dRonin - No longer just for the brave

[jipp]

thats cool the guys at dRonin are active. it will be interesting to see how this software grows.

its to bad about the how these groups can not get along.. im sure LP guys are the same who had a issue with tau.

my guess is they are all about robot stuff? im not sure what dRonin mission statement is..

but its a cool time to be alive and watch this stuff develop.

chris.

 

[jhitesma]

thats cool the guys at dRonin are active. it will be interesting to see how this software grows.

Active, friendly, and working on some cool stuff

its to bad about the how these groups can not get along.. im sure LP guys are the same who had a issue with tau.

LP is a lot of the core developers from OP so yeah, they still have a grudge against Tau. dRonin is trying hard to get away from the past and be known for what it's doing now not what people in past projects did. Some people like to keep digging up the past and complaining about it - but that isn't why LP and dRonin aren't working together. Basically it comes down to both groups have their own ways of working and goals and they just don't mesh. Which is fine, I'd be really upset if everyone wanted to do the same things for the same reasons...would make life boring

A big part of it comes down to control over the project though. dRonin forked from Tau in large part because the active developers were frustrated at being unable to get stuff done due to tau's management (which put a bunch of limits on tau because of the issues they had with OP that caused them to fork.) So the dRonin devs don't want to give up the freedom they got by going out on their own. LP basically formed because OP imploded (due to many of the same reasons that caused Tau to fork previously) and they're also enjoying their freedom and don't want to give it up.

my guess is they are all about robot stuff? im not sure what dRonin mission statement is..

Not sure what LP's goals are. dRonin wants to build the best code possible for basically everything. From the dronin.org site "Whether you want to push the limits of FPV racing or aspire to build advanced autonomous behaviors for aerial robotics, the dRonin flight stack makes it possible."

The main guy behind dRonin really wants to do some crazy advanced flying robot stuff. But he also wants to get away from PID's and develop some alternative control schemes as well. But before that stuff can really happen there's a lot of groundwork that has to be done which is what's happening now. Fixing old bugs, cleaning up base code, improving usability and performance...

Getting more people flying it and sharing experiences/bugs will help a lot of that happen quicker. Getting more developers interested and contributing will help even more

There's a really interesting side project for dronin going on right now that's slowing some of the core work...but when it's done will help advance things for all projects quite a bit. All I'll say for now is that the software is on github and first try at hardware is close. Once it's a bit more mature I'll be talking about it in detail for sure

but its a cool time to be alive and watch this stuff develop.

It is indeed!

 

[jipp]

thanks for the feed back.. yeah i can understand each team wanting the freedom and work flow as they like to work in
yeah op imploded, must of been one huge head ache for all. i look forward top dRonin future advances.
will be cool to advance past pids etc.. even if i just got the basic understanding of pids under my belt
thanks to Bruce recent YouTube video. still a bit confusing in some aspects but his video helped.

i was gonna use the sparky II on my next build. but have decided to save it for a future more advance AP build and by the time i get parts for that im sure dRonin will be ready for such tasks.. i have ordered a kiss fc for this build.. quick 46.00
but whatever, now if the kiss 24a esc will come to market even tho with everyone looking i doubt ill score any, so ill probably just look for a good deal on some others.

hopefully i can start working on my next build in 2 weeks or so.
chris.
p.s
i should mention i do have a RCX 175mm frame that just needs motors i have every thing else plus 5 OP FC on my desk so ill be using dRonin before the AP machine. like you said no shortage of projects on my desk that need a few parts here and there to start. this hobby is quite addicting. after all you only need one quad.

 

[jhitesma]

thanks for the feed back.. yeah i can understand each team wanting the freedom and work flow as they like to work in
yeah op imploded, must of been one huge head ache for all. i look forward top dRonin future advances.
will be cool to advance past pids etc.. even if i just got the basic understanding of pids under my belt
thanks to Bruce recent YouTube video. still a bit confusing in some aspects but his video helped.

