Harakiri aka multiwii port to stm32

e_lm_70 · Post by **e_lm_70** » Wed Jun 18, 2014 9:48 am

strips wrote:I do not know how the PH code works. But tuning in MW / Baseflight enables you to tune GPS versus acc effect.

So depending on how often the code reads / processes the GPS updates you could turn down dependence on accelerometer and rely more on GPS. Tempted to try the Ublox 8 with compass. To test on Harakiri, Baseflight and APM.

18Hz ... can also overload the GPS parsing code ...it also may need more then 115k serial speed.

Anyhow ... even at 5hz update ... the accelerometer filtering in theory does not need much work

Point is that all test code assume noise in the GPS signal and do an heavy filter down ... yes , at 18Hz a filter down will make the "delay" seen by Mr Fiero less visible ... 20 samples filter will be less then 1 seconds, then 4 seconds ... so quite a difference in the "GPS lag due to filtering"

Anyhow, for how it is designed the PH, mainly what matter is a very accurate GPS signal ... 5Hz is more then enough.

Harakiri could be customized in the PH ... APM is a spaghetti code that nobody want to touch ... except the cooking people from 3Drobotics

I expect the price of Neo 8 to be only 20$ in few months ... not going to buy it now from the rich Germans :mrgreen:

timecop · Post by **timecop** » Wed Jun 18, 2014 9:56 am

e_lm_70 wrote:Nice ... so soon NEO 8 and MAX 8 will be available at cheap price, so far these modules are made in Germany.

I like this one: http://www.ebay.com/itm/Ublox-MAX-M8Q-G ... 1489476543

That's CSG_EU, well known rcgroups scammer.
Made nowhere near germany, sometimes ships from latvia, sometimes from who knows where. Usually doesn't ship at all.

strips · Post by **strips** » Wed Jun 18, 2014 9:57 am

Right. Wish I had more time to dig in the code and figure things out. Kids do take too much time :-)

Wondering how Baseflight (Mumtiwii) is compared to Harakiri regarding GPS updates and filtering.

e_lm_70 · Post by **e_lm_70** » Wed Jun 18, 2014 10:27 am

timecop wrote:
e_lm_70 wrote:Nice ... so soon NEO 8 and MAX 8 will be available at cheap price, so far these modules are made in Germany.

I like this one: http://www.ebay.com/itm/Ublox-MAX-M8Q-G ... 1489476543

That's CSG_EU, well known rcgroups scammer.
Made nowhere near germany, sometimes ships from latvia, sometimes from who knows where. Usually doesn't ship at all.

Thanks for the info ... still eBay and PayPal ... are ultra safe, even if the guy try to cheat.

It sounds you have GPS experience after all :geek:

... so ... maybe one day you will officially support GPS on Naze32 too, even if you can sell them now like hot cakes without the need to bless anything

Ups ... maybe not ... I see the guy sell also some 100$ 10DOF that are equivalent of GY-80 (9$ module) : http://www.ebay.com/itm/HMC5883L-BMA180 ... 0766408107
LOL

So ... yes ... better to stay away from this seller ... and possibly you did not find him because of his GPS module

PS: As I write, just testing in the garden my stock NEO6M, stock NEO6M with big antenna suggested by Mr.Fiero .. and then gotta test MAX7
Running 10min sitting on the garden table outdoor .... logged via arduino open log .. power by 1S LiPo ...
The one that will keep the best stable position ... it the winner :mrgreen:

Not sure this is the right way to test GPS ... but it is a lazy way to check them ... on a multicopter with Video TX, Telemetry Rx ... the noise could make much more different results ... but ...

EDIT: WOW

... he sold 260 junky 10DOF for 100$ each .. these cost less then 10$ each ... this is really a good biznes :mrgreen:

...

e_lm_70 · Post by **e_lm_70** » Wed Jun 18, 2014 12:54 pm

Ok ... just a little finding about GPS ... tested while sitting on the table outdoor in a static position.

Testet
- NEO 6M ... with stock Antenna, and with Big antenna (see eBay link on a previous post)
- MAX 7M ... with stock (unchangeable) antenna.

And the winner is -- > MAX 7M

NEO 6M passing from stock to big antenna did win 1 full satellite more in average ... from 6.7 to 7.9 ... MAX7 has in average even a bit more satellite in view .. 7.99
Interesting, the number of sat in view has been never more then 8 in the test ... also never below 6 except for few rare occasion with stock NEO6 .. data are used after around 200sec from start up. I also stat the data of the last 200sec of recording.

What is a bit odd is the checking the average speed ... I would expect that the best GPS is the one that report the lowest average speed, and the lowest MAX speed ... odd enough the MAX7 report 3 time average and max speed compared to NEO6 .. maybe something is wrong in the configuration, since I did push the NEO6 configuration on the MAX7

Anyhow ... the verdict is about who reported accuracy ... NEO6 stock report 1.5m average accuracy ... NEO5 big antenna report 1.0m average accuracy ... while MAX7 report 0.92m average accuracy.

