of the Zero, but never got beyond 125 rounds per gun.

I was a bit surprised by your experience that the stall speed was 65 knots. I am getting MUCH lower numbers. I just went back to test after reading your post. I am getting around 62 knots in a fully loaded but clean condition and around 50-52 knots with full flaps and gear down which is closer to a carrier approach configuration. Without trim, there isn't enough control authority to hold the nose up below about 65 knots, but that isn't because the wing is stalled. I repeated the tests a bunch of times because the aeroplane gets very wobbly near the stall and there isn't a hard pitch down.

To trim for landing requires almost full nose up trim which isn't compatible with a take-off or normal flight. The difference is why I included a trim gauge so the pilot can reset back to neutral after landing.

Regarding Knots versus MPH or KPH, I believe this is very much a modern thing. Period flight testing, manuals and reports are pretty consistent at using MPH or KPH depending on the nationality. There are a few exceptions for USN tests though even some of the USN tests are reported in MPH especially if they involve Army or Foreign aircraft.

I went through an awful lot of very contradictory reports while working on the flight model. I figure those reports are "primary data" and they make very interesting reading.

Thanks!
- Ivan.

P.S. Does that mean you all aren't looking for more Pacific Theater equipment any more?
 
hubbabubba said:
As a target, it burns and explode with regularity and the smoke and fires effects were well noted by others. Most critics were complaining about her quick loss of "controllability" once pepered.

But it does punish as well. The canon are short lived and should only be used close and personal for closing arguments while the guns have plenty of ammo. Deflection shots did wonders in my case.
Click to expand...

I need to check as to why a couple hits on just about ANY aircraft degrade the roll rate so much.

As for the cannon, that is one area I goofed. A 20 mm cannon fires a fairly worthwhile explosive shell. Against flimsy aircraft structures, it should cause a bit more damage than something like a .50 cal solid shot. I had worked out some numbers I believed were reasonable, but since I did no combat testing of this aeroplane before release, I didn't notice the rather light hitting power. The numbers I came up with are similar to the ones used by the 1% folks but different in that I don't believe an explosive shell is enhanced so much by extra kinetic energy. Also the Round and Link weights are closer to reality.

- Ivan.
 
You must log in or register to reply here.
Back
Top