Hello, I have some trouble finding the right MOT_THST_EXPO for my MAD M6C12-170kV arm set with CB2 22x7 propellers. Here is my graph at expo 0,7:
Here is excel file Kopia ArduPilot Motor Thrust Fit - Google Sheets
Is there a parameter to rise the curve? How can I know that the fit is good enough?
I would greatly appreciate your help, thank you Marek
Our normal formular gives MOT_THST_EXPO,0.72 for 22inch props, so 0.70 is quite close. I would just go with 0.72 and it will be fine.
The ESC requires:
MOT_PWM_MIN,1100
MOT_PWM_MAX,1920
Worst case scenario is the ESC is similar to a T-Motor Alpha and you have to set MOT_THST_EXPO,0.60 but leave that alone unless you have particular instability issues after some tuning. Those circular MAD ESCs look completely different to the T-Motor Integrated Alpha ones so you should be OK.
Hello!
Dear MR. Xfacta, I wholeheartedly disagree.
First of all - we are not trying to make something that barely flies and is prone to failures becouse of f.e. wind gusts. We are making UAVs which are reliable enough to fly over cities in a safe, repeatable, manner.
Therefore a statement that the value 0.6 should be fine when the normalized thrust curve and the measured thrust curve is not even close to each other is not a satisfactory answer to our problem. We need a way to make both of this curves align as closely as possible to later tune the filters, PIDs and the rest of the settings in a way that is as close to impossible to fail as possible.
I see a few solutions to an issue presented by an author.
First:
Introduce a point when one function of an expo ends and another one, with a different value, starts. Simillar to TPA breakpoint in betaflight.
Second:
Add the possibilty to change the value of an endpoint for the mot_thst_expo function by adding a defined by user ammount of kilograms to the ammount of corrected thrust at PWM 2000 which will be scaled down up to value of 0 which would be corelated with the PWM value of 1000.
Third:
Rewrite the mot_thst_expo function in a way that would allow the user to directly input the kgs of thrust created per arm while a certain PWM is sent to the ESC with the linear increase of the value between the defined points.
I personally think that the third option would be the best solution, unfortunatelly we do not have any programmers on board and hence we are kindly asking our dearest community for help.
I think you misunderstood some of what I said.
And I’m not sure I understand where you are coming from, as if I am advocating for unreliable copters ??
Can you say what it is that I am so wrong about?
I advocate for using the normal/calculated MOT_THST_EXPO,0.72 for 22inch props. This will at least be a good starting point.
It’s a fact that the MAD integrated ESC requires:
MOT_PWM_MIN,1100
MOT_PWM_MAX,1920
The bit about expo of 0.60 and T-Motor ESCs may not be useful, but something to keep in mind if there are still some particular instabilities seen during ascent and descent - the point being that the expo can be dependant on how some ESCs work too.
EDIT:
Are you disagreeing with the use of the spreadsheet tool to map thrust and calc a thrust expo?
I can partially agree with you there, as there is some special points to be aware of when using that method that can easily make your calculated thrust expo quite wrong.
For a start, people are almost exclusively measuring static thrust, and often include a section of data right down in the low RPM region that is effectively useless.
As @xfacta said:
The part of the curve that needs to match is the region around the hover thrust and around the thrust at cruise speed
Low side of the curve is much, much less important.
The three points you list each have some merit.
Your Second point is effectively what we are doing normally.
Your Third point, while maybe more correct or ideal, it’s not very practical.
So what we have is a formula that gives us a reasonable thrust expo value based on prop size (in inches).
thrust expo = 0.15686 x LN( prop_size ) + 0.23693
This formula is used in the Initial Parameters calculator in MissionPlanner, and also the Methodical Configurator.
Dear XFacta,
My angle comes from the analysis of your responses in a lot of previous topics. I don’t speak much here, I’m sorry for that but specialising in tuning I have quaite a lot of work and quite a small amount of time, therefore usually any discussion done outside of the private conversations with people whos know how I have previously veryfied as beeing on high level. It does not change the fact that I am in this industry for the last 16 years and I read A LOT. During this time I have read a lot of your responses, in which you were showing the logs of UAV that lets just say were not flying flawlessly as anything I’d deem as reliable. That is why I said what I said. Of course I base this statement on my memory, which in fact can be wrong, and if my statement causes you any discomfort - I am more than willing to delete any parts of it that you’d deem as necessary.
As for starting point - I don’t fiddle around, I don’t just try inputing semi random things and observing the resoults. I use scientific method. Calculate the resoults before implementing them and if I am wrong - I update the models I use for calculating whatever I need after analysing the root of the problem.
It doesn’t change the fact that I messed up here and you are, in fact, very correct about you statement providing the info about the min PWM and max PWM. After inputing this data to the mot_thst_expo caculation sheet I was able to reach the conclusion that the proper mot_thst_value for this motor/prop/ESC combination is, in fact, 0,84. I am basing this resoult on the new thrust test, which
I will attach as a screenshot, made using a battery pack capable of providing more current.
“Are you disagreeing with the use of the spreadsheet tool to map thrust and calc a thrust expo?”
I don’t disagree as I use this spreedsheeet religiously basing the resoults on what this spreedshit allows me to calculate.
With your experience I’m fully prepared to learn from anything you have to offer.
I can certainly agree with you that measuring and calculating the thrust expo for your components is the right thing to do. The formula we developed suits most props, motors and ESCs, but may not be suitable in all cases.
I cant say I’ve posted logs of copters that dont fly well, at least not a lot. Usually I’m looking at others logs. Like many others I would have posted some logs that I needed help with, but who hasn’t.