Final and maybe more important .. is how LAT LONG change being on a stationary position.
Both did a good job ... error is only in the lowest digit or LAT LONG, and it is just +/- 1 ... with few exception of 3 unit gap between max and min.

So ... average error for
Stock NEO6 -> 0.00086 (this is Lat error + Long error average) (note the GPS accuracy is 0.001 there is no 0.000x in the NMEA format)
BigAnt NEO6 -> 0.00082
Stock MAX7 -> 0.00058

So, the big antenna helped to get more satellite but at the end did not drastically improve the accuracy.
MAX7 with a stock antenna is doing better

Finally ... quite disappointed with the error over the Altitude reported by all these GPS ... I see +/- 2.5m error, except for Stock NEO6 doing +/- 3.5m error

Somewhere I was reading of people having better altitude from GPS ... still quite useless the GPS altitude for fly low ... but for a fly high airplane the baro is not really mandatory I would say ... but with 5611 having 0.1m accuracy ...

Anyhow ... sorry for the long an bit off topic post ... but this clarify a bit a comparison between NEO6 and MAX7 ... MAX7 win hands down I would say ... even with stock antenna. MAX7 cost 17$ shipped, more or less the cost of a big GPS antenna.

I have also a MTK on my x900 copter, a FrSky mini GPS from telemetry (not sure if is a MAX6 or a NEO6) .. and maybe a LEA5 somewhere .. plus some other from different devices that I could hack away ... but ... anyhow ... NEO6 vs MAX7 sound interesting to me.

PS: Test has been done one after the other .. with a mini gap between NEO6 and MAX7 due to the discovery that my MAX7 lost my configuration ... for some strange reasons :oops:

... so after 10 min of garbage in the openLog I had to reconfigure and repeat the test ... weather was nice a clear sky ... location is near couple of big trees blocking some sky view 60deg in one direction .... vertical sky view was free.

Mr-Fiero · Post by **Mr-Fiero** » Wed Jun 18, 2014 5:28 pm

You can adjust your PDOP and HDOP mask with the larger antenna. more SV's will not ensure more accuracy but it will allow better signals which means you can reject better with settings in the MAX7. Also try changing your angle to horizon setting (cant remember exact name) to maybe 8 or 10 degrees and try the larger antenna and do your same test. This wont help with altitude reading though. I do find it weird you cant get more than 8 SV's. On a good day with my larger antenna I can see up to 12 at one time. On a bad day, with thick clouds/rain and between my trees I will see 6 to 8 SV's. Are you sure when you test you do it without any RF TX's running? Just a thought but maybe you have a harmonic to worry about, and if you do, the larger antenna can work against you and you will see worse results. There is also the slight chance if you have a strong Wifi device running while you test, the second harmonic could affect your GPS also. Its only slightly I find, but it will play more against an antenna that has more gain.

How are you measuring your average error? do you have a known fixed point your comparing against? if your reading the average error in the Ucenter it wont be right to real world because of different algorithms used in the averaging.

Nice testing. I like seeing your results.

I doubt anyone can get altitude better. always an issue with GPS's. The only way is if you can receive a ton of SV's at all angles of the horizon that have excellent orbit status, and are in good health. It will never compare to a baro.

When I get the replacement MAX7 I will re-post a new config, As I think some of the 7 series optimizations were lost in my uploads to the 6 series.

e_lm_70 · Post by **e_lm_70** » Wed Jun 18, 2014 10:50 pm

Yes max 8 sat is a bit strange, maybe is the location ... I always get better result at flying fields

I tested with 1s lipo to power both OpenLog (a atmega328 chip running 16mHz with a microSD card) and the GPS
This OpenLog sucking around 2mA should not make much rf noise .. my big pc was left all time back indoor, far from the gps .. and as well, I did not use u-center at all

Test was in the backyard, there is still some wifi signal both from my router and from couple of neighbors ... but can't be that much

I did not had any TX or RX noise around, nor I did use the base diy shield that I did use on my copter recently for my video on multiwii

About Lat - Lon error, I did use the error from the average result for the same gps module

So .. just calculated the average and then the classic formula ( (lat-latAv)^2 + (lon-lonAv)^2) ^ 0.5 .. all done over an xls ... I just imported directly the nmea txt in xls

I did exclude the first 100/200 sec of nmea log ... and I did calculate the data based on 9000 samples (around 15min data) .. and also the last 1000 sample ... so the last 100sec

About PDOP / HDOP .. this sound an alien language to me .. I just import your gps configuration , and just change back the gps serial speed, and as well filter out only the two nmea messages needed at code level: rmc and gga

I did push thw same config on the max7 ... but here I was getting some errors too ... I think there is some incompatibility ... also at start up my max7 report to be a prototype version ... with a quite old firmware inside ..

Anyhow, I don,t regret to have took the big antenna ... it did definitely improve the NEO6 ...

