Ok, here is another problem Ive spotted in the mind field of hardware issues and for people who continually bang their heads trying to fix them through PID tuning. .
If you have a housing that is covering over all of your electronics such as a plastic container,.. then when you descend or ascend quickly the pressure inside the container changes and gives your barometer an false reading.
For example if you descend quickly the pressure increases inside the container making the barometer think that you are descending faster than you actually are. So the APM will try to compensate this out and try to increase the throttle and this probably cause a bouncing behavior some people are finding happens .
Some may argue that the barometer cant be this sensitive, but I beg to differ because the pressure variation it can sense for the pressure variances in altitude that it works off is very fine. The pressure change inside the container even for horizontal flight will be enough to for it to sense the changes of pressure in the container.
This also includes pressure zones around the incoming and out going air stream around the props if they are to close to the container or even if you have an open center frame the pressure zone around the props can possibly still affect the barometer
I'm considering some fixes for this such as a tube that has a H tube branch at its end and that sits above or below the center board.
I would be interested to know what others view and ideas are on this. .
I think that you are right, and the problem may difficult to solve in multirotor. I'm studiing this from same weeks and seem that the pressure envelope during dynamic fligth is complex.
In my experience I found the pressure can drop going forward and . ramp going backward. then the quad go down in forward fly and go a little up in backward the same my happen going left and right.
Opening some large holes on the cover don't seems to solve the problem as I aspected ... !?
You can see this in this video:
I mesured the barometri altitude gap in 4 m/s forward fligth and I had more or less 1 m down in a 8 m/s I had close to 3 m.
Actualy I'm testing a tube that read pressure outside the box by same little pinholes
I think more peoples don't see this effect because they use the altitude hold only in near static position.
BTW My quad JJ-copter is a large customized MWii (was Aeroquad until last Dicember) with my own altitude hold sw. The MtoM size is 70 cm the props are 12x3.8, the baro is MS5611
The barometer measures air pressure and the new one on the APM 2 pretty accurately.
I believe that you have hit a on significant problem for Multicopters.
Finding an unaffected static air location is always hard even in conventional airplanes where it is also often used for determining air speed.
In Multicopters even without a dome, static air is really hard to come by and even if you did have an ability to position the sensor, it would be hard to know where to put it.
One effect that I have noticed is that if you have gusty updrafts or down drafts it also affects perceived altitude (ostensibly because of the relative air pressure difference) and the copter travels up or down accordingly potentially even to the ground.
Reality is that for tiny little very close to the ground vehicles like multicopters, an air pressure sensing device is a very inadequate means of gauging altitude especially on a moment to moment basis.
We have sonar but that only really works to about 12 feet or so AGL and it is essentially limited by physics in air from being effective at much greater distances.
What we need is a cheap and very small RADAR or LIDAR altimeter.
Why not just do like a Pitot tube? Planes seem to have no problem getting barometric even in 200+ mph airflow.
I think you guys are over thinking this.
Yes that was mentioned earlier. Not sure we are over think this, its better to explore and discuss it in every direction and sooner or later a simple solution might just pop up. The main point of this post is this is certainly an area to take into account that was never even considered previously. Try this easy test. Run your copter without launching and watch the barometer reading on MP. In most case you will see there is quite a pressure change. Ive noticed between 2 to 5 meters variance in the reading which may be why some launches re a bit wobbly. . Likely most of this is from ground wash affect. Now hold it or suspend it some how well off the ground and now check the barometer when you rev up the props and see what happens. With my small quad where the blades are close to the center plate I get some pressure variance when the throttle is increased and decreased between 1 to 2 meters. This isn't any where near as pronounced on my larger rigs where the props are well away from the housing. The sonar to some extent eliminates some of this affect up to 12 meters.
Pitot tube references local relative air pressure to pressure difference produced by ram air, so this will always show correct differential pressure.
Baro for altitude shows absolute air pressure which varies by local air pressure and aside from weather, local air pressure is affected by flight conditions, gusts, up and down drafts and in the case of multicopters, multiple sources of turbulent air.
Also air pressure difference indicating small altitude difference is very small.
Net result, even if baro is very accurate a small local absolute pressure difference can easily read as many tens of feet of altitude difference. EG: Auto compensation puts multicopter 10 feet underground.
I periodically have a small but noticeable updraft on my property from North winds coming up a small slope from the neighbors property and abit of a down "rotor" over the backside. When it gusts I get ballooning or sinking or more generally both in sequence.
In Stabilize mode I just compensate manually, but Alt Hold sees this as an air pressure difference equivalent to an altitude difference and compensates inappropriately.
The bottom line is that barometers are quite limited for reliable fine altitude maintenance due to variability of local air pressure caused by common physical phenomena especially at low altitudes.
Sure. I hope nobody really expects accurate (1 meter) altitude from a barometer. That should be pretty obvious. There's going to be noise from all sorts of sources and drift from atmospheric changes.
