Altitude Hold improvement solution

mahowik · Post by **mahowik** » Mon Nov 12, 2012 12:37 am

dramida wrote:And here is the video in witch the variation of altitude is controlled by throttle stick position

Many thanks for nice demo!

Only one note here. Althold point/zone - it's where it's activated, but not center position.

So I will add VARIO code to MultiWii_shared branch and will reduce pids to 15-10 because it seems little bit rapid/sharp for default behaviour...

update: code was added to MultiWii_shared branch

thx-
Alex

Crashpilot1000 · Post by **Crashpilot1000** » Mon Nov 12, 2012 10:35 am

I wonder if your solution will do a decent job on keeping the altitude during movement (real flying).

So long
Kraut Rob

dramida · Post by **dramida** » Mon Nov 12, 2012 1:28 pm

I tested in-flight altitude hold and in does a far better job than previous R1170. For me was good enough the altitude hold behavior during fast flight. I can see it compensate the throttle when leaning also.

Improvement: setting hower throttle to the point where Alt Hold is activated is strange , i prefer 50% throttle. (or at least a define)
Why? because i want to know it keeps its altitude even if it's far above and i can't say if it is climbing or not and after some ups and downs you may be loosing the alt hold spot on your remote.... (aerial photos)

I hope you get my point.
Mihai.

alexia · Post by **alexia** » Mon Nov 12, 2012 1:38 pm

i could only say you felicitation for all you hard working!
baro mode seem to be far better than original code.i am not a coder but i am very happy to see that this great project become more and more famous and better..and that is thanks to you.
i know you take your free time for this and i hope lots of people are reconnaissant(i don t know how to say that in english..lol.french word ).

diyboy · Post by **diyboy** » Mon Nov 12, 2012 2:21 pm

mahowik wrote:
diyboy wrote:
mahowik wrote:svn revision: http://code.google.com/p/multiwii/source/detail?r=1257
prepared archive: http://multiwii.googlecode.com/svn/bran ... _r1248.zip

Today I tested the r1248 with my 330frame X4 the BARO working fine.

Many thanks! I think you was first

I hope you are talking about testing "prepared archive" or "svn revision" from links above? i.e. not just r1248, because it hasn't latest changes...

Also could you tell more details or probably you have video how VARIO controlled altitude change works?

thx-
Alex

Hi Alex
I just tested the prepared archive "http://multiwii.googlecode.com/svn/branches/Mahowik/MultiWii_varioPID_based_on_r1248.zip".

My testing process: (Using the default vario PIDs)
1. Quadx fly to 5M high
2. Throttle at about 50% then active Baro, the QuadX hold on about 5M.
3. Throttle move up to 70%, the quadX goes up
4. Throttle return to 50%, the quadx hold on it's height
5. Throttle move down to 30%, the quadx goes down.
6. Throttle return to 50% the quadx hold on it's height

I also test Baro+GPS hold with about process.

But I haven't taken video.

DIY Boy

vpb · Post by **vpb** » Mon Nov 12, 2012 3:39 pm

Crashpilot1000 wrote:I wonder if your solution will do a decent job on keeping the altitude during movement (real flying).

So long
Kraut Rob

Hi Rob, I use Mahowik alt-hold 100% the time I fly arround, solid height holding, I just use left stick to yaw.

mahowik · Post by **mahowik** » Mon Nov 12, 2012 4:47 pm

Crashpilot1000 wrote:I wonder if your solution will do a decent job on keeping the altitude during movement (real flying).

In my 2.1_B[x] releases I tried to play with with solutions (my own and from alexmos)for throttle angle correction:
(+) it works for fast forward flight
(-) it make a big jumps (3-5m) when you change rapidly the direction of the flight

So we can introduce this feature with define to have a choice...

vpb wrote:I use Mahowik alt-hold 100% the time I fly arround, solid height holding, I just use left stick to yaw.