I guess soon you will get your own NEO7 ... you get already outstanding result, it will be interesting to see if you can get even better results

Mr-Fiero · Post by **Mr-Fiero** » Wed Jun 18, 2014 11:24 pm

If the configs are from other Ublox's, if a message is not recognized, the engines will just ignore the settings. I have never crashed a Ublox uploading and mixing configs. So you probably dont have to worry about that. errors while uploading will not harm the GPS, but they might however be missing some settings. Well, I'll Upload a 7 series config as soon as I get my GPS. Its the same settings, just slightly different in the cfg files but once I set the new GPS up, I can read it and save it so its more compatible.

I have sooo many GPS's, half have the larger antenna. I have 2x MTK (damaged), 4x NEO6's, 6 MAX7C's, 1x LEA6, and a NEO7 almost here, maybe today..... I find the ones that I upgrade to the 35mm antenna holds position tighter every time. The 35mm antenna's startup lots faster in less than ideal environments.

I wonder if your just seeing jitter/spikes from the GPS. I get it too but for the most part because of the averaging, it does not affect my flight. I think you might want to consider this. A GPS is not that accurate, but they do pretty good when everything is relative. What you have to look for in GPS is minimal delays (MTK is high compared to Ublox) . Then the averaging algorithms need to be fast updates, but as accurate as possible (these work against each other). And when you have a balance, the flight controller needs these position updates as fast as possible, and then it has its own delay issues to predictively try to fly. The next real important thing is then the number of SV's. If you can maximize this, then you have other options to refine "better" GPS data via config of the GPS engine. I can make a GPS really accurate, but when I do, I knock the "bad" SV's with masking, and other variables. But it makes it harder to get a fix, and hard to get a good number of SV's. (skyview becomes extremely critical) But when I do, it outperforms the specs of the GPS.

So, your measurements are really nice, but I have to stress they might be misleading as its not real world in the right application. You should test on your multicopter. What is nice, is from your testing you didnt notice anything weird. Check with Ucenter and see if the larger antenna helped signal strengths. If it did, its better, period.

What you need to look at, is quality of signals, strength of signals, number of SV;s and delays of position updates. And, of course, send the right information to the flight controller (messages). If you add any delays, say like lower baud rates, this is bad because, well these 32bit controllers handle higher bauds just fine. even if a packet had to resend, its still less delays than if you send slower baud. Everything has to be considered for such a dynamic application.

Maybe your GPS is doing better than you think. From your testing it sounds pretty good. All except I do wish you could get a sat or two more.... Try with 8 but let your GPS settle for 5 minutes on warmboot.

Mr-Fiero · Post by **Mr-Fiero** » Thu Jun 19, 2014 12:05 am

I had a thought.... the easy way to look at it, is you want relative, but very accurate position updates, with the smallest delay possible, over shortest term possible. Not over long term.

With that in mind, there is so many variables that come into play and if one were to test this, it would have to be with two identical GPS units at the same time, One as the fixed station and then you would have to compare between the two units. Believe it or not, that is the easy way, and the only way to get an idea. Even still, there could be slight multipath issue differences between the two.

Thats why when Peterkins wrote this, it means more than it sounds like....

by Peterkins » Sun Jun 08, 2014 5:34 pm

yes barking - but what?
the bird is up in the sky. you may encounter spots where the signal strength gets weak.
you get gps glitches when the constellation the satelites do change.

in the end you wanna look at raw gps meassges - keep an eye on the rtk lib.

He is right, and its a very complicated answer.

But I try to keep it simple, and really trust me, if your signal strengths are better, than it will perform better. Also newer generation GPS's like the 7 series will be better. For allot of reasons. You can also test to see and make sure your signals stay strong when you activate your other RF devices.

Once you have maximized your GPS, then the next focus is the settings in Harakiri. There is also a chance that its over reacting, or not reacting properly for your rig. Drifting however usually leads me back to the GPS, but again, look short term first.

e_lm_70 · Post by **e_lm_70** » Thu Jun 19, 2014 12:08 am

My comparison was pretty academical

I'm not after perfect gps hold .. +/- 2 or 3 meter is more then fine for me

My fly tests with NEO6 (after your help, even with small stock antenna) with my hacked simple multiwii was perfecrly fine to me

Actually .. I did cut away all the filtering of gps in my test, at multiwii level ... the copter take some time on its own to move, so ... a spike inside 200mSec will not be seen even without filter.

I did even took away other stuff too , and I could not notice any difference.

I like to have a more simpe code if possible .. and my test on KK2 using multiwii with just 32k flash was on edge

I want to move soon to harakiri, I should have maiden already .. anyhow ... for me my test was to prove that the good result got on neo6 can be repeated on max7 ... and for what I see max7 should do even a better job.

Clearly, with all the rf noise of a flying copter with possible video tx and telemetry, bluetooth, etc ... each copter and each gps may have his own 'story' ... the shield is looking to have made miracles on my old copter used in my multiwii kk2 hacking ... on other copters I may need to be more creative, or hope max7 is less rf noise sensible ... I want to put harakiri on a 250class copter with fpv gears, and frsky telemetry ... this is my end goal

Mr-Fiero · Post by **Mr-Fiero** » Thu Jun 19, 2014 12:10 am

Short term - you should be able to get better than .5 Meters over say 1 minute. 1M would be even easier! I have 6 units, and all are the same for me for accuracy. its very do-able! I did not get the GPS today, but as soon as I do I will post new configs compatible with the NEO7.

