GPS Track is constantly moving around the drone

Hi, new member and new pilot.

I’ve assembled a HMF S550 with a PixHawk running the latest copter, and with a lot of reading have almost got it ready to fly.
I have managed to get it off the ground after a few issues, worked through with help from old threads here - so many thanks for all the info so far, but i have one or two things i am uncertain of that have stopped me going for a full test flight.

On the flight data screen, my black line (GPS Track) is moving all around my copter when i test with or without main power supply attached (not motor / esc interfearance)
What does this suggest? It has 3d GPS Fix. and its heading is correct.
Would instability of this GPS Track cause me a fly away or similar issue?
Or is this a compass related issue that could cause me to yaw or toilet bowl?

Im very new to this, but have been reading for months so i hope i don’t fall into the idiot category, but a lot of this terminology is not fluent to me yet so i may have mis-understood something along the line.

Thanks in advance for any assistance you may be able to offer me!

1 Like

Short answer, I believe it is ok.

GPS position from the Satellites is not pinpoint accurate. GPS position is only accurate to, I believe to a 3m diameter circle. Arducopter uses the other sensors to hold a tighter position. UNLESS you are using a RTK GPS setup, but it does not sound like you are.
I am sure someone could give a better technical answer, but I think you are ok.

Thanks - im heading out to the field now for an attempt at a first proper flight

well i didnt achieve much - other than my throttle is ridiculously overpowered.
it flies - but without any motion from me it flew off for 20m ish - and i cut the throttle before it decided to go further.

ive started another thread to sort that issue out before i look at this again, but it certainly didnt hold any position :frowning:

Hi DoubleT,

I also have an HMF550 with Pixhawk and it’s flying extremely well on 3.5-rc1 with EKF3 enabled. I’m happy to help in any way to get you flying stable. Drop a log and I’ll take a look to see if there’s anything that jumps out.

Maybe you can share more info about your setup? Here’s mine:

  • HMF550
  • HMF650 Retractable Landing Gear
  • Pixhawk (Chinese Clone) with UBLOX M8N GPS
  • 1.3" OLED On-Board I2C Display (newly supported in ArduCopter 3.5)
  • Mauch Electronics Power Module (BEC + Current Sensor)
  • SunnySky X2212-980KV Motors
  • APC 9x4.5 props
  • Mystery “M-30” SimonK ESCs w/ BEC
  • FrSky X8R Receiver w/ FrSky Taranis Transmitter
  • Craft and Theory FrSky Telemetry Cable
  • Zhiyun Z1-Tiny2 3-axis GoPro Gimbal
  • Grayson Hobby 5Ah 40C 4S Batteries

Thanks John - good to know someone has achieved (im sure there are loads) what im trying to do.

My setup is:
HMF S550 F550 with standard carbon fibre T-Shape Landing Legs
JMT 920KV CW CCW Brushless Motor
30A Brushless ESC Speed Controller
9443 CW CCW Propellers
Pixhawk Flight Controller
6M GPS with Compass L5883 25cm Cable
DJI GPS Folding Antenna Mount Holder Metal
RadioLink AT10 2.4GHz 10CH Remote Control System
3DR Telemetry (although i ordered the wrong cable so this isnt in use yet)

Biggest concern atm is getting the flight controller to majorly reduce max throttle - i can see how to do it on the RC, but reading tells me this is not the thing to do. But i cant find thr_min or thr_max.
Hand test feels like all else is now resolved - i need a wind free day to be sure though :slight_smile:

Logs of a flight so short you can barely call it a flight!
https://goo.gl/yoYiAg

I’m looking into any london (or close) arducopter meetups as i’d like to learn from some pro’s but so far all the groups ive found seem to be centered around ready to fly or off the shelf models.

The setup looks good. What battery are you using with that setup and do you know the all-up-weight?

With that very heavy 4s battery and beefy gimbal and gopro, not to mention the additional servos for the retracts, my copter still hovers at about 40-45% throttle. This corresponds to the THR_MID parameter set in Mission Planner. Total weight of my setup with battery is about 2,444 grams.

