Originally Posted by NEAGRON

Hi Georgez,ive now got the usp-asp-windriver.2011-05-28 installed,have downloaded the KK2 latest firmware but it is still not downloading to the board,im doing some thing wrong,may be some one will point me in the right direction,your's,Marty.

Hi Marty

Try this, it should help

Flashing the KK2: How to update your firmware (4 min 51 sec)

This new FW works great on my quad but on my tri it seems as if the yaw gyro is reversed and it can not select level mode although I have that option on the radio and reciever and it works through testing, anyone else having this problem with tri's? Also is there a way to reverse the gyro?

I just received my KK2 board and wanted to share some observations.

I found two pieces of solder on the back of the board that potentially might cause a short.
Also this board was covered in flux which again could cause problems. I doubt very much that these boards are tested given what I have found.

I suggest before applying power to your new board or for those that may be having flaky problems that you inspect the board (both sides) with an eye loop looking for potential shorts.
Clean the bottom of the board with alcohol and (tooth)brush to remove the residual flux.

Otherwise, the board is quite nice and looks beautiful!

Well done KK - thank you!

HTH - Fred

Originally Posted by MIDNCOCO

This new FW works great on my quad but on my tri it seems as if the yaw gyro is reversed and it can not select level mode although I have that option on the radio and reciever and it works through testing, anyone else having this problem with tri's? Also is there a way to reverse the gyro?

On previous KK boards, you can reverse the yaw servo by connecting to M5, not M4. Not sure if that is the fix for KK2 but may work.

Originally Posted by MIDNCOCO

This new FW works great on my quad but on my tri it seems as if the yaw gyro is reversed and it can not select level mode although I have that option on the radio and reciever and it works through testing, anyone else having this problem with tri's? Also is there a way to reverse the gyro?

From what I've read the fix is to go into the mix editor and invert the yaw value (100 becomes -100 or the other way around).

That is correct as that is what i did as per instructions in the Q&A on Post 1 page 1 this thread Question 14.

Always useful to read first posts for general info and answers to popular questions.

I would reverse your gyro then fly it, you will probably notice Auto Level working now that the Tri is not continually spinning due to incorrect Gyro sense.

Made a quick test with an Octo.
Also tested with one motor US.

120815OctotestKK2 (2 min 28 sec)

KK2 with firmware 1.1 Octo. Homebuilt frame with glassfiber and aluminium arms.
First with 8 motors running, then with 7 motors running.
8 Rctimer Motor BC3536-11 750KV Outrunner Brushless Motor
2 Rctimer Programmable ESC 40A Brushless Motor Speed Controller
6 Rctimer RC Hobbies ESC 50A Brushless Motor Speed Controller
All ESC are flashed with SimonK
Total weight with 3s 3300mah 3Kg

I have now tested the latest FW, and jumped on the CPPM option immediately.
It works great, but there is an error in the CPPM settings menu item: The labels "Yaw" and "Throttle" needs to be interchanged (unless channel 3 & 4 has changed place in my PPM stream, which I doubt.
I had some serious problem after upgrading. It would oscillate heavily at gains above 30%, but was possible to fly below, but was very unstable. Then I figured out that I had forgotten to restore the most important setting, the X quad motor layout! It was at whatever is the default .
Now it works perfectly! Just need to tidy up the remaining RX cabling.


Fred

Originally Posted by flarssen

I have now tested the latest FW, and jumped on the CPPM option immediately.
It works great, but there is an error in the CPPM settings menu item: The labels "Yaw" and "Throttle" needs to be interchanged (unless channel 3 & 4 has changed place in my PPM stream, which I doubt.
I had some serious problem after upgrading. It would oscillate heavily at gains above 30%, but was possible to fly below, but was very unstable. Then I figured out that I had forgotten to restore the most important setting, the X quad motor layout! It was at whatever is the default .
Now it works perfectly! Just need to tidy up the remaining RX cabling.


Fred

The order of channels for PPM SUM varies from vendor to vendor, you have to adjust to suit your product.

Here are some popular combinations taken from the MultiWii source:

Select the right line depending on your radio brand. Feel free to modify the order in your PPM order is different */
//#define SERIAL_SUM_PPM PITCH,YAW,THROTTLE,ROLL,AUX1,AUX2,AUX3,AUX4 //For Graupner/Spektrum
//#define SERIAL_SUM_PPM ROLL,PITCH,THROTTLE,YAW,AUX1,AUX2,AUX3,AUX4 //For Robe/Hitec/Futaba
//#define SERIAL_SUM_PPM PITCH,ROLL,THROTTLE,YAW,AUX1,AUX2,AUX3,AUX4 //For some Hitec/Sanwa/Others

Just testing 1.2 on my HT-450 frame. Flash went fine but default PI is wobbly as all hell. V1 was rock solid, no wobble at all.