Hey, when you stated, when you move over to harakiri, does that mean you have been trying something else? Baseflight does not do GPS well, and TC does not have any interest in GPS. If you use position hold (or other GPS functions), you need Harakiri to see it work well. MultiWii was not great either for me. FYI

Also, having a signal meter in Ucenter is awesome, and allows anyone to test for sensitivity first, and then interference next.

e_lm_70 · Post by **e_lm_70** » Thu Jun 19, 2014 12:59 am

I'm still mainly using multiwii

GPS rth and ph works good enough for me on multiwii with stock pid
Altitude hold on 5611 is even perfect for my need

MultiWii has now some speed calculation and location prediction , and gps filtering ... it was looking a waste , using noisy data for make noisy prediction, wadting cpu and memory.. and I cut it away .. so .. my multiwii depend only on GPS for keep the location ... and it does the job for me

Harakiri, should (I did not check the code, I trust what I read for now, I check code only in case of issues) have more advanced gps handling ... I would expect it does predict location by a fusion of gps data and accelerometer data ... making a double integration with double level of noise potentially ... but naza proven acc data can be used reliable for position hold

In my book , only the acro pid should be copter specific .. all other pid should work in a wide range of copter .. else ... there is soemthing wrong.

About APM ... LOL .. if you followed the myGeekShow .. a guy (quite a 'poor' guy in term of skills, but with huge ambitions) that want to make cost to cost by rc plane using mission planer from apm ... LOL .. not only APM can crash land multicopter, but can do the same for airplane ... amazing ... inside the first 2 days the guy crashed both his 2 planes and aborted the mission , after very few success flights ... frustrated he gave up on a mission sponsored by 8k USD from brave followers ...

jihlein · Post by **jihlein** » Thu Jun 19, 2014 1:34 pm

I found the my NEO6 when mounted close to the FC board would perform terribly, low satellite count, low CNOs. From past experience, the low CNOs (around 30) was a flag that this wasn't going to work well at all. I added a ground plane to the antenna, and in the same mounting orientation, I now lock 10-12 satellites, with CNOs in the high 30's low 40's. I'm not sure if by shield you mean ground plane.

Here's a link to a picture of my test quad with the stacked boards and ground plane on the GPS:

http://aeroquad.com/showthread.php?7065 ... #post75935

strips · Post by **strips** » Thu Jun 26, 2014 10:48 am

I couldn't help my self :mrgreen:

Just recieved this one: http://www.csgshop.com/product.php?id_product=173

Got a few satellites and ok HDOP indoors:

Looking forward testing this on Harakiri.

strips · Post by **strips** » Thu Jun 26, 2014 11:32 am

I see on diydrones forums that they recommend Airborne 4G as Dynamic Platform model. Because this gives the APM the least filtered data so the APM filters can do their job. See comment from tridge: http://diydrones.com/profiles/blogs/ublox-m8-series-gps-glonass-receiver-test

How does this translate to Harakiri?

e_lm_70 · Post by **e_lm_70** » Thu Jun 26, 2014 2:23 pm

strips wrote:I see on diydrones forums that they recommend Airborne 4G as Dynamic Platform model. Because this gives the APM the least filtered data so the APM filters can do their job. See comment from tridge: http://diydrones.com/profiles/blogs/ublox-m8-series-gps-glonass-receiver-test

How does this translate to Harakiri?

The velocity accuracy is much more important than most people realise, as it is the GPS velocity readings that allow the autopilot to correct for inertial forces and thus controls whether the aircraft can get an accurate attitude. If the GPS produces a great position but terrible velocity numbers then the aircraft will know where it is but won't know which way is up. This has been the root cause of many crashes, as poor GPS velocity leads to a poor roll/pitch solution and the aircraft turns on its side and crashes.

OMG

I though Mr.Fiero was exaggerating a bit over APM ... but these APM developer will never stop to surprise me.
After the useless 6 position ACC calibration ... how dummy is this ... how could GPS sensed speed play any role

A control board should sense zero speed at take off .. store this, and then try to use GPS location and Accelerometer .. for fuse location, and speed ... location is the integration of speed, speed is the integration of acceleration ... for me .. it sound to integrate twice a noisy acc sensor ... but ... yes ... as shown by DJI, it can be done with decent result.

Ok .. back to Harakiri translation ... I guess you should not bother much ... More accurate is the GPS location, and better result you will get ... GPS filtering, actually happen both inside the GPS module and both in the control board ... it cause some "lag"/"comeback" in GPS-PH as shown by Mr.Fiero ... but ... nothing too serious.

