May 2007 Archives

We completed the last flights for today. I can say, we are back in the air!

Setting the gier integral value to zero worked wonder! Everything is back to normal and as long as you do not make too great movements with the sticks everything works fine.

It seems that the new software version 3.12 has problems with large stick movements so I switched on the "Specky-Bit" which reduces nick and roll signal to half the normal maximum.

Today I was able to complete my testflights without loosing a propeller!

Now watch my newest testflights:

As you can see it's stable enough again to try some outdoor flights. Sadly the weather seems to play bad and we hope to be able to fly outside to the end of the week.

Now that our UAVP fly again, eventually we can try to work on the software. I have several ideas I would like to try to implement. Attitude stabilization, coordinate translations using the compass, throttle adjustments and special throttle courves, GPS support, it all should be possible...

Today we continued our testflights. We still want to eliminate the periodic movement we have on the gier acis. We tested a lot of different parameters, bigger and smaller integregrals as well as no dampening by setting the differential values to zero.

It all seemed to no avail. We had that periodic movement all the time exept when we started perfectly so that no gier correction seems neccessary. As soon as some movements on the gier axis starts the periodic movement never ends again.

Strangly my brother does not experience this problem through he experiences another problem which I have as well. Suddenly without cause the whole UAVP flips or extremelt starts correcting in one of the nick or roll axis. This most often results in a crash or even a broken propeller.

We taped my tests again. You will see our testflights with the 3.12 software using different integrals and no differential. Later on we raised the differential parameter as well in the hope to dampen the periodic movement. In the last scene I rolled back to the 3.11 software to see if the periodic movement exist there as well but as you can see on the movie, it does not exist in the 3.11 software which let's us guess that the problem has something to do with the compass support which 3.12 introduces.

Interesting is that my brother is able to fly 3.12 without a periodic movement on gier, but on the other hand, his gier seems not to be very stable and we saw turn arounds of 180 degrees!

This where my testflights:

Later on we powered up my brother's UAVP and did further tests. It seems that the perodic movement of the gier axis can be eliminated by setting the integral value of the gear closed-loop control to zero.

More tests showed that even a value of one in the gier integral will result in small periodic movements.

Following we show yoy the later testflights my brother did. They were quite successful as you can see! Setting the gier integral to zero seems like a very good idea at the moment!

I think this movie shows quite good that the new setting with the gier integral value of zero is at least a good fix for the moment to counter the priodic movement problem.

The testflights from yesterday clearly showed the need for parameter tuning. So we dedicated today purely to finding the best parameters.

The be short, we failed. After much testing and tuning we can say, that we did not find the best parameters and even have more problems than yesterday.

Yesterday we noted that our gier was periodically swinging. So we tried to fix this today and had to notice that as soon as the gier integral gets a bit bigger the periodic swinging starts. Even when we configured a differential part of 80. Only reducing the integral part to -3 ot -5 helped and even then we had a small periodic movement on our gier.

Furthermore we now have a aperiodic tremble on nick and roll. We tried to compensate with a bit of differential and less proportional, but we could not eliminate it fully.

It seems our parameter tuning turned out to be a bad idea. On the whole the UAVP flies less stable than yesterday and we do not know why.

The problems started with the upgrade to the version 3.12 of the software. It's clear that the changes in the software required changes in the parameter (compass support & differences in P & D calculations of the closed-loop control). But up to now we could not find really perfect parameters.

Here come some movies of our testing today.

First we show you the flip my brother got today. We do not really know what caused it:

Next I show you some of my testflights. If you watch carefully, you see the tremble on nick and gier as well as the periodic movement on gier:

The result of my last flight was another broken propeller...

We will ponder the results of today tests and the try to fix the parameter debacle.
Expect more to follow...

Today the newest testflights have take place! I have to raport that our piloting skills are growing from non existant to the point of being quite miserable...

My brother was the first starting today and he was partly successful. His start was a bit shaky but when he got into the air he stayed there for a little while. Aproaching the trees he looses orientation and starts to correct into the wrong direction crashing streight into the bushes:

Then the problem arose how to recover the UAVP. It seemed lost in quite dense bushes which my brother didn't seem to mind and rushed trough them searching for the missing UAVP. In a short period of time he found it and brought it back:

After this my brother was out of the game for now and my time began...

I started my UAVP quite good and was able to hoover for some time in front of me. Then the UAVP aproached me a bit too much and I tried to stear against. This resulted in such a big changing angle that I automatically panic like an amateur and counter-steered even more in the other direction. It seems I steered a bit too much...

... the results were 1.5 loops and a defty crash:

After this I was out of the game and my brother who meanwhile recovered and fixed his dented UAVP took over. This time his flight was much more controled. He was in the air for a long time and finally landed when his battery run out:

The conclusions are that the new version 3.12 of the Wolferl software definitivly behaves different that the version before. The changes in the paramter P and D resulted in a very different flight behavior and will result in much parameter changing for us.

We both definitivly need more differential on the gier axis (it's oscilating!) and I surly need to lower the integral part of nick and roll, I don't want such a fast reaction to my flight control failures. And furthermore we both want to lower the proportional values of nick and roll too.

We will tune our settings and try out more in the days to come!
Expect more of the story here in my blog.

As you all know we are ready for the first test flights.

We want to keep you posted, so we thought we would not let you miss our latest experiences even trough they were not that successful and even a bit embarrassing. It seems piloting an airship is not such an easy endeavor for untrained guys like us, even trough the closed-loop control stabilizes and helps a lot! Things like mass, velocity and inertia keep you quite busy piloting an UAVP...