Actually the issue with reducing altitude exist when you start to move with inclination about >20degrees, but it's can be critical only on low altitudes...

thx-
Alex

Crashpilot1000 · Post by **Crashpilot1000** » Mon Nov 12, 2012 4:54 pm

@vpb @mahowik

Thank you for the info!
I guess i will have to test it myself

So long
Kraut Rob

mahowik · Post by **mahowik** » Mon Nov 12, 2012 6:11 pm

dramida wrote:I tested in-flight altitude hold and in does a far better job than previous R1170. For me was good enough the altitude hold behavior during fast flight. I can see it compensate the throttle when leaning also.

Improvement: setting hower throttle to the point where Alt Hold is activated is strange , i prefer 50% throttle. (or at least a define)
Why? because i want to know it keeps its altitude even if it's far above and i can't say if it is climbing or not and after some ups and downs you may be loosing the alt hold spot on your remote.... (aerial photos)

In first attempt the main goal was to check VARIO pid controller works or not. And of course now we can improve and add some features there

Let's discuss and sort out possible solutions, because i have too much points in my mind on this

Current implementation:
(+) when you activate the althold it will keep altitude immediately, i.e. it can be used to safe your copter during the activation
(+) initial throttle, i.e. throttle for hovering is taken as current throttle value on althold activation
(-) center is not 50% throttle, but if you need to have this in center, you can just reactivate the althold. I.e. just set the throttle stick to desired position and turn off/on the switch of althold quickly.
(-) as for now, predefined "throttle expo curve" also used for setting desired rate/vario and it's good if you activate the althold at the MID expo point (it's set in GUI and in right tuned config it equals to hover point, i.e. value of the initial throttle for hovering)
BUT if althold activated in point of throttle which in diff with MID expo point, it will be not usefull for vario regulation, e.g. MID_expo=40% and althold_activation=60%, it means ascending will be too fast, descending will be too low, because you are not in the center of expo curve...

Desired implementation:
1) center is 50% throttle
2) max rate/vario set with define
3) to have a beeps/light signaling during ascend/descend altitude OR best choice to have feedback/telemetry from flight controller to pilot (i'm experimenting with frsky downlink telemetry to have beeps on my TX)
4) need to predefine own expo curve with 50% center OR probably we don't need expo for rate/vario stick controlling at all, if all range about +/-3..5m/s?
5) (not obligatory but nice to have) safe function when althold deactivated... e.g. when you turn off the althold, throttle will be changed smoothly (in 2-3 seconds) from last althold's throttle to current throttle value to have time to react and adapt throttle after althold deactivation...

(+) proportional stick movement to have the same ascend/descend rate
(-) you need to tune initial throttle, i.e. throttle for hovering, because with this approach you should have possibility to activate althold with any thottle stick position. E.g. with #define INITIAL_THROTTLE_HOLD_FROM_MID_EXPO_POINT (Predefined initial throttle for AltHold will be calculated from MID (middle/hover) point of expo throttle from GUI).
But probable it has more advantage than disadvantage

(-) visible disadvantage - it will required to move stick to center after activation

Pls. provide you thoughts on this. I.e. we need to collect requirements before the next implementation step

thx-
Alex

scrat · Post by **scrat** » Mon Nov 12, 2012 8:39 pm

Hi guys.

If I download mahowik's zip file: http://code.google.com/p/multiwii/sourc ... _r1248.zip

Is it ok to use it or it's not good. Then which MultiWii Conf to use with it?

Thx!

dramida · Post by **dramida** » Mon Nov 12, 2012 10:09 pm

It works fine that release. One point to mention, i observed in this last version that arming the copter with yaw right in not always working.
Sometimes when i try to arm the copter, i hear a very short beep and green led from Crius AIOP blinks also shortly. Nothing happens after. After several atempts, copter arms. I checked all EPA of remote and the signals are right, between 1000 and 1100us for left and 1900 and 2000us for right yaw.
Needless to say that when i upload old R1177, things comes back to normal. Anyone has the same issue?