PS: I see that is possible to order these GPS directly from the shop, avoiding eBay ... this 1.5g GPS is looking amazing to me: http://www.csgshop.com/product.php?id_product=172 ... quite expensive 65$+7$ .. but it has also i2c exposed interface ... maybe if once back from my vacation I can consider to get one

strips · Post by **strips** » Thu Jun 26, 2014 2:49 pm

e_lm_70 wrote:
strips wrote:I see on diydrones forums that they recommend Airborne 4G as Dynamic Platform model. Because this gives the APM the least filtered data so the APM filters can do their job. See comment from tridge: http://diydrones.com/profiles/blogs/ublox-m8-series-gps-glonass-receiver-test

How does this translate to Harakiri?

The velocity accuracy is much more important than most people realise, as it is the GPS velocity readings that allow the autopilot to correct for inertial forces and thus controls whether the aircraft can get an accurate attitude. If the GPS produces a great position but terrible velocity numbers then the aircraft will know where it is but won't know which way is up. This has been the root cause of many crashes, as poor GPS velocity leads to a poor roll/pitch solution and the aircraft turns on its side and crashes.

OMG

I though Mr.Fiero was exaggerating a bit over APM ... but these APM developer will never stop to surprise me.
After the useless 6 position ACC calibration ... how dummy is this ... how could GPS sensed speed play any role

A control board should sense zero speed at take off .. store this, and then try to use GPS location and Accelerometer .. for fuse location, and speed ... location is the integration of speed, speed is the integration of acceleration ... for me .. it sound to integrate twice a noisy acc sensor ... but ... yes ... as shown by DJI, it can be done with decent result.

Ok .. back to Harakiri translation ... I guess you should not bother much ... More accurate is the GPS location, and better result you will get ... GPS filtering, actually happen both inside the GPS module and both in the control board ... it cause some "lag"/"comeback" in GPS-PH as shown by Mr.Fiero ... but ... nothing too serious.

PS: I see that is possible to order these GPS directly from the shop, avoiding eBay ... this 1.5g GPS is looking amazing to me: http://www.csgshop.com/product.php?id_product=172 ... quite expensive 65$+7$ .. but it has also i2c exposed interface ... maybe if once back from my vacation I can consider to get one

Thanks.

I saw a post in that or another thread on diydrones that the helical had poorer performance.

timecop · Post by **timecop** » Thu Jun 26, 2014 2:50 pm

btw, tridge is correct.

e_lm_70 · Post by **e_lm_70** » Thu Jun 26, 2014 3:00 pm

strips wrote:...
I saw a post in that or another thread on diydrones that the helical had poorer performance.

Thanks ... I will dig more about it.

Possibly I will save 70$ .. for stay on my old GPS :geek:

Mr-Fiero · Post by **Mr-Fiero** » Fri Jun 27, 2014 3:13 am

I second that tridge is right.

That is why I said if you did your GPS testing at a stationary point thats great, but you still needed two gps's to get proper results on true accuracy. Again, its not all about accuracy. To be clear, I am not referring to differential GPS. But I was trying to keep it simple. I see tridge says the same exact point.

Nice link/read, I guess tridge said it better......the main thing I prefer now is the pedestrian model, because Its not just averaging filters, there are much more considerations on the GPS models used, and they need to fit your application. 50M/s is more than enough for speed, but the algorithms just seem to perform better for smaller/slower aircraft. At Least it seems that way in my tests for me. The only thing a flight controller does for filtering as far as i can tell, is some averaging of the data, but the bulk of the processing for position data is the GPS. The FC needs this to be in best form so the FC can predictively position the craft. With this, there is some trade off's that have to be used to get flyable GPS data. I always find it amazing that the codes work with the FC's to accomplish this task. Any code that flies always impresses me. Adding GPS and having it still fly is also very interesting, especially when you get to see the results.

OK. As of today, I got the new GPS in. Started setting it up on the 35mm antenna. WOW, so much better but not tested yet so we will see what real world performance is. I will post new configs for Ucenter probably once the GPS has been flight tested.

I am putting the new GPS on the new hex. I have to install my stm32f3 first. Tried APM2.5 again....dumb, but yep, I'm done. I was even trying to compile the newest firmwares, and its such a bitch to fit into the APM. Its just outgrown the platform. I wont even start on the details of all the failures with the APM's this time, but its alot! Just waiting now for headers for my new FC and the hex will be flying soon with Harakiri. First time for me for Harakiri and a Hex. All my other units are quads. Hopefully the hex will be my new favorite unit. Right now, one of my crusty quads is my fav.

--> strips
nice GPS! should work very well for you! hey, does that have a saw filter on that board? I cant see it and I am not sure if the ublox8 engine has one inside of the unit. I have to read up more, but the examples I saw was daughter boards with the saw filter. But the ublox states that it is included, so one would think it was onboard the GPS engine itself. Or, I think I see one labled FL1 on that board.....maybe.

hinkel · Post by **hinkel** » Sat Jun 28, 2014 9:32 pm

hinkel wrote:Hi Crashpilot1000 !
Because your Althold is very fine.
I am testing an automatik althold support. If you are flying and you don't move the throttle stick for 2000ms, althold is engage autonomously and
disable if throttle stick move more then deadband. The goal behind is to keep a stable altitude during flite and don''t move some switch. .
Code: Select all
cfg.al_suptime  = 2000; // 0 = disable // althold is engage autonomously after for ex: 2000 ms when throttle stick is not moving deadband Throttle = 60 
    cfg.al_deadsup                = 60;         // Throttle deadband when Althold support is use  
    cfg.al_maxthrsup              = 1790;       // Over this throttle value Althold support is disangage
It is like a poor man sonar , I don't know !

