Would someone be so kind as to have a look at this log.
Heres the background. I had crashed this drone and after repairing it I was doing my second test flight.
I got a GPS lock and took off. I flew about 12 or so minutes. I suddenly got this strange feeling to put my fingers on the Flight mode switch…it was flying in Poshold. The winds where high and it was fighting them well. But this odd feelign came over me and so I grabbed the switch and waited. Suddenly the green status LED turned yellow and the wind started to carry the drone away.
I switched to stabilize and landed. No damage all looked good.
Oh and I had a battery checker plugged in to see how long the 10000mah would last before having to land.
I saw 3.9 per cell when it landed.
I am puzzled by this…and suck at log reading.
So any input would be appreciated.
Your battery failsafe was set to 21.6 and RTL. Battery voltage dropped below 21.6 volt and copter switched to RTL mode and started to raise to 15meters (rtl altitude) but you stopped it within 2 second by switching toi stabilise. Thats all.
Thank you Andras. I went through the log and came to the same conclusion…then started to doubt myself because the per cell voltage was 3.9…which is 23.4…well above fail safe…so I am suspecting a bad calibration on the voltage sense.
Thank you for your input…its greatly appreciated.
Hi Ricky - You asked 
Yes, I always calibrate the voltage at the lowest you’d expect to see since usually the voltage sensing is a bit non-linear. The lower voltages are the most critical
Current monitoring seems wrong too, maybe a 10th of what it should be ??
Motor out 4 is a bit lower than the rest, maybe a bit of weight imbalance.
Vibe X and Y are bad, like props out of balance
Vibe Z is not good, but not as bad as X and Y.
Vcc (5volts DC) to the FC is a bit low, and goes crazy during failsafe like you’ve got an onboard buzzer - probably disable that with:
NTF_BUZZ_ENABLE,0
You can almost never hear those little buzzers in-flight anyway, and on the ground you can tell what’s happening without the buzzer.
I’d definitely do something with the 5 volts too.
I have a Pixfalcon that had the same Vcc issues and I had so solder the power wiring to the PDB and to the FC circuit board with a MPX connector in the middle to allow disconnection. Put an electrolytic cap in the modified wiring harness on the FC side of the MPX connector. The connectors are just too small, so I bypassed them completely.
The servo out and GPS connector kept giving intermittent issues too and they now hot-melt-glued in.
I’d be wary of INS_GYRO_FILTER being at 32 (as per the guide) - I came to the conclusion it’s better left at 20 (or slightly less for larger frames) UNLESS you’re really good at manually tuning PIDs.
ATC_THR_MIX_MAN could be 0.5
PSC_ACCZ_P could be 0.2
PSC_ACCZ_I could be 0.4
AUTOTUNE_AGGR could be 0.1 and run autotunes again since it already flies OK.
Good to see the HNTCH enabled.
ATC_INPUT_TC is very high - is that OK like that? Usually we’d see around 0.2 for mid-sized frames and up to about 0.3 for larger frames… Is something else making this quad all twitchy that you tried to tune out with ATC_INPUT_TC ?
Wow thank you Shawn
As always very detailed analysis, exactly the sort of help I was after…thank you.
So first balance…ok your suggesting the rear right motor is working harder…will check the balance.
Vibes…we have a look. I just put 15 inch props on it and hate them. They are coming off tomorrow…assuming I can find my other props…my preferred.
Will pop a log on after my next flight.
Will disable the buzzer…they are useless at distance.
Not sure what your getting at around the VCC. I think your saying the connectivity (wiring harness) doesn’t cut it. So you soldered directly to the flight controller with the mpx in between the pdb and the fc. With a Capacitor in between to smooth stuff.
IF yo soldered to the FC…how and where did you do that
I will hot glue the connectors in place tomorrow as well.
AS for the INS_GYRO_FILTER I can change that to 20 tomorrow.
The other settings I will adjust.
The ATC input…not sure I changed it but will set it to .2 as you suggest.
VCC stuff worries me as I already had one crash due to power.
Not sure what you are doing with the Notch filter Ricky. You have it enabled, set for batch logging 3 IMU’s (you have 2), Throttle referenced yet no hover thrust value set. And the Batch logging Option of 3 I don’t think is a valid value (both Sensor Rate and Post filtering). Perhaps you should disable it while you work thru these issues.
You’re correct about wiring and connectors not cutting it. It was tricky to solder to test points on the underside of the PDB and test points on the underside of the Pixfalcon board. I couldnt get the connectors off without risk of destroying the board. I had to glue the wires in place so they dont rip off the near-microscopic test points. I actually joined to the original harness too, so the connectors + soldered wires are all in place at each end and join to the MPX connector at half-way.
This assumes that the PDB (or BEC) can actually provide the required 5.2volts DC (or there-abouts).
It might be easier to move to a different PDB or BEC.
That was just my particular situation and I mentioned the servo and GPS connectors as demonstration that the Pixfalcon has such small connectors - they’re about half the size of DF13.