Sunday was a bad day for us test pilots... We went to the football training ground which we elected as ideal start place. On the way there we had to notice that there was quite a bit of wind (later that night a small storm hit our area). Silly us we decided to fly anyway...

First, Taarek tried to fly and crashed when he bounced on the ground and his antenna touched his rotor. If you listen carefully you hear the "click" when the antenna hits the rotor and gets cut in half:

Later on I tried my own luck and went into the air with my own UAVP. I was no more successful than my brother and crashed after 22 seconds:

Something went terribly wrong. When I touched the ground you hear the beeper start and then my motors stop and it crashes. It seems as if I lost my battery or my receiver when I crashed. having switched it off and on everything is fine again. I can not yet really explain what happened.

Next we will upgrade our UAVP software to version 3.12 which will support the compass. This will make sure the UAVP's orientation will be stable all the time, since the compass is a reference sensor with no integral error.

Be sure to hear more about our next flights in my next blog entry!

It flies! It flies!

But let's start at the beginning...

After having built our UAVP as described in my last article, today we fixed the last problems and did the first testflights indoors!

Everything went well after we fixed some bad soldering on Taarek's board and after having found out that one of the sensors of my own board was not sitting streight.

We made a small movie showing Taarek's UAVP flying in my living room:

Then finally, when it was nearly dark, we went out for the first outdoor testflight! There was barly enough light to make some camara pictures and so we failed to put the first outdoor flight on movie tape... :(

But to give you an insight in what went on, take a look at these outdoor pictures!

My own UAVP stands ready for it's first integral outdoor flight:

And up it went! As you can see it really flies:

And it can gain height quite fast, I was astonished myself:

And here you see Taarek's UAVP standing ready for it's first real outdoor testflight:

And it jumped into the air as well:

Then we needed to change our location since it was too dark for unexperienced pilots like us. We went up to the street, where there's some light from the street lamps.

Taarek successfully completed several flights near the street in front of the next building:

These first testflights were very successfull and showed the considerable potential of these UAVP constructs!

We were astonished about their ability to gain height within a very short time and about their considerable agility in the air. They are so agile that we even feared several times to crash accidentally against trees, buildings and the like, just because we could not assess how fast the UAVPs change direction, height or orientation.

Be sure to hear more about our next flights in my next blog entry!

As the honored reader probably already knows we finished flight controller and frame. For those new to the story I must mention that in my last two blog entries I talked about the construction of the flight controller and then described how we built our frames.

Our UAVPs already started to resemble quadcopters.

So next we started mounting the four brushless motor controllers which we intended to use. Both UAVPs will use YGE30 brushless motor controllers which are able to handle current up to 30A and should be fast enough for a quadcopter with 54cm diameter.

This will allow us to use a wide range of motors:

We added the flight controller in the middle of the frame and we put some Klett bands under the frame to be able to attach the Lithium ion polymer battery pack we intend to use:

When all connectors and power cable were connected it looked like this:

The second UAVP looks similar:

Next we started pondering the problem of crashes and how to protect the fragile flight controller from impact forces.

To be sure to have the electrics protected all the time we designed a sort of cap for it which surely will absorb some of the impact should it happen to fall on it:

Looking at it on the second UAVP it shows how tightly everything fits:

Mounting the Lithium ion polymer battery pack did not take long.

Finally ready to for flight we connected the serial connection to our laptop and started to tune the flight controller's parameters for it's PID closed-loop controls:

Here is another detail picture:

The second UAVP with mounted Lithium ion polymer battery pack looks similar:

With this we finally finished the construction of our two UAVPs. All pre-flight checks check out and it seems the two quadcopters are fully functional and ready for flight.

The new frames and flight controllers will have to prove themselfes in their first flights. I will tell you more about success or failure of our test flights in my next blog entry!

As I told you in my last blog, we successfully finished two fully functional Wolferl flight-controller boards for our future UAVPs:

Next we wanted to build two stable and strong frames for our quadrocopters.

We bought several four cornered aluminium poles which we used to build a aluminium cross by milling notches into the middle of the poles so that they fit seemlessly. We then secured them in their place by scews and small aluminium plates which we drilled holes into accordingly.

We built two different frame types, one built from solid aluminium poles and one built from hallow aluminium poles.

The hallow variation has a weight of 232g and looked like this:

The solid cross weights 524g and looks not really different:

As you can see the weight difference is nearly 300g so we decided to use the hallow version for our first try on a frame.

Next we drilled holes for the motor anchorage and mounted the motors:

One of the two UAVPs will use Hacker 20-20L motors with EPP 1045 propellers which will give it a thrust of over 600g per rotor. This will result in a total thrust of over 2.4kg for all four of them. Our expected total weight will be around 1kg so we will have some 1.4kg thrust in reserve:

The second one uses AXI 2217/20 motors and looked like this at that stage:

It's motors will do also around 600g thrust per motor with the same total thrust of 2.4kg:

With mounted EPP 1045 propellers the Hacker version looked like this:

Next we unmounted all motors again and started building some landing gear for the quadcopters:

In the end we went for a simple contuction which will also help absorbing some of the deformation energies of hard landings:

When we combined the frame and the landing gear it finally started to resemble something like a quadcopter:

This will serve as a frame for our future UAVPs.

The next step to completion will be adding all the powerlines and mounting the flight-controller in the middle of the construct.

I will tell you more in my next blog entry!