Anytime you've got air blowing on your baro you're going to get a lot of noise and errors. That's why I think most of the issues here just come from unrealistic expectations and poor design of enclosures.
You either have to be out of the airflow (not possible in most cases) or in a proper enclosure that is vented perpendicular to the airflow. Once you do that you're not going to get any better readings.
That's why I think this thread is way over-thinking things. Simple problem, simple solution. Use a non-vented (but not air-tight) enclosure or else make sure the vents are perpendicular to the airflow. Whatever results you get from that are the limits of the underlying technology and physics.
Jake I think you are confusing the difference between air pressure and air movement here. Regardless of having a vent perpendicular to the air movement it wont eliminate pressure changes around that air movement. For example if you have a cover as in this picture and the copter descends, the increase in pressure affects the baro and even if you have the baro covered to prevent the air movement hitting it directly the pressure will still affect it. On some of the tests I did you can get increased reading anomalies between 5 to 8 meters which is a lot. Once I vented the cover it came down quite a lot. So as you see the flow of air was the same in both tests but preventing the pressure build up in the cover by venting it help fix it so some extent. . You are spot on, the issue is about poor design of enclosures, but then whats the best way to fix it to a avoid pressure variances when the craft is flying? The other issue is if the props are to close to the center as well. Why I posted this was mostly from seeing so many posts about people struggling with PID tuning when in may cases I see this issue may be partly a cause. Its like thinking you can fix a flat tyre by pumping it up,.. you need to fix the hole first.
If you have a vent hole perpendicular to the airflow...
Maybe the physics are beyond us here. But you remove the pressure noise from the buffeting and hopefully the venturi effect cancels out the increased pressure from being in the airflow zone.
In any case, as long as you aren't directly in the airflow you should have no problems. If you measure the static pressure you should be fine. It seems to me the only problems anyone is having is from poorly designed enclosures that put the baro directly in the airflow or in some strange enclosure.
Just remove the airflow and you'll be fine. I don't think there's any need to worry about being in the increased pressure under your rotor and there's nothing you can do about it anyways unless you want to extend some sort of tube above your props.
Jake imagine the housing cover is like some of us use such as a CD case. Similar in this drawing. Now cover the bottom of the cover with a thin plastic film, similar to what you wrap sandwiches in so as to stop air flowing into the cover, Now if you descend quickly there is no air going into the cover because of the film however the thin film will push inwards increasing the pressure inside the cover and this would give the baro reading it is descending faster than it actually is. This will make it bounce up and down which is exactly what some people had. This isn't air movement inside the cover but increase in air pressure and this happens without the film as well. You can easily stop wind and air hitting the baro using foam and other devices, but air pressure need to access it what ever the case be for it to work. You are probably better of not having a cover at all and just a frame around the electronics. I have a vario altimeter that is fully protected from wind I use for paragliding and I placed it near the center frame of my small quad and it picks up pressure changes not from the wind but from the change in air pressure around the prop zones in the areas not even in the prop wash. This is very sensitive up to 0.2 meters. . You need to make clear definition between wind and pressure change as they are not always related
Hi John, I agree that a lot of people are probably mistakenly thinking they can tune this out with a PID when the reality is it is a real phenomenon associated with the way altitude is detected and the realities of normal local variances in air pressure.
I wonder if some compensation couldn't be provided by cross checking the GPS altitude with the baro and at least limiting undesirable excursions with this information.
Far from ideal since GPS altitude resolution is pretty limited but it could serve as a basis for reducing if not actually ignoring spurious altitude deviations caused by local air pressure changes.
Of course in reality we would very much like to stabilize within a foot or 2 and even closer if possible, certainly for camera work and the 12 foot altitude limit practically possible with SONAR isn't really adequate.
A kludgey work around might use the optical flow device and a couple "beacons" for triangulation of both position and altitude.
A collimated IR semiconductor laser reflecting back to a linear (X) photo detector could also make a useful finer resolution altimeter. Poor mans LIDAR.
From what I understand the baro reading is mixed with the sonar to try and give some stability and accuracy up to 12 meters. Im going to fit the optical sensor to my quad and test how well that works. I did some reading on Lidar and this seems to be the future once an affordable compact unit is available
You would want to avoid the design you just mentioned. I don't think you would have the same problem if you had the flexible area or vents on the side perpendicular to the airflow.
Actually, I think you made my point for me. If your box had a flexible piece of sandwich wrap on the side rather than the bottom it would not flex in on climb or decent. It might flap a little from the airflow, but wouldn't be pushed in very much.
The tendency of the pressure from the airflow to enter the container would be offset by the Venturi effect and/or Bernoulli's principle. Perhaps Brad or some other physics expert could explain this better, but I believe as long as your vent is perpendicular to the airflow you aren't going to have the added pressure effect you're talking about.
Otherwise how could you get a static pressure from a pitot tube? I'm not saying that it will give you an exact barometric reading, but it should be pretty close. Certainly a lot better than the problem enclosures mentioned in this thread.