Issue: I was able to disarm copter on the ground in baro mode. So if you put throttle down, yaw left, even in Baro mode, copter disarms.

Another issue/bug not regarding Mahowik branch: I can't change the voltage prescaller in software so regardless what VbatSCALE value i wrote, the voltage shown in GUI stays the same. Am i missingng something? What it does VBATNOMINAL...is it really necessary? Do i have to erase eeprom to change the scale? Wouldn't be straight forward to have in GUI a box to enter the scale for voltage sensor (voltage divider) and current also?

Code: Select all

#define VBAT              // uncomment this line to activate the vbat code
    #define VBATSCALE     58 // change this value if readed Battery voltage is different than real voltage
    #define VBATNOMINAL   168 // 
    #define VBATLEVEL1_3S 140 // 
    #define VBATLEVEL2_3S 132 // 
    #define VBATLEVEL3_3S 125  //
    #define VBATLEVEL4_3S 120  // if vbat ever goes below this value, permanent alarm is triggered
    #define NO_VBAT       60 // Avoid beeping without any battery

mahowik · Post by **mahowik** » Mon Nov 12, 2012 10:14 pm

dramida wrote:It works fine that release. One point to mention, i observed in this last version that arming the copter with yaw right in not always working.
Sometimes when i try to arm the copter, i hear a very short beep and green led from Crius AIOP blinks also shortly. Nothing happens after. After several atempts, copter arms. I checked all EPA of remote and the signals are right, between 1000 and 1100us for left and 1900 and 2000us for right yaw.
Needless to say that when i upload old R1177, things comes back to normal. Anyone has the same issue?

Probably because of 3rd point from viewtopic.php?f=8&t=2371&start=320#p26473

i.e.:
3) protection from ARM if BARO activated

dramida · Post by **dramida** » Mon Nov 12, 2012 10:20 pm

mahowik wrote:
dramida wrote:It works fine that release. One point to mention, i observed in this last version that arming the copter with yaw right in not always working.
Sometimes when i try to arm the copter, i hear a very short beep and green led from Crius AIOP blinks also shortly. Nothing happens after. After several atempts, copter arms. I checked all EPA of remote and the signals are right, between 1000 and 1100us for left and 1900 and 2000us for right yaw.
Needless to say that when i upload old R1177, things comes back to normal. Anyone has the same issue?

Probably because of 3rd point from viewtopic.php?f=8&t=2371&start=320#p26473
i.e.:
3) protection from ARM if BARO activated

I think something works not as expected in this protection routine, i can assure i wasn't in baro mode and all i had to do was to try a few times and then armed...

mahowik · Post by **mahowik** » Mon Nov 12, 2012 10:26 pm

dramida wrote:I think something works not as expected in this protection routine, i can assure i wasn't in baro mode and all i had to do was to try a few times and then armed...

ok... thanks... I will re-check at home...

BTW what do you think on this viewtopic.php?f=8&t=2371&start=330#p26627?
we need to gather requirements

vpb · Post by **vpb** » Tue Nov 13, 2012 8:39 am

dramida wrote:It works fine that release. One point to mention, i observed in this last version that arming the copter with yaw right in not always working.
Sometimes when i try to arm the copter, i hear a very short beep and green led from Crius AIOP blinks also shortly. Nothing happens after. After several atempts, copter arms. I checked all EPA of remote and the signals are right, between 1000 and 1100us for left and 1900 and 2000us for right yaw.
Needless to say that when i upload old R1177, things comes back to normal. Anyone has the same issue?

I dont have that issues, but with new code, I cannot trim ACC :-|

, led doesnot flash each time I trim via the combo.

dramida · Post by **dramida** » Tue Nov 13, 2012 10:16 pm

mahowik wrote:
dramida wrote:I tested in-flight altitude hold and in does a far better job than previous R1170. For me was good enough the altitude hold behavior during fast flight. I can see it compensate the throttle when leaning also.