Interestingly Peabody has actually got some of the new non-PID based stuff flying on Tau this week.

https://www.instagram.com/p/BD9y3S7lqvj/?taken-by=peabody124

He's using an experimental LQG controller https://en.wikipedia.org/wiki/Linear-quadratic-Gaussian_control

This is the same basic path that rs2k wants to take raceflight down and icee wants to take dRonin down, though I'm pretty sure they're talking about different control schemes than LQG - but still predictive schemes that don't use PIDs

The catch is that those non PID systems all need a good system model to work well and that has to be specific to each individual craft. Which is why PID's aren't completely eliminated. The way Peabody has LQG working is he does one flight with the PID controller to run autotune which generates the system model. That's basically how autotune works, it's just that so far that system model has only been used to calculate optimal PID's - Peabody is now using it for the LQG controller instead of using it to calculate PID's. That's the same thing icee wants to do with dRonin. I'm not sure what rs2k's plan is to do this on raceflight since they don't have the autotune code that generates the system model.

Oh, the other catch...this isn't going to work on F1 targets. They just don't have the power. Running autotune on F1 is already hitting 99% utilization of CPU, it's maxed out. Peabody's new stuff only works on sparky2/F4 it's currently too demanding even for F3. That could change with optimization but we'll have to wait and see.


So yeah, exciting times!

 

[jipp]

thats great to hear.. well to be honest. people are just gonna have to except the FC needs to be upgraded to F4 and above as the technology grows etc..
i think most people are willing to update a fc.. since you need one with each build. the problem i guess comes down to people getting complaisant and comfortable with what they know and have.. so some will refuse to fly anything but a naze rev 6 for example.

should be fun to see how these guys work there magic.
i can only imagine a auto tune must of been a lot of work to get it to work correctly.

chris.

 

[jipp]

hey jason, is there a reason why newer FC do not add a on board bec? its annoying to have to come up with 5v, i have several pdb here but none have 5v on them so i had to order one so i can start to work on my 120mm insect. i can only think it must have to do with clean energy. and a bec would give this in spades.. but it sure would be nice if the FC could handle this too.

chris.

 

[jhitesma]

thats great to hear.. well to be honest. people are just gonna have to except the FC needs to be upgraded to F4 and above as the technology grows etc..
i think most people are willing to update a fc.. since you need one with each build. the problem i guess comes down to people getting complaisant and comfortable with what they know and have.. so some will refuse to fly anything but a naze rev 6 for example.

I think a lot of people abandon frames quicker than they abandon FC's. But that's just my guess. The CC3D and Naze are popular in no small part because of how cheap they are now. That and having been "first" helped a lot.

should be fun to see how these guys work there magic.
i can only imagine a auto tune must of been a lot of work to get it to work correctly.

Jump into some of the IRC channels if you really want to see the magic happen. Just trying to keep up with some of the discussions on there is a great way to learn a lot

As for autotune...here's the details:
https://github.com/d-ronin/dRonin/blob/next/flight/Doc/Autotuning Derivation.pdf

It's not exactly "light reading" though!

hey jason, is there a reason why newer FC do not add a on board bec? its annoying to have to come up with 5v, i have several pdb here but none have 5v on them so i had to order one so i can start to work on my 120mm insect. i can only think it must have to do with clean energy. and a bec would give this in spades.. but it sure would be nice if the FC could handle this too.

Heat and noise are the two biggies. A linear regulator like many ESC's use for a bec gives off a lot of heat. A switching regulator like a pololu gives off a lot of noise. Neither are things you really want on your FC for a number of reasons.

That said I know of at least one upcoming FC that's currently in the hands of some early testers (hint: such OSD wow) which doubles as a PDB with built in bec. I'm kind of excited about it, those I know who've got them are pretty impressed and it's designed by someone with a very good track record. I believe it uses a switching regulator with some extra filtering to deal with the noise issues. Should be a very nice board for small clean setups when it comes out.

 

[jipp]

yeah, i can see heat and noise being a issue. i guess that is why some ESC use switching and others use linear i think that is the other type.. i just know if your ESC are switching like emax Simon K for example you do not have to remove the power from 3 of the ESC.. but if they are not switching you do.. or at least that is what i have read.. maybe newer FC can deal with either type of power now.

yeah i should check into IRC its been a while since iv used it. i use to be addicted to irc in the 90s

i just assumed because FC are cheap ( this is sorta new thing tho and recent.. since i paid good money for my sparky II yet it has gotten cheaper already, guess compared to not to long ago.. i assume that is why a lot choose to use the KK board because it was the best bang for your buck back then ) they use a new one with each build.. but i guess it could go either way..
i guess switching parts over to a new frame is most common. here i am building new ones for no reason other than to learn, and i find it fun to build.

sounds like the new FC is gonna be cool. iv seen PDB with on board OSD but cant say iv seen any FC with OSD yet.

chris.