MPX connectors were what I had available and frankly the gold plated contacts are going to be much more reliable than tin plated.
The FC could possibly work ok at the sub-optimal voltage, but by the time your 4.8 volts makes it up to the GPS (and whatever else is plugged in) it’d be getting too low to actually work, like 4.6 volts or less. I suspect it wouldnt take much (change of temperature, humidity, phases of the moon…) and the voltages will go downhill and lead to at least bad GPS or at worst a crash.
I’m focused on voltages and currents since these flying-rotating-knives depend on them 
Good points Dave - I must admit I neglected to look into the HNTCH values, I just noticed it was enabled.
Yea, not sure if it has anything to do with Ricky’s issue but invalid parameter values I suppose could be trouble.I thought something was odd when it took >2 minutes to load the the log in APM Planner!
Lately the new parameter are making APMplanner take ages to load a log - it doesnt understand them all and pauses and logs an error about each one.
Thanks all.
Not sure what this incorrect value is…if someone can let me know what it is I will correct it.
I get Shawn what your talking about. I will have a look into it.
I lost this once to power issues…I need to make sure I don’t again.
This is my task list so far
INS_GYRO_FILTER set to 20
Replace the props…balance them
Calibrate the esc to dshot
ATC_THR_MIX_MAN could be 0.5
PSC_ACCZ_P could be 0.2
PSC_ACCZ_I could be 0.4
AUTOTUNE_AGGR could be 0.1
ATC_INPUT_TC .2
Add capacitor to the power in to the FC.
Any other input is appreciated.
There are 3. You have batch logging enabled for 3 IMU’s and the FC only has 2. Enable only the 1st one. Then you have throttle referenced notch configured but 0 where you should have the MOT_THST_HOVER value set as reference. And for logging options you have the bit mask set for both pre and post filter logging. It’s either one or the other.
Thanks David.
Um but I think half of that went over my head.
So I have 3 IMUs being logged but there are only 1 on the FC…so I have to turn 2 off…ok…need to figure out how thats done. More reading to do.
the whole mot thrust thing…that lost me.
The logging I think I can figure that out.
Batch logging- Set INS_BAT_MASK to 1 to log the 1st IMU which is all you need.
Notch throttle reference: Set INS_HNTCH_REF to your MOT_THRST_HOVER value in your parameter file.
Logging options: Set INS_LOG_BAT_OPT to either 1 or 2 depending on what you are doing. Set to 1 to log sensor rate (pre filter) to determine FFT characteristics for setting the notch frequency and bandwidth. Set it to 2 after you have set these parameters and want to review the result.
When you are done configuring the notch filter turn off batch logging (0) so you don’t log huge files that you are not using the data for.
I think Dave means the board has 2 IMUs and you only need to log 1 or maybe 2.
Set this:
INS_LOG_BAT_MASK,1
If you’re going to dig into the depths then
INS_LOG_BAT_MASK,2 or 3
It depends if IMU 1 or 2 are better or worse with vibrations or you want to log both.
Edit: Dave beat me to it
Set INS_LOG_BAT_MASK, to 1.
Set INS_LOG_BAT_OPT to 1
Set INS_HNTCH_REF to the value for MOT_THST_HOVER,0.2308475
Ok I think I have it. Will implement the changes…thanks guys I super appreciate your input.
Yes, that’s right, Then make a simple flight Hovering in AltHold for a minute or so. Then review that log in the FFT analyzer. Or post it here we’ll check it out.
It occurs to me that this probably has nothing to do with your initial post. I got sidetracked looking at the parameter file and re-directed your thread!
Hey Dave.
Any feedback is appreciated. I am more of a hardware guy then a software one. So your support is really appreciated.
Tomorrow I will finish the power changes that Shawn called out…at least changes that may make a difference that remains to be seen at this point.
Also balance if the props…so will get the bird all updated then out to the field to test it out again.
Ricky @rickyg32,
This is what the FFT did for my quad when I followed Dave’s recommendations
PS. Excellent catch with that stabilize, though might have not been needed but you didn’t know that at that time, nice one!
Wow thats a big change to the vibrations. I hope to have the same result. I am glad he was able to help
@dkemxr @xfacta
Shawn, Dave thanks again for the input. I have now implemented all the config changes recommended.
I have moved the FC to a new BEC as the current one when I measured the non loaded voltage it was 5 vdc and the new bec is 5.26vdc no load.
I moved all ancillary equipment but the GPS to a separate bec. While there wasn’t a lot. Receiver and the external LED…but I am adding a lidar and they are a huge consumer of current so better its not on the fc.
My logic was that if I loose the fc bec I loose the bird, but if I loose the other bec I at least have an FC and GPS working to bring her home in RTL.
I will look at adding in the redundant power.
I just need to hot glue the connectors…thats later today.
Also I have balanced all the props.What a pain that was…but its done.
I think thats everything…just need to get to the field and do the test.
Thanks all…will report back after the next flight.