Improvement: setting hower throttle to the point where Alt Hold is activated is strange , i prefer 50% throttle. (or at least a define)
Why? because i want to know it keeps its altitude even if it's far above and i can't say if it is climbing or not and after some ups and downs you may be loosing the alt hold spot on your remote.... (aerial photos)

In first attempt the main goal was to check VARIO pid controller works or not. And of course now we can improve and add some features there

Let's discuss and sort out possible solutions, because i have too much points in my mind on this

Current implementation:
(+) when you activate the althold it will keep altitude immediately, i.e. it can be used to safe your copter during the activation
(+) initial throttle, i.e. throttle for hovering is taken as current throttle value on althold activation
(-) center is not 50% throttle, but if you need to have this in center, you can just reactivate the althold. I.e. just set the throttle stick to desired position and turn off/on the switch of althold quickly.
(-) as for now, predefined "throttle expo curve" also used for setting desired rate/vario and it's good if you activate the althold at the MID expo point (it's set in GUI and in right tuned config it equals to hover point, i.e. value of the initial throttle for hovering)
BUT if althold activated in point of throttle which in diff with MID expo point, it will be not usefull for vario regulation, e.g. MID_expo=40% and althold_activation=60%, it means ascending will be too fast, descending will be too low, because you are not in the center of expo curve...

Desired implementation:
1) center is 50% throttle
2) max rate/vario set with define
3) to have a beeps/light signaling during ascend/descend altitude OR best choice to have feedback/telemetry from flight controller to pilot (i'm experimenting with frsky downlink telemetry to have beeps on my TX)
4) need to predefine own expo curve with 50% center OR probably we don't need expo for rate/vario stick controlling at all, if all range about +/-3..5m/s?
5) (not obligatory but nice to have) safe function when althold deactivated... e.g. when you turn off the althold, throttle will be changed smoothly (in 2-3 seconds) from last althold's throttle to current throttle value to have time to react and adapt throttle after althold deactivation...

(+) proportional stick movement to have the same ascend/descend rate
(-) you need to tune initial throttle, i.e. throttle for hovering, because with this approach you should have possibility to activate althold with any thottle stick position. E.g. with #define INITIAL_THROTTLE_HOLD_FROM_MID_EXPO_POINT (Predefined initial throttle for AltHold will be calculated from MID (middle/hover) point of expo throttle from GUI).
But probable it has more advantage than disadvantage
(-) visible disadvantage - it will required to move stick to center after activation

Pls. provide you thoughts on this. I.e. we need to collect requirements before the next implementation step
Please check
thx-
Alex

The desired implementation is what I need. If anyone thinks different, let's hear a second oppinion plea
se.
About disarming in baro mode- bad things could happen thie way.
About arming in baro mode could be ok for your desired implementation.
Right now i have issues with arming the copter ...i wrote earlier about this.

mahowik · Post by **mahowik** » Tue Nov 13, 2012 10:53 pm

dramida wrote:Right now i have issues with arming the copter ...i wrote earlier about this.

I have not reproduced this issue with my AIO...

vpb · Post by **vpb** » Wed Nov 14, 2012 6:22 am

I've tried new alt-hold in 1268dev, it's ok, still not strictly test cause I cannot trim acc. But it's a bit unconfortable when I cannot arm with baro on.

vpb · Post by **vpb** » Wed Nov 14, 2012 7:01 am

+1 on feedback via frsky telemetry. I'm using that feature now and always keep my eyes on alt-info.

Shogun · Post by **Shogun** » Wed Nov 14, 2012 10:06 am

mahowik wrote:
dramida wrote:Right now i have issues with arming the copter ...i wrote earlier about this.

I have not reproduced this issue with my AIO...

