Check your static pressure

This may be of relevance to Dynon and MGL autopilots (if the built in AP is used in case of the MGL).

During tests we discovered an interesting effect related to static pressure. This tends to affect only a few aircraft and is directly related to the static system and where and how static pressure is measured.

The new AP systems tend to be extremely good in resolving static pressure to very fine resolution. This can be a problem. Depending on type and placement of the static port, even very small changes in pitch attitude can lead to noticable changes in measured static pressure. This may be because the port is located in an area that is affected by turbulence caused by appendages or very sensitive to actual AOA. The changes in pressure are very small and do not really show on VSI or altimeter (not in a way that you would be bothered) - but the AP does see this and will initiate a correction (that it should not). This can be the start of an endless series of attempts by the AP to exactly capture the desired altitude - but it never manages, the static pressure keeps running away...

If you have issues like this that you cannot calibrate out by the AP's sensitivity settings, try the folowing:
For a test, vent the static directly to the cabin (static in the cabin does not seem to be influenced greatly by small attitude changes).

If this fixes or improves it. Time to find a better place for your static port...

Rainier
CEO MGL Avionics

Rainier Lamers said:

If this fixes or improves it. Time to find a better place for your static port...

Click to expand...

This problem has been seen on some type certificated aircraft, and the autopilot manufacturers have been successful in adapting the autopilot control laws to deal with it.

It isn't really practical to change static port locations in most cases, as it is very difficult to find a location that has acceptable position error over the whole airspeed range. If an autopilot is so sensitive to this issue, I think the proper fix is in the autopilot.

Kevin Horton said:

This problem has been seen on some type certificated aircraft, and the autopilot manufacturers have been successful in adapting the autopilot control laws to deal with it.

It isn't really practical to change static port locations in most cases, as it is very difficult to find a location that has acceptable position error over the whole airspeed range. If an autopilot is so sensitive to this issue, I think the proper fix is in the autopilot.

Click to expand...

Yes, you can desensitize, delay reaction, add dead bands and do all sorts of things.
But none of this fixes the real issue which is an induced measurement error.
Doing any of the usual fixes unfortunately reduces the AP's overall response and accuracy (at least during the tracking phase, you have to accept a greater altitude error, longer time to fix it).
But yes, this can be an acceptable trade-off if you don't want to change the static system.
Having said that, it does look like for all practical purposes a good location can easily be found - just in some instances not much thought is given to the location of the static port and sometimes less that ideal locations are used.

My post was intended to give a simple explanation and an even simpler method of finding out IF a problem is related to static, also in view of recent posts. Sometimes it is not obvious...

Rainier
CEO MGL Avionics

I was having an issue with the trutrak altitude hold that was related to the static port. The airplane would make these little up and down movements for no apparent reason. I simply removed the static system from the autopilot and left the port open and vented to the cockpit. The problem went away. I have been flying like this all summer and I have yet to see a disadvantage. My set altitudes match with what I am seeing on my instruments and even if it did vary slightly due to outside pressure changes this is easily adjusted.

Back when I was trying to get the Dynon AP to work I tried venting the static pressure to the cabin multiple times and saw no noticeable improvement in behavior.

I am more and more convinced that Dynon is simply missing part of the equation. There is no other possible explanation since the trutrak worked in my airplane on the first try. The Dynon and Trutrak servo installations and mounting hardware are identical, so that pretty much narrows it down in my book.

Posted from my iPhone so I apologize in advance for fat fingered typing.

Small problem

with taking static from the cabin, at least on my buddy's -7A comes when you open and close the cockpit vents. It apparently changes pressure in the cabin and causes the A/P to climb/descend until it equalizes

small problem A#2

Tom Martin said:

I was having an issue with the trutrak altitude hold that was related to the static port. The airplane would make these little up and down movements for no apparent reason. I simply removed the static system from the autopilot and left the port open and vented to the cockpit. The problem went away. I have been flying like this all summer and I have yet to see a disadvantage. My set altitudes match with what I am seeing on my instruments and even if it did vary slightly due to outside pressure changes this is easily adjusted.

Click to expand...

Just my theory, but if you change airspeed while asking for constant altitude this may not work quite as well as through the static system. I use the AP when doing speed tests.

A different approach

One has to think about the ultimate goal of the autopilot:

Is it to be "completely accurate within 1 foot, while jittering the stick 10 times per second"? While this may be the goal for a cruise missile, it's not too pleasurable for manned flight.

Or is it to be "very smooth, like an old hand pilot"? Every thing I've ever read, heard or been taught about maneuvering flight is "conservation of energy" using the smallest, smoothest movements that accomplish the goal. Sometimes it's a yank and bank, but still you strive to do it smoothly.
And yes, straight and level is considered a maneuver.

While I can appreciate the idea of sub-foot accuracy in an autopilot, is it really necessary?

My RV is already near perfectly stable in vertical pitch once trimmed, only requiring a gentle two finger adjustment every few seconds to stay within +/- 10 feet, without the autopilot engaged. Seems like the airplane is already doing 99 percent of the work.

My personal preference would be for an autopilot that is "smooth". I know my passenger(s) will prefer it. If we occasionally excurse a few feet (due to a disturbance, etc) I'm perfectly fine with that, as long as the recovery is "smooth".

