PDA

View Full Version : Gettin rid of the shimmy shakes



OBIO
February 26th, 2009, 21:00
In K. Malinowski's PLZ P37A and P37B is what I am trying to do. They are the CFS2/FS2002 versions downloaded from Simviation.com. The planes work well in FS2004, but they shake when they get to 250 mph. I am clueless about what needs to be changed to get the planes to fly without the shimmy shakes. I changed the MOI to those recommended in Air Wrench, but the planes still shake.

OBIO

sparks
February 27th, 2009, 12:18
When the 'shakes' occur in relatively low speed flight, it's usually related to the aspect ratio of the wing. Every sim version except FSX seems to develop the shakes if the aspect ratio is lower than 5 or so. What are your wingspan, chord and wing surface area?

OBIO
February 27th, 2009, 15:11
Wingspan = 67.58

Surface Area = 610 square feet

Cord = 9.67

Aspect Ratio = 7.49

The air and config files are CFS2/FS2002 formatted...even doing a resave of the files in FS2004 format (via Air Wrench) yeilds a shakey plane.

OBIO

sparks
February 27th, 2009, 20:08
Thanks for the info on the model. I downloaded the PZL P37A package to see things first hand.

I did a quick search on the web for the specs and spent some time updating the dimensions and performance specs with AirWrench. When I was satisfied the wing dimensions wouldn't be a problem, I loaded it in FS9. The first thing I noticed was 'shakes' on the ground.

Most of the time I've seen this, it's been a problem with the landing gear contact points. I changed them to increase the amount of wheel travel. I usually try to keep the dimensions in line with the animation, but in this case I was more concerned with getting it to work. The bottom line with this approach was that it was an improvement, but not the cure.

The next thing I changed was the MOI values. The quick and dirty way to check something like this is to substitute values from a similar size aircraft that you know works, so I used the same values as the stock DC-3. That fixed the shakes on the ground.

In the air, it'll do well over 400 mph in power dive and still not shake, so I'm done. If you want to see exactly what I did with it, you can download the flight model files here:

http://www.mudpond.org/aw_pzl_p37a.zip (http://www.mudpond.org/aw_pzl_p37a.zip)

This isn't tuned at all for handling feel, it's just a first cut for the sole purpose of curing the shakes.

aeromed202
February 27th, 2009, 21:47
I've also noticed a conection between the yaw damp in flight tuning and the yaw MOI. Too much damp for an MOI will produce yaw shimmy in flight. They seem to feed back each other so check on those ratios too. Had a yaw shimmy with a CH46 when on the ground which was fixed with increased yaw MOI value, then a tweak of the stick input corrected for that increase for good handling. Good luck.

Ivan
March 24th, 2009, 15:07
Hello Sparks,
Is the aspect ratio a calculated number or is it a specific record in the AIR file? If it is a specific record, then what is the record number? I tend to tune via AirEd or FDE and want to know what I am looking for.

Thanks in advance.
- Ivan.

fliger747
March 24th, 2009, 19:35
Span devided by the mean aerodynamic chord.....

Ivan
March 24th, 2009, 19:39
Hi Fliger747,
I know what aspect ratio means. I was just wondering whether this was a value that is calculated from other parameters in the AIR file or whether it is directly specified in the AIR file, and if so, where it is. If this is part of record 1204, then I already know about it. If not, then I hope to learn more.

Thanks.
- Ivan.

sparks
March 28th, 2009, 08:07
There is no specific value for aspect ratio in the air file; however, oswald efficiency effectively modifies (reduces) the aspect ratio. (BTW, when
oswald_efficiency_factor in FS is greater than 1, it does not increase the aspect ratio. It will decrease AR)

As far as 1204 is concerned, it's used in some FS/CFS versions and not others. I'm not certain where the break occurred, but AirWrench always sets it. In 1204, oswald efficiency is encoded in 'Planform Adjustment'.

fliger747
March 28th, 2009, 09:26
Sparks:

Does FSX handle the aspect ratio of Biplanes in any better (different) way than FS9? What is the optimum solution for a bipe? Been along time since I have worked with one of those!

Cheers: T.

Ivan
August 1st, 2009, 20:56
I guess I might be reviving a dead post, but in re-reading, I saw that the original problem was an aircraft that bounced a bit on the ground. What I found with CFS is that the spring rates and damping also have a great effect on this. I wrote up a little spreadsheet that calculates the percentage of the aircraft's basic weight that is on each contact point taking into account the ground angle. Unfortunately it doesn't calculate fuel or ammunition loads, or how the angle changes when the struts compress, but seems to work anyway. The basic idea is that if the spring rate is proportional to the percentage of weight on that wheel and the damping is about 1/5 or so (need to confirm in notes) of the spring rate, the aircraft doesn't bounce. I have only tried this on about a dozen aircraft but it has worked every time.

BTW, In checking stock values in 1204 for Planform adjustment, the only stock aircraft that falls below '5000' is the P-51D which IMO is a bit screwed up anyway.

- Ivan.

Milton Shupe
August 3rd, 2009, 05:13
Abacus and Tom Goodrick put out a paper back in the FS98 days. Here is an excerpt regarding gear spring and damping that makes for an interesting read:

"9. LANDING GEAR DESIGN
Proper landing gear design is one of the most important things you can do using FDE.