I have the same arming issues and need to do this via switch. Have this since r1240, r1177 was OK.

vpb · Post by **vpb** » Wed Nov 14, 2012 11:13 am

Hi, maybe you should hold the arming-stick longer than before. That way I can trim ACC now.

vpb · Post by **vpb** » Wed Nov 14, 2012 11:18 am

@Mahowik: About % throttle at the hover point, how about hard-coded mid RC throttle value? Will it be more precise than percentage? With frsky telemetry we can read the throttle value at hover point.

dramida · Post by **dramida** » Fri Nov 16, 2012 8:36 am

Hello, I noticed that latest MWC zip file that i previously tested it was removed from Mahowik branch due to some unresolved "issues". Do we have anything new to test this coming weekend?

vpb · Post by **vpb** » Fri Nov 16, 2012 9:14 am

On r1271, mahowik has rolled back with log message: "rollback changes as not stable yet...".
I'm still flying ok, hello Mahowik?

wilco1967 · Post by **wilco1967** » Sat Nov 17, 2012 9:00 am

no issues here also on version just before rollback..... flies just fine (better then ever before in fact...).

What where the issues ?

dramida · Post by **dramida** » Sat Nov 17, 2012 1:16 pm

The altitude control works very good, the only issues where about arming/disarming that could be inherited from main trunk.Specifically, on Crius AIO PRO board, quad mode with camstab, buzzer, GPS, Vbat- enabled, i couldn't arm the copter from the first time, i had to try several times. Also in Altitude control mode, i was able to disarm the copter with yaw right.
But i repeat, the main objectives where achieved, meaning a better altitude control. I would be very happy that MWC 2.2 prerelease could include these new features

dramida · Post by **dramida** » Mon Nov 19, 2012 9:51 pm

I don't get it:

After successfully implementation of altitude variation control by Mahowik almost two weeks ago, the shared trunk code remains the same. Let me remind you how nice the altitude control works, here is the video:

http://www.youtube.com/watch?v=7Sa-oyJ4Ljs

And in case the code is no longer availabe on Mahowik branch, i attached a zip file with the MWC source code you see flying in video above.

jevermeister · Post by **jevermeister** » Mon Nov 19, 2012 10:26 pm

hi,
there is a list of people who can alter code over at the google code repository. I looked into my mqilbox today and found no mail regarding this.
I did not follow this thread closely enough to feel adressed to upload this code.

I would suggest to contact one of the devs to merge the code. How should those guys know
about that. There was no command to do so.
Do you expect them to sit and wait for every idea and development to be ready to merge?

please contact one of the devs directly and wait for negative or positive feedback.
I can only speak for myself here: I always have a hard time merging code that is to complicated for me to understand. I could fuck up the code without knowing why and how should I fix thise bugs...

Nils
ps.: I do not want to rant, just want you to understand some devs. ;-)

pps.:I like your fantastic work - If only I could understand these algorythms :-\

vpb · Post by **vpb** » Tue Nov 20, 2012 7:03 am

dramida wrote:I don't get it:

And in case the code is no longer availabe on Mahowik branch, i attached a zip file with the MWC source code you see flying in video above.

Hi dramida, if you want a merged version that Mahowik did on the main trunk, just checkout the version before he removed it

Code: Select all

svn co http://multiwii.googlecode.com/svn/trunk/MultiWii_shared -r 1268

And mahowik branch, with the zip file

Code: Select all

svn co http://multiwii.googlecode.com/svn/branches/Mahowik -r 1268

It's always there.

vpb · Post by **vpb** » Tue Nov 20, 2012 7:27 am

jevermeister wrote:I can only speak for myself here: I always have a hard time merging code that is to complicated for me to understand. I could fuck up the code without knowing why and how should I fix thise bugs...

