Why?
Has ANYONE who wrench tightened their props had them fly off yet? I have read and read and searched and searched and NEVER found a single report of a wrench tightened prop flying off.
Why complicate it with locktite?
URGENT: Tighten your props with wrench
- Thread starter yorlik
- Start date
Why?
Has ANYONE who wrench tightened their props had them fly off yet? I have read and read and searched and searched and NEVER found a single report of a wrench tightened prop flying off.
Why complicate it with locktite?
Frank,
When I asked ken about the potential validity of this SDK data, he had no qualms about it being as accurate as the variables reported. He was confident HIS app had no effect on the data. That said, you later said you felt you have rearranged the columns better and felt 95% confident on the data. I still have zero reason to believe the SDK presents wrong data for currentCurrent. I am awaiting dji to give their feedback. Meanwhile, with your 95% confidence report (sorry, could not find where you stated that) the data is identical to before. I flew today and the variable continues to report inverted polarity, and not just twos compliment numbers. I think we need to wait for dji feedback.
Isn't each quad independently reporting there own results real time via the SDK Frank?
Couldn't his bird be on the edge of a fly away "edge" while you test device is more middle ground.
Results reported are device independent.
Yorlik's reports are per his - which is not what you see.
Ala the difference.
Ala uncovering an issue!
Couldn't his bird be on the edge of a fly away "edge" while you test device is more middle ground.
Results reported are device independent.
Yorlik's reports are per his - which is not what you see.
Ala the difference.
Ala uncovering an issue!
So why not some of you folks who have logs POST them here? We can compare notes and see if mine is unique?
So I got some time to play with numbers some more. I took one of my latest logs, from one of the latest versions of UF, and added a few columns to show voltage and current difference from each 1 sec row recording to the next. I then used my battery initernal resistance value to 'correct' the rerported current values based on voltage droop each second. I believe it shows that the data is indeed valid, not corrupt, and definitely not the simple twos compliment overflow we were heading towards.
Based on the following, I would say anyone flying without wrench tightening their props is the same as driving a car without your seatbelt. Just asking for more trouble. But that is just my opinion and I am not telling others they have to tighten their props, just making a strong suggestion.
That said, here are my latest calcs. Do what you want with them.
I am not allowed to upload an xls sheet, so will post a link to it instead:
https://www.dropbox.com/s/dcj8u6xxof6zsq1/2015.03.28 15.45.37mk.xls?dl=0
Based on the following, I would say anyone flying without wrench tightening their props is the same as driving a car without your seatbelt. Just asking for more trouble. But that is just my opinion and I am not telling others they have to tighten their props, just making a strong suggestion.
That said, here are my latest calcs. Do what you want with them.
- Add delta V column to spreadsheet @AB to spreadsheet 2015.03.28 15.45.37mk.xls
- Add delta C column to spreadsheet @AE
- Calculate "expected Vdroop" column @AA
- based on (deltaC)*AA2 cell -> where I can play whatif and put in different ir values (internal battery total resistance, in mohms)
- Note: I previously calculated my ir on this battery as approx. 0.030 ohms. This value does not change much in these lipo batteries between 20-100% charge, and does not droop much in extra heavy current upto peak discharge rating of batt, so is valid to use.
- Next ASSUME two's compliment error plus other faults in currentCurrent value reported via SDK and calculate corrected Delta C @AF based on actual delta V and ir=30mohm (or present value @AA2).
- Add corrected Vdroop @Z showing Vdroop with corrected current based on actual delta V.
- Add blue correctedCurrentCurrent @AG.....
- this blue column should have REMOVED any erroneous SDK reported currentCurrent values with more realistic ones based on actual Vdroop shown.....
- Stand back and get a macro look at column AF to see how erroneous currentCurrent values play out......
- @time10968 into flight it appears a glitch went thru that caused motors to slow down suddenly when motors were drawing -13amps and suddenly switched to +9amps. This sudden 22amp swing conceivably is enough to unscrew a loose (weak hand tightened) prop.
- @time253162 into flight it appears a glitch went thru that caused motors to slow down suddenly when motors were drawing -17amps and suddenly switched to +22amps. This sudden 37amp swing conceivably is enough to unscrew a loose (weak hand tightened) prop. Looking at velocity Z column, it appears the vertical ascend also corresponds to this current change, climbing at 3m/sec, then during this current event it drops to 0.1m/sec, then next readings shows it goes back to 2m/sec rise. Ditto remainingPower data shows that when it did this quick slow down even during middle of fast rise that the battery was pumped back up - ie., charged up - by this sudden current reversal as the motors turned into generators and pumped energy back to it.
- This possible 37amp quick swing maybe high enough to be close to the level where the smart battery decides there is a short and shuts off - ie., battery shutting down and P2V+ falling out of sky issue.
- Much more analysis can be done using the other columns present for data confirmation, same as I did using voltage and currentCurrent data.
I am not allowed to upload an xls sheet, so will post a link to it instead:
https://www.dropbox.com/s/dcj8u6xxof6zsq1/2015.03.28 15.45.37mk.xls?dl=0
Last edited by a moderator:
Wow lots of opinions... I use throttle down when handing catching otherwise CSC I have never had an issue with tip over CSC always works flawlessly. I asked because I was wondering if CSC would possibly cause some sort of internal damage but I have gathered that if you're not a goon with the sticks CSC is just fine
I cant wade through your penchant for cutesy argot and love of acronyms to decipher what your point is
Ok, I get it you love the cutesy argot, the acronyms, and your charts but I cant grasp how powering into reverse threads can loosen the props. Can you in real terms explain how this phenomenon is manifested? Please forgo the minutiae.