Almost all aircraft converted from earlier versions of FS have poor landing
dynamics in FS98 because the conversion process does not properly set the landing gear parameters for spring constant and damping coefficient.

Evidently the FS designers made a significant change in the gear simulation and forgot to tell the aircraft designers. Even the original aircraft supplied with FS98 land poorly on stiff gear. This makes them bounce a lot. Curiously, the problem becomes worse the smoother the landing. Since all new designs must first be set up in the earlier version, FS5.1, and then converted to FS98, this is a big problem. For example, the XAS jet bounced so badly just taxiing, it could not be loaded without a crash. It could start parked with the engine off and it would rock fore and aft a lttle, then a lot and finally jump up and crash on its tail or nose!


Parameters


In FDEditor, someone identified two parameters, spring loading and damping, for the main gear and for the center gear. These parameters had been mis-identified in other editors. Shortly after seeing this listing of parameters, the author was able to fix all his hopping problems.

Though the mains only carry part of the static load, and during landing the wings are still contributing significant lift, the main gear springs are sized on the assumption they support the whole aircraft in a dynamic condition at landing. If they supported the entire aircraft in a static load, the force
would be W, the total landing weight of the aircraft.

But in a dynamic condition, we assume the springs are fully stretched and unloaded when the wheels touch.

Then the aircraft falls onto the springs which compresses them until their force has stopped the fall. In this motion problem the potential energy of the fall is W*z and the energy absorbed by the spring is 0.5*k*z^2 where z is the total compression.


Potential energy passing into the vehicle equals energy absorbed:
W*z = 0.5 * k* z^2 (k is the spring constant )

Define z1 where k * z1 = W, the static load deflection.
Then we find 2* W = W * z / z1 and 2 * z1 = z.

This means the dynamic load is twice the static load or 2*W since the deflection is twice the static deflection when the potential energy is absorbed. We can play safe, allowing for a little extra kinetic energy by
increasing the “spring loading” a little beyond twice the weight.

Flight Dynamics by Tom Goodrick

Stopping Hopping

Here’s the solution: First, figure the total weight by adding the fuel weight to the dry weight as found in the .air file. We’ll call that W. Then calculate and enter the following values in the Landing Gear Section of FDE:


Main Gear Spring Loading Factor = 2.72 * W
Main Gear Damping Factor = 0.34 * W
Center Gear Spring Loading Factor = 2.04 * W
Center Gear Damping Factor = 0.24 * W

You may want to adjust these values after trying some landings. However, these settings will be good enough for safe landings. These exact values were taken from the author’s Aerospatiale Trinidad.

Factors from a heavier jet are: 2.01, 0.34, 0.86 and 0.11 which are the values for the 35,000 lb Galaxy jet.

You have to watch for excessive compression or excessive stiffness. With excessive stiffness you will hear a series of strong chirps as the wheels barely touch the runway on a smooth landing. In bad cases the plane will bounce fore and aft or left and right (in roll) as the wheels chatter on the runway.

With these good values, any landing with touchdown velocity between 0 and -400 fpm will look and sound smooth.

Ivan
August 5th, 2009, 13:03
Hello Mr. Shupe,

I found that reference a while back also. It is a pretty good related discussion but isn't really what I am describing here:

Sometimes one encounters aircraft that seem a little wobbly but on closer observation only the tail is bouncing around a bit. Or perhaps only the mains are bouncing. In cases like that, the ratio of the spring rates should be proportional to the weight distribution on each wheel. The discussion in your message describes a ratio of

2.72 * weight
and
2.04 * weight

which probably fit a particular aircraft. What I am describing is using the locations of the contact points in relation to the aircraft's CoG to calculate the percentage of the aircraft's weight that is on each wheel and using those ratios to calculate the spring rates. It isn't quite the same problem you are addressing.

- Ivan.

hairyspin
August 7th, 2009, 09:22
If the problem is shakes on the ground, there may be a scrape point that is touching the ground when the aircraft is at rest and the model may be rocking on tailwheel and scrape points (for instance). Some quite bizarre behaviour can be found in these circumstances.

Ivan
August 7th, 2009, 12:15
Hi Hairyspin,

Have you actually seen this happen? I don't believe that I have. When the scrape points are on the ground or below, I have seen some really poor acceleration or no acceleration at all.

- Ivan.

hairyspin
August 8th, 2009, 04:47
Yes, while setting up contact and scrape points for a new model using a borrowed aircraft.cfg and finely adjusting via Notepad. I wish I'd taken screenshots, there were some spectacular reactions at times.

I should say this was for CFS3. Also that I use ACM nowadays - Notepad lacks a few visual tools...

Ivan
August 12th, 2009, 18:00
Is there still a place ACM can be found for download?

- Ivan.

hairyspin
August 15th, 2009, 12:41
Google a search for Aircraft Container Manager and you'll find it. I got mine from a Rapidshare link, but that link for ACM 2.5 is now dead - not used enough, apparently - but there are others. Be sure to scan anything you download for viruses or spyware etc.

If you're still having bother, PM me with a usable email address and I'll send it.