You merge the code manually by hand? without the confliction code, you can always merge code by command.

jevermeister · Post by **jevermeister** » Tue Nov 20, 2012 10:44 am

I am not merging by hand but I want to understand the code I contribute. I wanted to say, that I have a bad feeling everytime I commit something for someone I do not fully understand.

nils

mahowik · Post by **mahowik** » Tue Nov 20, 2012 5:42 pm

Hi guys,

Sorry for delays... My plans was little bit changed...
To be honest I did enough for the mwii project and comunity and it's time to do something for myself...
So now I'm trying to copy naza behaviour related to AH functions and altitude ascend/descend and going to distribute (individually) the firmware in *.hex format with flash tool/script...

I hope for your understanding...

thx-
Alex

copterrichie · Post by **copterrichie** » Tue Nov 20, 2012 5:47 pm

I am not surprised, wish you well.

mahowik · Post by **mahowik** » Tue Nov 20, 2012 5:52 pm

what you are not surprised?! You know me very well?

copterrichie · Post by **copterrichie** » Tue Nov 20, 2012 5:53 pm

I am not surprised at your course of action, nothing else to say.

alexia · Post by **alexia** » Tue Nov 20, 2012 6:23 pm

mahowik wrote:Hi guys,

Sorry for delays... My plans was little bit changed...
To be honest I did enough for the mwii project and comunity and it's time to do something for myself...
So now I'm trying to copy naza behaviour related to AH functions and altitude ascend/descend and going to distribute (individually) the firmware in *.hex format with flash tool/script...

I hope for your understanding...

thx-
Alex

thanks for you hard working.
i understand if you want working for yourself.you are already made lots of for the communauty
thanks again
alexia

vpb · Post by **vpb** » Tue Nov 20, 2012 6:29 pm

You've done a big improvement Mahowik, I guess dramida will miss you much

. Have luck with your plan!

kataventos · Post by **kataventos** » Tue Nov 20, 2012 6:47 pm

What a f

k is happenning here? :shock:

Can someone explain?

alexia · Post by **alexia** » Tue Nov 20, 2012 7:06 pm

nothing..just mahowick want to use is free time for other things
i hope that he doesn t move out of this incredible project

kipkool · Post by **kipkool** » Tue Nov 20, 2012 7:14 pm

That's a pity but it's your choice of course.

As Multiwii is GPL, binary distribution without source is not an option, does this means you will develop a new flightcontroller code from scratch ?

mahowik · Post by **mahowik** » Tue Nov 20, 2012 9:24 pm

kipkool wrote:As Multiwii is GPL, binary distribution without source is not an option, does this means you will develop a new flightcontroller code from scratch ?

nothing decided exactly yet... but in any case thanks for the info...

NikTheGreek · Post by **NikTheGreek** » Tue Nov 20, 2012 9:27 pm

Probably you will not count my opinion..

But ..... there is any chance to reconsider your decision ?

Polleke · Post by **Polleke** » Wed Nov 21, 2012 8:46 am

Hi Mohawik,
Thanks for the great work you've done

Wishing you all the best.
Hoped you'd say: "I'll be back!"

timecop · Post by **timecop** » Wed Nov 21, 2012 9:18 am

mahowik wrote:Hi guys,

Sorry for delays... My plans was little bit changed...
To be honest I did enough for the mwii project and comunity and it's time to do something for myself...
So now I'm trying to copy naza behaviour related to AH functions and altitude ascend/descend and going to distribute (individually) the firmware in *.hex format with flash tool/script...

I hope for your understanding...

thx-
Alex

lol, naza.
will you add the magic wobble too?
can't wait!

dramida · Post by **dramida** » Wed Nov 21, 2012 9:51 am

Mahowik, move your ass back here and stop fooling around with binary distribution nonsense witch contradicts the GPL licence. If you are not working in a team, you'll end up wasteing your time.
If you want to do some good money with Multiwii, offer technical support for local/global users and/or do aerial pictures or video ( i do the same here http://www.fae-drones.com ). Open your own topic/branch with a big donate button on first post also will help. For more marketing strategies i would have to charge you

dr.tom · Post by **dr.tom** » Wed Nov 21, 2012 1:12 pm

don't kill the messenger,
I know that this my video is not taken by multirotor,
but just wanted to show how accurate can an old BMP085 sensor be, I still can't believe it.
on my multirotors I have MS5611 and although it does work better than BMP, it's still no 'Naza like' alt. hold.