Depending on what battery you’re using and the fact that it doesn’t look like you’re running a gimbal/camera your power to weight is going to be crazy and it will fly and hover at a very low throttle percentage.

It looks like this may be the case since looking at your logs even at low throttle input, the outputs appear to be well above your throttle position. I would try to reduce the THR_MID parameter to whatever the minimum allowed value is, or add some weight.

Makes sense with what I had read, thanks for confirming.
Sadly I only have thr mid acro,no min or max or mid.

Battery is 3s 4400mah
Not sure on weight can find out tomorrow

Yep…looks like that’s been replaced with a parameter called MOT_THST_HOVER and MOT_HOVER_LEARN. Your MOT_THST_HOVER is at 0.6875 and with MOT_HOVER_LEARN you are setup to automatically learn the hover percentage. However, this “learning” only takes place in non-manual flight modes and I think in stabilize flight modes the MOT_SPIN_HOVER doesn’t really matter as motor speed is being controlled by the RC Transmitter.

I’m still curious as to why the motor output is higher than the RC3 (throttle) input. It looks like the graphs for ThrIn match up with the graph for ThrOut perfectly, but it just looks like the pulse length of the output signal it not matching up with the pulse length of the throttle signal that’s been input. I think it has something to do with the motor scaling described here: http://ardupilot.org/copter/docs/motor-thrust-scaling.html. It’s causing the startup power of the motor to kick in right away as soon as you apply any throttle, and because it’s too powerful for the weight it feels like it’s super sensitive. But I’m honestly just making some educated guesses here.

I stand by the suggestion to try to get your input endpoints to closer to 1000 min and 2000 max. There are lots of guides online for how to do that with the AT10. Once you’ve done that I would re-calibrate your ESCs and try again. If it’s still too powerful, maybe try to lower the MOT_THST_MIN parameter or adjust the MOT_THST_EXPO parameter closer to 0 so the response is more linear at the low-end of the throttle input.

I’m learning on-the-go here with some of the changes made in AC3.4/3.5. Been spending too much time on racing quads with Betaflight!

Brilliant - Thanks John
Will let you know how i get on, probably the weekend before i can put it up again, but this has been a huge help.
I now understand why i was seeing the behaviour i was - and all looks to be okay actually.

Telemetry cable arrives tomorrow with a bit of luck - time to find a bigger field and trust in Mission Planner a bit more.

It is normal for the indicated GPS position of a stationary receiver to “wonder” due to GPS errors from various sources. Sometimes the error can be great. This “feels odd,” but is normal.

There is error in every position the GPS receiver calculates and this error affects where the GPS/autopilot thinks it is vs where it really is.

Consumer-grade GPS receivers apparently smooth out the calculated position and, I assume, show an average position, or snap the apparent position to the nearest roadway on the map in the case of GPS receivers for cars.

Survey-grade GPS receivers correct the error with RTK (Google it).

The amount of error in a GPS position depends on, amount other things, the spread between GPS satellites used. Each satellite is in its own orbit and there are times when most of the available satellites are bunched up in one part of the sky. This will yield poor GPS accuracy for the 5-minutes or so that they are bunched up. You may/should avoid flying during such times. Trimble has a free app for planning the best time to do a GPS survey that may be of use.

Don’t assume GPS hardware provides an accurate position all the time. For navigation while driving or walking, the error isn’t a problem - the GPS will normally get you close enough to where you want to be that you can find what you are after. For drones using GPS, expect them to wonder unless you are using RTK.

I suspect you’re using AC3.4.5, the latest official version, rather than AC3.5 (which is still in beta testing) so I’m going to move this discussion to the Copter-3.4 category if that’s ok. If you’re using MP you can see the version at the very top and center of the window.

For new users, it’s definitely better to use the latest official version rather than a beta so we can separate the regular initial setup issues from the beta testing issues.

Made some progress but still not completely happy or confident with it.
Had a couple of rough landings and have managed to snap the main power distro board :frowning:
Few weeks for a new one to arrive. I thought i might get through some rotos, but never a main board… hey ho.

Thanks for all the info above, its been a massive help.