Kiss ESC and Quatos

The new control algorithm for AutoQuad flight controllers.

Re: Kiss ESC and Quatos

Postby Cookiemonsta » Thu Jan 21, 2016 6:56 pm

Hi guys :)

I recently built an experimental frame (12mm tubes and self constructed) and tried to get quatos dialed in but it doesn't fly stable. It tries to yaw and/or dips to front/side...

My Setup:
-Fully calibrated AQ6 with DIMU and GPS (tried PID but those just made it oszillate in high frequency which I couldn't get rid of, other than that flew ok, no drift)
-Kiss 18A V1.2 ESC
-Tiger Air gear motors with original 9x5 Props
-D4r-II Frsky Rx
-Micro minim OSD (not yet flashed or connected to FC)
-FT951
-Pz04020h FPV Cam
-SLS 4s 7000mAh LiPo (583gr)

All together 1332gr

20160121_183213.jpg

20160121_183223.jpg


I bought a 2,5Kg Hobby license and made this xml-file for it
Trainingsquad.xml
XML quad
(1.75 KiB) Downloaded 102 times

I calibrated the KISS Esc, and inserted these A1 and A2 values with help of Jörg's calibration data
Kopie von thrustCalc.xls
Thrust Calculations A1 A2
(40 KiB) Downloaded 133 times

The RPM_SCALE_ is 10500 according to this thread. ESC_Type is 0 for external

As I previously told I tried to fly (in my appartment, with GPS Antenna disconnected) but it either tried to yaw or dip to one side...completely random.

Any hints? Did I do something wrong?

Thanks in advance! :)
Tobi
Cookiemonsta
 
Posts: 126
Joined: Thu Apr 11, 2013 3:31 pm

Re: Kiss ESC and Quatos

Postby aBUGSworstnightmare » Thu Jan 21, 2016 8:46 pm

Hi Tobi,

visualization of the craft looks o.k., although you should try to balance it a little more (weight distribution on the craft).

As I already told you: the RPM_SCALE from the spreadsheet is only valid when using ESC32Vx.

EDIT:
Don't know your parameters, so lets assume:
MOT_MAX 1950
MOT_MIN 1000

--> MOT_VALUE_SCAL = MOT_MAX-MOT_MIN = 950.
Calculate like this when NOT using ESC32 with QUATOS

EDIT2: You may want to
- increase QUATOS_MAX_OUT to 6
- set QUATOS_PROP_K1 to 14.5

It is possible that it want's to YAW; you need to 'dial in' your AM1/AM2 Values. A good starting point for your setup should be
AM2 = 12
AM1 = 6

When're able to hover let us have a log-file please (place it in Dropbox folder i.e.).

Which FW are you using (Mennos HOTT)? Post your parameter file please.

Joerg
aBUGSworstnightmare
 
Posts: 1460
Joined: Fri Jun 22, 2012 5:24 pm

Re: Kiss ESC and Quatos

Postby Max » Thu Jan 21, 2016 10:23 pm

I wonder what the vibrations look like on that frame. I'm hopefully wrong but at first glance, with the skinny long arms and minimal center plates, it looks like that could be a problem.

I've found so far that flying with PID controller first (before Quatos) can be good indicator of the overall system health. In other words if it isn't going to fly with PID, it isn't going to fly great (or at all) with Quatos either. Quatos might be able to compensate for a lot of issues, but ultimately it can't work magic (fix physical problems with the build). So when you say that it doesn't fly with PID, that suggests deeper issues than just the Quatos settings. High frequency oscillations you describe are a prime example of "too many" vibrations (can usually damp it down by adjusting Tilt Rate D and friends).

Cheers,
-Max
Max
 
Posts: 2814
Joined: Mon Aug 13, 2012 9:45 pm
Location: Near Ithaca, NY, USA

Re: Kiss ESC and Quatos

Postby Cookiemonsta » Fri Jan 22, 2016 12:09 am

Hi Jörg and Max my dear fellows :)

Last time I tried your values, Jörg, like 4 potences lower it almost flew through the roof even in idle...no matter if I make MOT_VALUE_SCAL 950 or 10500 so I placed your thrust and rpm measures in the example excel sheet and inserted those.

I tried to dial it down with AM1/AM2 values 8/15 respectively but thats how it is. I'll try that other values you mentioned.
I use mennos HOTT firmware that's right. Latest.

Max...you could be right, too...though I made it as stiff as it could get and this huge lipo should dampen a little bit just because of its mass...I tried lowering / increasing tiltrate D whenusing the PID version from 2500 to 20000 but it just got a little bit better somewhere...

Bit at least the model is ok...I'll try those other two values tomorrow

Good night and thanks so far, I'll keep you updated ;)

Tobi
Cookiemonsta
 
Posts: 126
Joined: Thu Apr 11, 2013 3:31 pm

Re: Kiss ESC and Quatos

Postby aBUGSworstnightmare » Fri Jan 22, 2016 7:36 am

Cookiemonsta wrote:Hi Jörg and Max my dear fellows :)

Last time I tried your values, Jörg, like 4 potences lower it almost flew through the roof even in idle...no matter if I make MOT_VALUE_SCAL 950 or 10500 so I placed your thrust and rpm measures in the example excel sheet and inserted those.