 

[Snarls]

The KISS FC has a built in BEC. People say it gets warm along with the usb port. Not sure how that affects the boards performance though.

 

[Twitchity]

I'm pretty sure this isn't the board Jason is talking about, but it follows the same idea as an all in one solution FC/OSD/PDB http://www.readytoflyquads.com/flip32-aio

For me I've stuck with the Naze for so long for a few reasons. It's been good to me in the past, they don't cost an arm and a leg, and there's tons of documentation out there for them. The last reason is the main deciding factor for me. I'm an idiot when it comes to getting the FC configured for my setup, and I can do a quick Google search and find my answer within minutes for the Naze. I really wish dronin officially supported this board as I'm struggling to get it running on my Tricopter (which is pretty much all on me for being dumb and inexperienced when it comes to these things). My setups on new FC typically involves hooking up a flying quad, going into CLI in cleanflight, running a dump, and then copying and pasting that to the new board. Everything is configured within minutes and I'm on my way, but no auto tune which is why I want dronin so badly which goes back to the part about me being dumb and not knowing how PIDs work and how to tune. And yes, I've seen the videos explaining PIDs, it's just not a concept I can get my head around at this point haha

 

[cranialrectosis]

I think the Naze is cheap and good enough for 90% of pilots out there. If want cutting edge you can go minimalist like the tornado or large like DJI or a Brain.

If all I wanted was to fly, a KK2 or Naze on BF or CF works fine for me. I tinker with dRonin so I can work with the robotics geeks at the local high school, and so I can document my attempts here. I figure if a butthead like me can do it, anyone can.

Oh and like Twitch said, tuning PIDs blows. Yeah autotune!

 

[jhitesma]

I'm pretty sure this isn't the board Jason is talking about, but it follows the same idea as an all in one solution FC/OSD/PDB http://www.readytoflyquads.com/flip32-aio

Nope, that's not the one The one I'm talking about isn't out yet and has a better processor and no OSD but will probably have a mating OSD given the designers history.

Since one of the people with one posted a photo to reddit I guess I can share it here:

I really wish dronin officially supported this board as I'm struggling to get it running on my Tricopter (which is pretty much all on me for being dumb and inexperienced when it comes to these things). My setups on new FC typically involves hooking up a flying quad, going into CLI in cleanflight, running a dump, and then copying and pasting that to the new board. Everything is configured within minutes and I'm on my way, but no auto tune which is why I want dronin so badly which goes back to the part about me being dumb and not knowing how PIDs work and how to tune. And yes, I've seen the videos explaining PIDs, it's just not a concept I can get my head around at this point haha

Well, Naze32 is officially supported...it's just that due to the design of the Naze32 support isn't at the same level as other boards. dRonin usually uses a bootloader which greatly simplifies flashing and initial setup. But that can't be done on the Naze32 because of the CP2102 chip between the F1 and the USB port instead of using the F1's built in USB functions (the same old issue with Naze32 I've been complaining about since I first tried one.) That's why GCS can't flash the Naze32 directly and you have to use a *flight configurator to flash the dronin .hex file.

Also instead of CLI dumps dRonin uses XML formatted UAVO's which make exporting and importing settings far more reliable and make the new auto-upgrade process super slick. Unfortunately due to the hardware limitations of the Naze32 auto-update doesn't work on it. But on other boards (even CC3D) it works amazing. Even when dRonin makes changes to the internal configuration you don't have to worry about it - when auto update runs it uploads your current settings to the cloud - runs them through a converter - and spits out a new configuration file updated to work on the new version. the auto-update really is like magic. Just plug in your board, power up the newest GCS, hit ok, and boom it does everything else for you.