Regards
hinkel

Just a Test Video.

Bamfax · Post by **Bamfax** » Wed Jul 02, 2014 8:53 am

Just some Harakiri PH + Althold.

bulesz · Post by **bulesz** » Thu Jul 17, 2014 9:48 pm

Heyho,

My full Naze32 boards are on the way (for my micro H and for my fat AUW 2,5 kg SS X3508 700kv bird).
I have some GPS+MAG modules (NEO6) from the APM era, can I use them with the Naze32 (plan to use Harakiri FW)?
Should I use the external magnetometer or using the Naze32 "onboard" MAG is better?

Many thanks guys,
B

ps: One more thing: with the APM I have been using OrangeRX OpenLRS RX/TX modules with PPM Sum, could I use it fine with the Naze32? I mean with the Crius AIOP V2 I couldnt use the PPM Sum with the MultiWii 2.3...

Bamfax · Post by **Bamfax** » Thu Jul 17, 2014 9:52 pm

Hi bulesz,

many people have done better with an external mag, I personally never had problems with the mag on the flip32+ boards I tend to use.

Just have a go with the internal, you will see if it has outside influence by throttle, current, etc. If it doesn't work out, you can still switch.

Cheers,
Olli

bulesz · Post by **bulesz** » Fri Jul 18, 2014 12:07 pm

Bamfax wrote:Hi bulesz,

many people have done better with an external mag, I personally never had problems with the mag on the flip32+ boards I tend to use.

Just have a go with the internal, you will see if it has outside influence by throttle, current, etc. If it doesn't work out, you can still switch.

Cheers,
Olli

Thanks Olli,

I have used the external mag on the APM, so it is the same on the Naze32? Enable in the settings, calibrate (set the orientation) and enjoy?

Cheerz,
B

Bamfax · Post by **Bamfax** » Fri Jul 18, 2014 1:35 pm

No, it's more like hotair the internal one out

strips · Post by **strips** » Sun Jul 20, 2014 9:56 pm

I'm out in deep water here. Just got my first Mac and wondering what is a good MW GUI for Mac that works with Harakiri? Only need it for RX/TX verification, AUX and map visualisation for mag/GPS. Anything else I use CLI.

Mr-Fiero · Post by **Mr-Fiero** » Tue Jul 22, 2014 2:19 am

To Crashpilot1000.

I need to try to use the gimbal with a flip32+ but have a question.

Do I use motor outputs 5, and 6 while its on the quadcopter? I know that i need to enable the feature but is that all that is needed to start using a gimbal that has pan/tilt only with two servo's?

BTW, my units are still flying,,,never an unexpected drop. Still running Testcode3 but as soon as I get some good fly time in, I will re-visit the Testcode4. I have never had sooo much success in a flight controller and now I have a massive amount of fly time. I do use the GPS heavily as I am into photography, and its really nice to be able to use the PH mode. Its still working on all my units, as good as the videos I had posted before. Its been a very tight PH every time (granted as long as I get a good lock first).

Mr-Fiero · Post by **Mr-Fiero** » Wed Jul 23, 2014 9:07 pm

OK after some searching I found this for what you had typed about the gimbal before Rob.

enable feature "SERVO_TILT" first. Then the first 2 PWMs are used for your gimbal and the motorchannels are shifted 2 places higher accordingly.

I will continue to take that route, as indicated in that post that you do not run gimbal. Therefore my focus will need to be looking at Baseflight code as its used in Harakiri untested.

bulesz · Post by **bulesz** » Wed Jul 23, 2014 9:34 pm

Mr-Fiero wrote:OK after some searching I found this for what you had typed about the gimbal before Rob.

enable feature "SERVO_TILT" first. Then the first 2 PWMs are used for your gimbal and the motorchannels are shifted 2 places higher accordingly.

I will continue to take that route, as indicated in that post that you do not run gimbal. Therefore my focus will need to be looking at Baseflight code as its used in Harakiri untested.

Hey Mate,

Please keep us informed, because I plan to do the same...big quad using Naze32 with Harakiri + gimbal with tilt control...

Crashpilot1000 · Post by **Crashpilot1000** » Wed Jul 23, 2014 10:10 pm

Hi! Sorry for long not posting or doing any copterstuff at all.... I have no gimbal running so I can only repost: viewtopic.php?f=23&t=3758.
The parameternames have been changed:

Code: Select all

    // gimbal
    cfg.gbl_pgn                   = 10;
    cfg.gbl_rgn                   = 10;
    cfg.gbl_flg                   = GIMBAL_NORMAL;
    cfg.gbl_pmn                   = 1020;
    cfg.gbl_pmx                   = 2000;
    cfg.gbl_pmd                   = 1500;
    cfg.gbl_rmn                   = 1020;
    cfg.gbl_rmx                   = 2000;
    cfg.gbl_rmd                   = 1500;

Happy flying!
Cheers Rob

Mr-Fiero · Post by **Mr-Fiero** » Thu Jul 24, 2014 5:39 am

Yep, Thanks Rob.