I tried to dial it down with AM1/AM2 values 8/15 respectively but thats how it is. I'll try that other values you mentioned.
I use mennos HOTT firmware that's right. Latest.

Max...you could be right, too...though I made it as stiff as it could get and this huge lipo should dampen a little bit just because of its mass...I tried lowering / increasing tiltrate D whenusing the PID version from 2500 to 20000 but it just got a little bit better somewhere...

Bit at least the model is ok...I'll try those other two values tomorrow

Good night and thanks so far, I'll keep you updated ;)

Tobi


MOTOR_SCALE_VALUE in Mennos HOTT FW for QUATOS needs to be 950 (or MOT_MAX - MOT_MIN); that's all I can say here.

Please note that my thrust values were generated using a ESC32V3 --> I pretty much expect that a different ESC can give you different values.

Be aware of those motors @ 4S! They are real beasts! I would not dare with a 4S in my living room!
aBUGSworstnightmare
 
Posts: 1460
Joined: Fri Jun 22, 2012 5:24 pm

Re: Kiss ESC and Quatos

Postby Cookiemonsta » Wed Jan 27, 2016 7:30 pm

Or should I insert the pwm values starting from 0 to 950 (1000-1950 ) for the rpm entries in the excel sheet and take those a1 a2 values then?
Cookiemonsta
 
Posts: 126
Joined: Thu Apr 11, 2013 3:31 pm

Re: Kiss ESC and Quatos

Postby Max » Wed Jan 27, 2016 10:17 pm

The whole PWM thing is just needlessly complicated. As is that spreadsheet. Use the attached one instead. Read the notes at the bottom.

If you have PWM values for your data now (instead of duty %), just convert those to duty first before plugging the numbers into this example. It would be Duty = PWM / (PWM_MAX - PWM_MIN) * 100 or in your case D = PWM / 950 * 100

I hope this helps.
-Max

V2T_poly_calc.gif
Attachments
Value to Thrust poly calc.xlsx
(12.14 KiB) Downloaded 116 times
Max
 
Posts: 2814
Joined: Mon Aug 13, 2012 9:45 pm
Location: Near Ithaca, NY, USA

Re: Kiss ESC and Quatos

Postby Astudillo » Fri Mar 04, 2016 10:01 pm

Hi Max, Its me again! :P

I received some new triple blade propellers in the mail, so today was thrust measuring day. And i came across your Excel file which was looking less complicated than the rpm one. So i gave it a try, and the graph is working just fine. But somehow the A1 and A2 values stay the default numbers. I tried this on osx and windows, with the same result.
Also i assumed that you can use the Excel for measurements taken with a esc32, right?

Cheers,
Ray
Astudillo
 
Posts: 139
Joined: Thu Apr 25, 2013 8:23 am
Location: Amsterdam, The Netherlands

Re: Kiss ESC and Quatos

Postby Max » Fri Mar 04, 2016 10:51 pm

Hi Ray,

I don't know... you must be pasting wrong :)

EDIT: wait... do you mean the numbers in the graph aren't updating, or values in the Notes section below? The Notes won't change, just the graph and those values should.

Just post the data here, I'll plug it in.

The formula/graph will work for anything in the "Duty" column (RPM, PWM, voltage even) as long as the MOT_VALUE_SCAL is set appropriately. So for ESC32 measuring RPM vs. thrust, it would be maximum RPM (as per Quatos instructions). Or for PWM it would be the maximum range, typically 950 or 1000us.

For ESC32 measuring RPM vs. thrust is going to be more accurate because (a calibrated) ESC32 can set the RPM exactly whereas thrust at a certain duty % is harder to predict.

-Max
Max
 
Posts: 2814
Joined: Mon Aug 13, 2012 9:45 pm
Location: Near Ithaca, NY, USA

Re: Kiss ESC and Quatos

Postby Astudillo » Sat Mar 05, 2016 2:02 am

Hi Max,

I meant these numbers:
excel.jpg


Those are suppose to change right?

Ah yes that would have been my next question about MOT_VALUE_SCAL, so with this method the resolution you work with is quite a bit less than with a full rpm/thrust measurement of lets say 8000rpm instead of the Duty of 100.

Ill do a full rpm/thrust measurement!
I'm just thinking out loud here, but it would be a awesome feature for the ECU to have a option to run a batch of different rpm's with a tap on the space bar or mouse click. Or maybe automated with a small interval like in the calibration process so you can go through all the rpm's a lot quicker.

Cheers,
Ray
Astudillo
 
Posts: 139
Joined: Thu Apr 25, 2013 8:23 am
Location: Amsterdam, The Netherlands

PreviousNext

Return to Quatos Adaptive Control

Who is online

Users browsing this forum: No registered users and 1 guest

cron