If you go to the beginning of the thread, it's more on topic, at least for me. Here in the later posts they have gotten into some mind boggling mumbo jumbo.
It's like when professors sit around and chat about quarks while having tea.
No offence yorlik, or any of the rest of you who went way over my head.
Yea, azisbest, look at the start of the thread, it's all there.
Azi I will try to Layman what yorlik has described. Not being super technical will lead to some things being not 100% correct, but I am attempting to simplify.
What yorlik found was a large amperage reversal that he believes happens during high performance maneuvers. Left stock up to the stop for example.
They discussed if the information in the log was an anamoly of the software development kit used to create the log, or if the information was truly genuine. It could have been an artifact of the code given the large numbers involved. They believe the numbers are indeed genuine. If so, this could explain a lot of battery issues.
The motors spin in the direction that constantly tightens the props. Except when they don't. If you just sit the prop on the prop head the motors will screw the head into the props until they bottom out. At that point the props will soon the the direction of the motor spin. All is right in the world. What yorlik believes he had found is instances where the current completely reverses polarity stopping the motor and possibly spinning it reverse of its normal direction of spin.
There are two possible ways of prop loss. The first is the motor spinning at maximum revolutions stops briefly, the reversed current acting as a brake against the momentum of the prop, allowing the inertia of the prop to spin itself off the prop head. The second theoretical prop loss is the reversal lasts longer than just an instant and the motor actually spins in reverse therefore actively working to throw the prop.
After looking at all the presented material the small charging that is occurring during this time actually bolsters the first position. Instead of the circuitry reversing the polarity, the current is cut allowing it to freewheel and act as a generator.
I am not a scientist full time, but I do software development. I am not trying to advocate, I am trying to explain a complicated issue that, if proven, would cause us all a lot of concern.
How did I do azi? Yorlik?
What yorlik found was a large amperage reversal that he believes happens during high performance maneuvers. Left stock up to the stop for example.
They discussed if the information in the log was an anamoly of the software development kit used to create the log, or if the information was truly genuine. It could have been an artifact of the code given the large numbers involved. They believe the numbers are indeed genuine. If so, this could explain a lot of battery issues.
The motors spin in the direction that constantly tightens the props. Except when they don't. If you just sit the prop on the prop head the motors will screw the head into the props until they bottom out. At that point the props will soon the the direction of the motor spin. All is right in the world. What yorlik believes he had found is instances where the current completely reverses polarity stopping the motor and possibly spinning it reverse of its normal direction of spin.
There are two possible ways of prop loss. The first is the motor spinning at maximum revolutions stops briefly, the reversed current acting as a brake against the momentum of the prop, allowing the inertia of the prop to spin itself off the prop head. The second theoretical prop loss is the reversal lasts longer than just an instant and the motor actually spins in reverse therefore actively working to throw the prop.
After looking at all the presented material the small charging that is occurring during this time actually bolsters the first position. Instead of the circuitry reversing the polarity, the current is cut allowing it to freewheel and act as a generator.
I am not a scientist full time, but I do software development. I am not trying to advocate, I am trying to explain a complicated issue that, if proven, would cause us all a lot of concern.
How did I do azi? Yorlik?
Spiritskeeper, even though you didn't ask me, I think you did just fine.
That is what I thought happened in the previous thread:
http://www.phantompilots.com/threads/wth-just-happened-props-just-flew-off.36530/
about props flying off, which yorlik joined on March 16th. Thank you yorlik.
If I am not mistaken, the conclusions drawn to in that thread, led to this thread which may have actually found an answer to the question; wth just happened...
Ya guys gave me a headache reading that stuff, but thanks.
I like my prop wrench, and my dumb battery.
That is what I thought happened in the previous thread:
http://www.phantompilots.com/threads/wth-just-happened-props-just-flew-off.36530/
about props flying off, which yorlik joined on March 16th. Thank you yorlik.
If I am not mistaken, the conclusions drawn to in that thread, led to this thread which may have actually found an answer to the question; wth just happened...
Ya guys gave me a headache reading that stuff, but thanks.
I like my prop wrench, and my dumb battery.
It's interesting to note that DJI issued prop locks for the Inspire that prevents a prop from spinning off while in the air. Obviously, DJI knows of a situation that can back off their self-tightening props. I'm guessing they probably couldn't design locks that would work on a phantom. See the Inspire prop locks at 12:45 into this video.
I bet whatever new model DJI is about to release comes with prop locks.
I bet whatever new model DJI is about to release comes with prop locks.
Last edited:
What am I missing? Your ESC fails, you lose the prop and you crash. Forget the prop, an ESC fails and you crash. Seriously, if an ESC fails and I crash, missing a prop will be the least of my problems.
If that were the only reason for props unscrewing, absolutely! But alas, that is not so.
So do we if an ESC glitch happens. Proof is not only in the data (sorry for the minutia) showing battery current reversal, but also showing the battery voltage being charged as the motor dumps energy back into it, and the vertical climb rate dropping during the glitch also....
Does each motor have enough torque to unscrew a self tightening prop? I'd say yes:
https://www.dropbox.com/s/yhuporg3bzi4yxz/props calcs.ods - OpenOffice Beta Calc.jpg?dl=0
https://www.dropbox.com/s/yhuporg3bzi4yxz/props calcs.ods - OpenOffice Beta Calc.jpg?dl=0
Yorlik the minutia is where the details live. I am glad to see the thread back on track.
Steve the data is showing the esc doesn't always fail and cause a crash. The data shows that the esc glitch causes the prop loss and that results in a tumbling crash.
Steve the data is showing the esc doesn't always fail and cause a crash. The data shows that the esc glitch causes the prop loss and that results in a tumbling crash.