Can someone post what the old default settings were for me to try please.

Here's an idea for kaptainkuk; at the moment you can only use the aux channel to switch one function. When i watch the reciever test and operate a 3 pos switch off is low, mid is neither off nor on and high is on.

Why not allow mid to arm and high to switch on level. Even without a 3 pos switch (dx6i etc) you can use mixing to allow 2 switches to set low mid high on one channel.

Well ive got the drivers installed but cant get the KK2.2 installed,any idea's please,Marty.

Originally Posted by e_lm_70

...Yes, in theory it is enough that X and Y are sensing 0 ... Z ... ideally should sense only G ... but more then autolevel, sensing G it means keep constant the speed of of going up or down ...

Anyhow, in the wind with autolevel from stick, my quad does raise the propeller into the wind, and it goes with the wind ... at "double" speed .. one speed the wind, the second speed for the fact that the propeller are not level and are not opposing to the wind direction...

Finally I took a look at imu.asm (same in 1.1 and 1.2) and what you said makes full sense.

The code only uses ACC x and y to approximate angle of tilting. The comment says good approximation up to 45 degrees, but our math teachers probably won't agree. Anyway, if the quad is level but is pushed forwards by a bit of wind, ACC x reading decreases. The code would detect the same thing when the quad is tilting backwards as it looks at x alone. Hence the KK2 will try to tilt the quad forwards making it move forwards even faster...

BTW, the gain of ACC is also somewhat multiplied. It looks like it would be half of P of self level times P for regular pitch and roll, when the cyclic stick stays in neutral. It is humongous if we use the default values. An acceleration picked up by the ACC could easily thow it into the limiters or violent oscillations, but it does not necessarily mean the quad is tilting at all, the acceleration could be caused by something else...

Enjoy.

Originally Posted by Wasp09

Finally I took a look at imu.asm (same in 1.1 and 1.2) and what you said makes full sense.

The code only uses ACC x and y to approximate angle of tilting. The comment says good approximation up to 45 degrees, but our math teachers probably won't agree. Anyway, if the quad is level but is pushed forwards by a bit of wind, ACC x reading decreases. The code would detect the same thing when the quad is tilting backwards as it looks at x alone. Hence the KK2 will try to tilt the quad forwards making it move forwards even faster...

That reasoning doesn't sound right, if you push a hovering quad forward, it will indicate a forward acceleration on the x axis, if you tilt a quad backwards it will indicate a backward acceleration on the x axis due to gravity.

Originally Posted by NEAGRON

Well ive got the drivers installed but cant get the KK2.2 installed,any idea's please,Marty.

search your computer for KKmulticopterflashtool.cmd if your windows create a shortcut to it and put it on your desktop and double click it. If your on a Mac your on your own until a Macker answers you. You have the lates Java on your machine? Make sure that you have the cable installed correctly on the KK2

Originally Posted by Erknie

That reasoning doesn't sound right, if you push a hovering quad forward, it will indicate a forward acceleration on the x axis, if you tilt a quad backwards it will indicate a backward acceleration on the x axis due to gravity.

We are confused by the term acceleration.

Suppose we are inside an elevator. If the elevator accelerates up, we would feel as if there is more gravity, i.e. pulled more towards the bottom. Now label the top of the elevator as front and the bottom as back. So when the elevator accelerates up (labelled as front), we would feel more pull towards the bottom (labelled as the back.) The exact pull is not due to gravity. It comes from acceleration towards the top (labelled as front.)

Now we turn the elevator 90 degrees, we would not feel any pull towards the bottom (labelled as back) nor the top (labelled as front). Then we tilt the elevator back a little such that the bottom is slightly lower that the top. We would feel we are pulled towards the bottom (labelled as back) again, i.e. the pull towards the bottom (labelled as back) increases on the tilt. In this case the part labelled front is slightly higher than the part labelled as back. In flying terms, we are tiliting the elevator towards the back from level.

Hence acceleration towards the front should feel like tilting it towards the back from level if we just consider the top-bottom or front-back axis.

The other way to check is to go the sensor test menu. Trying accelerating the quad towards the front for a period of time such that it is long enough for you to read the change in ACC X reading. (One way is doing a free fall while looking at the quad with the front pointing down, i.e. getting a big forward push by gravity. )

Check if the change goes in the same direction of tilting the quad towards the back from level.

I believe I saw a decrease in ACC X in both cases.

In simple words, we are falling more towards the back of the quad in both cases along the ACC x axis. The ACC does not measure movement to determine acceleration like an outside observer. It detects forces sitting inside the quad. When the quad is pushed to accelerate forwards, the ACC would detect a force pulling it towards the back on x axis. It is the same direction of force derived from gravity the ACC would detect when the quad is tilted backwards from horizontal.

Have fun.