MP 2128G2 Heli Sensors calibration

[Janusz Ochoński]

Hello Everyone!

I try to run the autopilot MP2128HeliG2 in trueHWIL setup. After load a simulation to the xPC Target computer I run the Autopilot and I got a lots of errors from sensors. The message gathered from Terminal is:
[1B][2J[1B][0;0HFATAL ERRORS:
Watchdog CRC Mismatch[1B][K
x accl failed [1B][K
Accl Temperatures Failed[1B][K
GPS config error[1B][K
Compass is enabled but is not installed.[1B][K
[1B][K
WARNINGS:
Autopilot loaded with non-release code.[1B][K
[1B][K
Lockd 11.3V 0.0V
spd= 17
z acc=854
compass hdg=0 0 0 0
I tried to run the calibration trueHwil Control Centre program to calibrate sensors. When I load the callibration_2128.thprj file and tried to load it to run the simulation I got the error of different numbers of parameters between downloaded model to target PC and parameter file. In result I cannot load the model to xPC Target computer to run the calibration. In xPC Target I have this message:

Download started...
Download finished
Mismatch between model and kernel versions

Is the calibration proces necessary for those Horizon fatal errors? If Yes then how I can solve the problem with calibration program problem? 

Thanks for any advices
Regards

Janusz

[Janusz Ochoński]

Hi Kevin, 

Thanks for the answers. I prepared the package for You with the screenshots of oscilloscope signals and the Terminal log from autopilot startup.  I hope that this give You some clues what happened with this pitch gyro sensor. 

[Kevin Surminski]

Hi Janusz,

The ST signals look good and your RRRR and SSSS reports also look good. While I'm investigating this failure further, I'd like you do to a test. Perform the standard start up procedure. Connect in Horizon. Open sensor monitor and add field 2080. Right click on it's value, write to, 0. This should clear the fatal error. Then click power on in the THWIL center, then arm and takeoff in Horizon. Let me know if it flies normally. If not, download the datalog, VRS and fly files and send it to me. Also send the error monitor log or a picture of the error monitor.

[Janusz Ochoński]

Hi Kevin, 

The error was cleared successfully. In the THWIL center i pressed power on button and pressed ARM button in Horizon program, then Throttle indicator starts increase up to the 100% and climb value was oscilate at range from -8 up to 20 m/min. The Takeoff button didn't show up. In TargetPC Scopes the Throttle values switched constantly between 0 and 1. The animation didn't move further. I uploaded the package with the screens and log files.

Regards,

Janusz

[Kevin Surminski]

Hi Janusz,

1) Looking at the datalog, AGL is all over the place which isn't right. I'm not sure why this is. Is the high time of the AGL signal when viewer with the oscilloscope stable? Can you perform the standard startup procedure, then record a ~1 minute video of the signal for me? If you disconnect J17 and probe J17 pin 3, does the signal look cleaner and/or more stable? The purpose of this test is to see if the autopilot is impacting the AGL signal at all.

2) In the datalog, everything looks normal until 54.926 s, then pretty much everything goes awry. What happened at this time? Is this when you clicked power on? Is this when you clicked arm? Is this when the autopilot finished initializing?

3) When you load the model, do you have "load model only" or "load model and custom parameters" selected? Try both and send datalogs of both tests.

[Janusz Ochoński]

Hi Kevin, 

I prepared some results for You with two movies:

https://drive.google.com/file/d/1RESofN_pMxUf0CQFPQzh7CqIi1OoFBlg/v...

https://drive.google.com/file/d/1RDu2JGdPtx-vPofbyb_ykksBUhInXD-K/v...

The oscilloscope screenshots and movies were taken with the "load model only" option was choosen. In the first movie at 33 second the terminal shows that message:
Running pre-release build.. exercise safety precautions during use.................................................................OK
Wait for GPS to lock and Gyros to Stabilize

and after 1 minute and 20 seconds the GPS locked position.

The second movie is the short record of the TargetPC screen when some parameters were constant and some of them fluctuate without reason.

Second shorter datalog is from test with chosing the "load model and custom parameters" option. About 85 second of simulation the Autopilot was powered. After 122 second It display "Wait for GPS to lock..." and at 159 second of simulation the GPS was locked. I stopped the simulation after 409 seconds. The signals from AGL on the oscilloscope were similar to those from previous case. 

During whole tests I didn't press "Power On" or "Arm" buttons. 
I hope that all of this will show You the influence of the autopilot on the AGL signal.

Regards,

Janusz

[Kevin Surminski]

Hi Janusz,

I'd like to narrow down to if the issue lies in the model or not. I'd like you try a specific fixed wing model that we know works. You can find it on Insight under Downloads, CUSTOM tab, Software, Customer specific downloads, 2015b THWIL2 Fixed Wing Model with GPS ACK packet. Make sure the version is 2023-12-08. Henceforth we will refer to this as the 2023-12-08 model. Load THWIL_Fixed_Wing.vrs that you'll find in the .zip. Transfer over your calibration factors. Run a sim and let me know the results. 

Also where do you have the ground of your oscilloscope connected to? 