watch the second half of the movie, on ground it says 0,when raised it says 2m, no drift and accurate,
in flight it can maintain altitude +/-1m even in 10km long run, no drift!
http://www.youtube.com/watch?v=zbk3ipTg ... 9A7DD761A7

It must be the code that makes all that difference, if someone is willing to dig deeper...

this OSD http://www.himodel.com/FPV_Telemetry/CY ... W_GPS.html
author of software is: Extraline http://www.rcgroups.com/forums/member.php?u=72169

(alt is not maintained by gps, I have also previous version 'Nova osd', it used gps for that and had a lot of oscilations, this uses baro for that)

copterrichie · Post by **copterrichie** » Wed Nov 21, 2012 1:38 pm

Within any given group of people and there is no exception here, there is going to be differences of opinions and personality clashes if GROUP THINK is not at play. There will even be a few trolls :mrgreen:

, so lets not allow this fine outstanding group accomplishment to become tainted by harsh feelings. Now get back to work, all of you!!!

Groupthink is a psychological phenomenon that occurs within groups of people, in which the desire for harmony in a decision-making group overrides a realistic appraisal of alternatives. Group members try to minimize conflict and reach a consensus decision without critical evaluation of alternative ideas or viewpoints. Antecedent factors such as group cohesiveness, structural faults, and situational context play into the likelihood of whether or not groupthink will impact the decision-making process.

http://en.wikipedia.org/wiki/Group_think

kataventos · Post by **kataventos** » Wed Nov 21, 2012 2:53 pm

dramida wrote:Mahowik, move your ass back here and stop fooling around with binary distribution nonsense witch contradicts the GPL licence. If you are not working in a team, you'll end up wasteing your time.
If you want to do some good money with Multiwii, offer technical support for local/global users and/or do aerial pictures or video ( i do the same here http://www.fae-drones.com ). Open your own topic/branch with a big donate button on first post also will help. For more marketing strategies i would have to charge you

kataventos · Post by **kataventos** » Wed Nov 21, 2012 3:04 pm

dr.tom wrote:don't kill the messenger,
I know that this my video is not taken by multirotor,
but just wanted to show how accurate can an old BMP085 sensor be, I still can't believe it.
on my multirotors I have MS5611 and although it does work better than BMP, it's still no 'Naza like' alt. hold.

watch the second half of the movie, on ground it says 0,when raised it says 2m, no drift and accurate,
in flight it can maintain altitude +/-1m even in 10km long run, no drift!
http://www.youtube.com/watch?v=zbk3ipTg ... 9A7DD761A7

It must be the code that makes all that difference, if someone is willing to dig deeper...

this OSD http://www.himodel.com/FPV_Telemetry/CY ... W_GPS.html
author of software is: Extraline http://www.rcgroups.com/forums/member.php?u=72169

(alt is not maintained by gps, I have also previous version 'Nova osd', it used gps for that and had a lot of oscilations, this uses baro for that)

What is so yuuupi on that? This is how multiwii ALT code is working with this OSD.

http://youtu.be/QQKaCK15iKQ

dr.tom · Post by **dr.tom** » Wed Nov 21, 2012 4:13 pm

nice,
but I am talking about 'alt hold', not 'alt display'.
in GUI my MWii also responses to 10cm change, nice to watch, but in air it's not so good. drifts by time...

...back to your OSD
how does it fly?
it's under development as written in video,
this is tested and flying as we speek. and controls altitude very good.

that's why I linked author of it so if someone has skill to look into his code(which works on common sensor used in mwii)
and see what could eventually be applied to mwc code.

I like the plug in option of your OSD, because that saves cable clutter, weight and price
(no need for double sensors onboard)

keep up the good work