I will post my results when tested.

Mr-Fiero · Post by **Mr-Fiero** » Thu Jul 24, 2014 7:06 am

gimbal is non-functional in Harakiri. TestCode3 - atleast for me.......MinumOSD works VERY well though!!!

I tried on two micro servo's that I know work and they will slightly move, like they want to. I cranked gain cfg.gbl_pgn and fg.gbl_rgn all the way to 100 but it made no change. The servo's however when FC is booted, will center on FC boot, and hold with PWM, but after boot, there is only very slight movement and its hard to detect its so small. even tried arming just to make sure.

Thats OK, I will just get a gimbal controller and setup the Pan/Tilt to keep isolated away from Flight Controller on separate power as well. Its probably better this way as there are some really nice gimbal controllers out these days. They usually have excellent GUI for setup as well.

bulesz · Post by **bulesz** » Thu Jul 24, 2014 9:40 am

Thanks for trying it out. I think I will do the same...separated gimbal controller.

Dude, are you using MinimOSD with KV OSD software on it?

Thanks,
B

Crashpilot1000 · Post by **Crashpilot1000** » Thu Jul 24, 2014 1:00 pm

Thx for the feedback!
I looked at the gimbal code without finding an obvious error - that doesn't mean there is none, of course

. Since that is mwii, did you activate CAMSTAB in the aux boxes?
Cheers Rob

gray100 · Post by **gray100** » Thu Jul 24, 2014 1:17 pm

Anyone know a version of MinimOSD that works with Testcode 3/4 ?

I tried Kv Team 2.3 and MW OSD 1.1 and some of it works, but always says disarmed. and wont save changes by the tx sticks via the osd gui.

Just wondering if there is a version people are using which works better with Harakiri?

osd's work fine with all functions in baseflight

Been using Testcode 4 mostly and works so well, going try testcode 3 and see if there is any difference, as from reading this thread seems most are using testcode 3.

Thanks everyone :-)

Mr-Fiero · Post by **Mr-Fiero** » Thu Jul 24, 2014 3:14 pm

Rob, good call on the CAMSTAB in AUX . I missed that as I was not looking there. I was searching for other settings, but somehow never saw that while testing, but I have seen it many times before in AUX. I will re-test tonight. saves me from using the Ardupilot as a gimbal controller....LOL Now again I have no purpose for the poor Ardupilots......

bulesz, Yes MinimOSD with KV Mod V1.0 as its from readytoflyquads. http://witespyquad.gostorego.com/minimosd.html
I had to drop my serial buad rate to 57600 on the FLIP32+ so the OSD would work, otherwise I would have had to re-compile for the OSD and who knows if it would handle higher baud. cfg.serial_baudrate = 115200; so CLI command would be "set serial_baudrate = 57600"
wired as per this diagram http://parkflysac.com/download/file.php ... &mode=view

bulesz · Post by **bulesz** » Thu Jul 24, 2014 3:53 pm

Mr-Fiero wrote:bulesz, Yes MinimOSD with KV Mod V1.0 as its from readytoflyquads. http://witespyquad.gostorego.com/minimosd.html
I had to drop my serial buad rate to 57600 on the FLIP32+ so the OSD would work, otherwise I would have had to re-compile for the OSD and who knows if it would handle higher baud. cfg.serial_baudrate = 115200; so CLI command would be "set serial_baudrate = 57600"
wired as per this diagram http://parkflysac.com/download/file.php ... &mode=view

Heyho,

Thanks broda for the infos! How could I change the baudrate of the serial? In the config.h? or in command line?

Many thanks,
Bulesz

Mr-Fiero · Post by **Mr-Fiero** » Thu Jul 24, 2014 3:57 pm

No prob

CLI this command should do it.
"set serial_baudrate = 57600" - or whatever your OSD defaults to but mine was 57600

Remember to save and reboot FC
"save"

Oh, yes, I should add, the OSD will not receive the mavlink unless its armed. When I was setting up that threw me off as bench testing was not working till I tried it armed.

I even got the battery voltage working. Next is to get my RSSI working with my OpenLRSng. That brings up a question that I have yet to research, what does the Harakiri want to see for a signal, is it 0v to 3.3v analog, or is it possible to have it read the PPM?