We're getting there, but I think sometimes us highbrow engineer types get buried in the minutiae and forget the ultimate goal, which is to enhance our flying experience and reduce pilot workload.

These RV's are good birds, light in the controls and generally need to be flown with a smooth, light touch.

BTW, Rainier, this is in no way a denigration of your excellent work, it's just my personal point of view.

Sometimes slower gets there faster.
Often, less is more.

Food for thought.

YMMV,

Rainier Lamers said:

Yes, you can desensitize, delay reaction, add dead bands and do all sorts of things.
But none of this fixes the real issue which is an induced measurement error.
Doing any of the usual fixes unfortunately reduces the AP's overall response and accuracy (at least during the tracking phase, you have to accept a greater altitude error, longer time to fix it).
But yes, this can be an acceptable trade-off if you don't want to change the static system.
Having said that, it does look like for all practical purposes a good location can easily be found - just in some instances not much thought is given to the location of the static port and sometimes less that ideal locations are used.

My post was intended to give a simple explanation and an even simpler method of finding out IF a problem is related to static, also in view of recent posts. Sometimes it is not obvious...

Rainier
CEO MGL Avionics

Click to expand...

Tom Martin said:

I was having an issue with the trutrak altitude hold that was related to the static port. The airplane would make these little up and down movements for no apparent reason. I simply removed the static system from the autopilot and left the port open and vented to the cockpit. The problem went away. I have been flying like this all summer and I have yet to see a disadvantage. My set altitudes match with what I am seeing on my instruments and even if it did vary slightly due to outside pressure changes this is easily adjusted.

Click to expand...

Yep. Now, with the AP engaged, close all of your vents - but be ready when you close the LAST ONE.



Not so bad if you are ready for it, but pretty exciting the first time in IFR.

This solution won't work for folks whose AP is also their EFIS, as it will lie about your altitude. But it can make life nice with a Trio or TT.

Two of my coworkers were doing some autopilot testing on a Canadair CL-415 water bomber. The autopilot had altitude hold, but no altitude preselect mode. Bombardier had difficulties with the autopilot not liking noise in the static system, so they removed the autopilot from the static pressure system and vented the autopilot's static port to the cabin. One day, with the autopilot engaged in altitude hold mode, the flight test engineer, who had consumed large quantities of Mexican food and weak fizzy beer the night before, experienced an intestinal tract overpressure event, and his intestinal relief valve cracked, releasing a large quantity of noxious gasses into the cockpit. This rendered the cockpit atmosphere uninhabitable, so the Bombardier test pilot opened the cockpit window to get some fresh air. This lowered the cockpit pressure, and the autopilot dutifully did an eye-watering 1500 ft descent (actually, their eyes were watering before the descent). After the cockpit atmosphere was breathable again, the cockpit window was closed, which triggered an aggressive pitch up to climb 1500 ft.

Bombardier agreed that these significant altitude deviations were unacceptable, and the autopilot control laws were tweaked to allow it to live with the noise seen on the static system.

Good one !
Thanks for the laugh

Rainier

Kevin Horton said:

Two of my coworkers were doing some autopilot testing on a Canadair CL-415 water bomber. The autopilot had altitude hold, but no altitude preselect mode. Bombardier had difficulties with the autopilot not liking noise in the static system, so they removed the autopilot from the static pressure system and vented the autopilot's static port to the cabin. One day, with the autopilot engaged in altitude hold mode, the flight test engineer, who had consumed large quantities of Mexican food and weak fizzy beer the night before, experienced an intestinal tract overpressure event, and his intestinal relief valve cracked, releasing a large quantity of noxious gasses into the cockpit. This rendered the cockpit atmosphere uninhabitable, so the Bombardier test pilot opened the cockpit window to get some fresh air. This lowered the cockpit pressure, and the autopilot dutifully did an eye-watering 1500 ft descent (actually, their eyes were watering before the descent). After the cockpit atmosphere was breathable again, the cockpit window was closed, which triggered an aggressive pitch up to climb 1500 ft.

Bombardier agreed that these significant altitude deviations were unacceptable, and the autopilot control laws were tweaked to allow it to live with the noise seen on the static system.

Click to expand...

I took the plane out last night to go to a local BQ. On my way there the air was a bit bumpy and I tested the altitude hold while opening and closing the air vent. Yes there was an effect but it created no more altitude changes then the slight bumps I was going through.
After the meal, (no mexican food was consumed), I tested the system in calm air. When a vent was opened or closed the autopilot would sense a change and the altitude would go up, or down, about 50 feet. This was not an abrupt movement, say 5 to 10 seconds. After about 30 seconds it would go back to the original setting.
Yes, opening and closing a vent will make a difference to the auto pilot. However I, or my passenger, control this event and I had no control over the bobbing up and down that the unit used to do. Previously, with the autopilot connected to the static system, there was a bunch of little up and downs, the kind of thing that would make my passenger sick after a while. I would MUCH rather have the unit operating as it is now, especially in IFR conditions then I would with the unit hooked to the static system.
Of course this could be completely different in another aircraft. I totally agree with a previous poster, smooth is better.