[Janusz Ochoński]

Hi Kevin, 

We downloaded the files, and i uploaded the vrs and fly file to the autopilot. After whole start procedure, i get the same fatal error in Horizon: 

Fatal error: Pitch gyro is broken.: 90

Whole error logs I send You in attachment. I reseted the error with help of sensor monitor (set value 0 in sensor 2080). Then I pressed Power Up button, and then plane started rotating clockwise. The Yaw value was increased constantly, the Pitch oscylated around 4.57, and Roll value oscylated around -0.5. The good sign was that the Throttle didn't increased and stayed at 0%. 

I couldn't pressed the Arm button becasuse of new compass fatal error in Horizon.

But the AGL value in Target PC oscylate about 0.161, it looks stable. I checked The noise level in AGL signal and it was similar like in previous experiment. During measurements I used an AGL J17 Pin 1 as a ground, or GRD spot on the top PCB plate in Autopilot (this one near P2 green plug). Both ground spots gave the sililar results.

Regards,
Janusz 

[Kevin Surminski]

Hi Janusz,

Did you transfer over your calibration factors from when you did the calibration awhile back? 

In the POST log, I see that the default report is the GGGG (GPS report). But in the VRS that you were supposed to load, THWIL_Fixed_Wing.vrs, field 585=0 which mean the initial report should be the LCD report. Did you change field 585? If not, please download the .vrs from the autopilot and send it to me. Or perhaps you didn't actually load THWIL_Fixed_Wing.vrs? 

Was the current yaw (field 1060) and compass heading (field 1208) stable until you clicked "Power on" in THWIL center? 

Regardless, I'd like you to record a datalog. Start it recording (go to autopilot, start datalog in Horizon). You can look at field 2009 and/or 2010 to see if the datalog has started. When these fields are incrementing, that means the datalog is recording. So once the log has started recording, wait for about 2 minutes. Then click power on and wait another 2 minutes. Then download the log, vrs, and fly files and send to me. 

J17 pin 1 is fine to use as a ground.  

[Janusz Ochoński]

Hi Kevin, 

I used the parameters file from previous positive callibration proces. 
You were right, i used the older vrs file with similar name. Now i checked cerfully the file names and uploaded the correct one. 
During simulation, before the Power on button pressed, the current Yaw and compass heading were stable. You will see it on the sensor monitor screen shots. Only unstable parameters at Target PC monitor were Throttle, ailerons and Rudder. Other parameters on the scopes values were stable.
Also I sent to You two datalogs from two experiments, the only difference was to clear Pitch Gyro error in the second test. In first one I didn't clear this error. In addition I checked the AGL signal on the oscilloscope and it looks great, without big noise. So it was probably the model influence.

Best regards,

Janusz 

[Kevin Surminski]

Hi Janusz,

I'd like you to check the servo outputs from the autopilot which are input to the digital interface.

Perform the standard startup procedure.
Connect in Horizon and wait for ready.
Do not click POWER ON in THWIL center.
Disconnect Horizon.
Click edit under the list of VRS files.
Select default.vrs.
Select the servos tab
Beside "Echo servo adjustments", select "Echo servo zeros".
Click ok.
The Aileron zero should be 1.5 ms. The Ailerons gauge on the target PC should be around 0.
Move the Aileron zero slider so the zero value is 2.000 ms. The Ailerons gauge on the target PC should be -30.
Move the Aileron zero slider so the zero value is 1.000 ms. The Ailerons gauge on the target PC should be 30.

Repeat for the Elevator and rudder.

The Throttle zero should be 1.0 ms. The Throttle gauge on the target PC should be around 0.
Move the Throttle zero slider so the zero value is 1.500 ms. The Throttle gauge on the target PC should be 0.5.
Move the Throttle zero slider so the zero value is 2.000 ms. The  Throttle gauge on the target PC should be 1.

If a gauge doesn't change, or the wrong gauge changes, look at the servo board to digital interface board wiring.

Let me know your results.

[Janusz Ochoński]

Hello Kevin, 

When I follow with this procedure i didn't get any visible reactions from Target PC readings. The Ailerons and Rudder switch between -30 and 30, Elevator was constantly 30, and Throttle switch between 0 and 1. I checked connections and I didn't see any errors, I connected:
J18 to S1, J14 to S2, J15 to S3, and J16 to S4. The Yellow wires as the signals and black wires as grounds. For comparizon I stored shapes of those signals with oscilloscope. Did those signals should be the constant voltage values or the PWM signals? 

To verify the connection of J1 plug take a look in this photo:
https://drive.google.com/file/d/1SzLBCKbviyo23l5bHbp41tnEakWgBu9X/v...

Could I verify those four signals on the J1 plug as the direct output from the autopilot?

Regards,

Janusz

[Janusz Ochoński]

To make results of this experiment more clear, when I used any of those four sliders the xPC values didn't respond. They didn't react for seting zeroes, and also the values didn't stabilize when sliders goes to the left or right side. The values of Ailerons and Rudder changed between -30 and 30 without passing middle positions. Throttle changed similary between 0 and 1 without passing middle positions.