Anyways, off topic I have had another success. I am only posting here because I am not part of any other forum, or topic, or care to be either. this is my only hang-out.
Its concerning bearings, as my unit with the most flight time has bad bearings, I am still flying, but the bearings are still lasting since my video that you can hear them. I started using this lube when my bearings were already shot http://poweruplube.com/store/r-c-l-1000-2 I already know this lubricant has weird properties in other applications and works abnormally well. So I tried it. Usually when my bearings make noise as bad as they were, I only have maybe 5-10 more hours max before having to change the bearings. But now I am still flying at almost 4 months later on the already bad bearings. Since I have been using RCL1000 my bearing noise has been reduced by %80 as well and they are really smooth. I usually spray top and bottom bearings every hour of flight, and they just keep working! I am not in any way trying to sell powerup, but I am saying that it seems to work VERY well for this application. I used to have to change my bearing every couple of months and its pretty easy, but I have been so impressed that I have changed NONE in 4 months. I am going to fly the bad bearing quadcopter till the bearings totally fail just to see how long RCL1000 extended my bearing life. When I fly, I go for 2 hours of flight or more each time(max batteries avail), and sometimes several times in one day (up to 6 hours flight time a day). Good thing the wife is forgiving....LOL

Mr-Fiero · Post by **Mr-Fiero** » Fri Jul 25, 2014 12:50 am

Update after CAMSTAB advice ...

Holy crap, Gimbal works great on TestCode3 with my servo's! I just have to tweak params so it stays level and works with my hardware with proper extents now.

bulesz · Post by **bulesz** » Fri Jul 25, 2014 12:01 pm

Hi Guys,

I have just got my Naze32 boards, Rev5. I have flashed the testcode 4 on them, calib the acc and mag in the multiwiiconf.

But I can't get in to the CLI mode...nor in Putty (I have only strange characters....), or in the Baseflight Configurator (cant connect at all).

I have the BaseflightGui2 too, but there is only one exe, the STMFlashLoader.exe, and it is disappear right after ran it. I have run it in CMD, and as I see it is for flashing not reaching the CLI mode.. ??

So, how can I go into the CLI mode after flashing the Harakiri?

Many thanks,
B

Crashpilot1000 · Post by **Crashpilot1000** » Fri Jul 25, 2014 2:21 pm

mrfiero: Sometimes it is the most obvious stuff that is the hardest. Good to know gimbal works!
bulesz: You see the mavlink heartbeat. To enter cli type 3 times enter ("<Enter><Enter><Enter>") or 3 times # ("###").

(Best terminal prg IMHO: https://sites.google.com/site/terminalbpp/)

bulesz · Post by **bulesz** » Fri Jul 25, 2014 3:01 pm

Thanks mate for the quick reply!

bulesz · Post by **bulesz** » Sat Jul 26, 2014 10:34 am

Dunno if it is offtopic, but my bird hovers at 55% but it was hovering at 45% with APM... is there an initial CLI command or any settings what I have missed?

It flies very nicely, only the extra power is missing.... ???

The bird config is the exact same except the FC board...

Many thanks,
B

strips · Post by **strips** » Sat Jul 26, 2014 11:53 am

bulesz wrote:Dunno if it is offtopic, but my bird hovers at 55% but it was hovering at 45% with APM... is there an initial CLI command or any settings what I have missed?

It flies very nicely, only the extra power is missing.... ???

The bird config is the exact same except the FC board...

Many thanks,
B

You asking the same question in at least three forums/threads. Harakiri is a fork of Baseflight. So if you figure out Baseflight you figure out Harakiri.

Have you calibrated ESCs and set correct endpoints on radio?

bulesz · Post by **bulesz** » Sat Jul 26, 2014 12:21 pm

Yo, coz I havent got any usable answer...and it is FW related question...coz it is different waz to set up.... I have red the documentations, but I still don't know how to set the endpoints correctly..I mean the docs are about the cable and modules plugs, and not about the proper settings..

Coz I'm from APM I dont know what MUST be done in Naze32 FW settings... ???????

At APM you calibrated the Radio and all set (I mean the radio related stuff....)

So, how can I set properly the endpoints? And how can be sure that the endpoints are correct? because the behaves of the bird is nice. Flies very good, so NOTHING seems to be need to change/missing with the radio related settings...except the missing ~10% of throttle...

****************
EDIT: I have changed the default max throttle from 1850 to 1980 and it is mucho better!

Crashpilot1000 · Post by **Crashpilot1000** » Sat Jul 26, 2014 10:47 pm

Dude, default for esc_max is 1950 https://github.com/Crashpilot1000/TestC ... fig.c#L216. However the mixer in mwii is rather crude - there is room for improvement...

bulesz · Post by **bulesz** » Sun Jul 27, 2014 12:18 am

yeye, thanks mate for the reply. I have learned a lot in these two days...but it was worth it, my GPS and OSD is working and I have made my first acro movements with my big bird (500 class) so it was nice...

akcom · Post by **akcom** » Sun Jul 27, 2014 2:47 pm

Unfortunately my german is subpar and reading that translation on fpvtreff.de made my brain hurt. Can someone just tell me very simply what is the advantage of using Harakiri over baseflight trunk?

strips · Post by **strips** » Sun Jul 27, 2014 6:33 pm

akcom wrote:Unfortunately my german is subpar and reading that translation on fpvtreff.de made my brain hurt. Can someone just tell me very simply what is the advantage of using Harakiri over baseflight trunk?

GPS hold and RTL works wery well. In Baseflight you can get it to work but no where near the same precision.