CC3D and Naze32 are very similar boards...but for dRonin CC3D is a somewhat better choice because it uses the F1's built in USB so the bootloader works and dRonin can flash it directly and do auto-updates. But the F1's days are really numbered. Autotune BARELY works on F1. icee did some amazing optimizing and used some impressive tricks to get it to run and even so the CPU is hitting 99% utilization when running autotune - and that only at the lowest gyro rate (think looptime in *flight) setting. The boards can run higher gyro rates...but then there isn't enough CPU left to run autotune so you're back to manual tuning

Sparky1 clones are out there for about $23 now and are a very nice F3 board that's fully featured with dRonin and supported in most *flight versions. It's probably the best "budget" option for a new board IMHO with the widest choice in firmwares. And in some ways the F3 is better than the F4. It's a newer chip and has better UARTS. It just doesn't have as much memory or as high of a clock rate as the F4 chips. (why STM released an F4 before a F3 I have no clue...such a confusing move that one.)

One of the dRonin developers made a great easy to understand post on reddit yesterday about autotune and the benefits of autotune over manual tuning and why chasing looptimes and ultra high gyro rates gives greatly diminishing returns: https://www.reddit.com/r/Multicopte...gotten_into_dronin_yet_if_so_what_are/d2jfitn


Naze32 boards fly great. But all F1 boards days are numbered and due to it's wonky USB setup the Naze is never going to be as easy to use with dRonin as even the CC3D. That's just the unfortunate reality due to the hardware limitations. There are things the dRonin devs could do to get GCS to be able to flash Naze32 directly...but it's a lot of work and is hard to justify since the F1 itself is so limited even though there are a lot of naze boards out there...and due to the USB setup still wouldn't be able to do auto-update stuff on Naze32.

 

[jipp]

yeah that is what i figured about auto tune it could do a better job than us. math the universal baseball bat!
on the other hand to the guys who hand tune and say they can do better, i think
its just that is what they are use to flying so why would they not think that is the best..

also cc3 FC are f1 chips correct? i have 4 of 5 of them on my desk. and one sparky II, and a kiss FC in the mail.

so ill just get a sparky 1 when i need to a FC.

chris.

 

[jhitesma]

also cc3 FC are f1 chips correct? i have 4 of 5 of them on my desk. and one sparky II, and a kiss FC in the mail.

Yes, the CC3D's are F1 based. But since they use a bootloader and the built in F1 USB support instead of a separate chip like Naze32 they're better supported in dRonin. In fact nearly half the boards currently running dRonin (that we're aware of - there could be more if people opt out of the data sharing) are CC3D's...much to the dismay of the devs since the F1 is already maxed out

 

[jipp]

Yes, the CC3D's are F1 based. But since they use a bootloader and the built in F1 USB support instead of a separate chip like Naze32 they're better supported in dRonin. In fact nearly half the boards currently running dRonin (that we're aware of - there could be more if people opt out of the data sharing) are CC3D's...much to the dismay of the devs since the F1 is already maxed out

thanks for the info, yeah i bet the devs are like sigh damn the cc3d just will not go away. laughs. i guess the main problem is price! and also popularity.. as dRonin is not as well know as OP, or LP.

so that will take time, then the main issue i guess would be all the kist sold with a cc3d.

id say its 75 percent cc3d the rest of the kits on ebay are naze. so yeah, they need to upgrade to at least a sparky II that would be cool.. but cc3d is here to stay until its pushed to the side by what people will buy, and then this comes back to people knowing the benefits of a F4 chip, and firmware like dRonin.

chris.

 

[evilmerlin]

So what's the best value board for dRonin now? Sparky 1? Or 2? Could get a Sparky 2 clone for about $50 which seems quite reasonable. Or should I suck it up and wait for the BrainFPV RE1?

 

[cranialrectosis]

So what's the best value board for dRonin now? Sparky 1? Or 2? Could get a Sparky 2 clone for about $50 which seems quite reasonable. Or should I suck it up and wait for the BrainFPV RE1?

I think it depends on what you want from your board. I love my Brains because of the onboard OSD and the flat underside. It makes for a super-clean build. I like my Sparky2 but have to have a separate OSD if I want OSD and the Pololu won't fit under it because of all the bottom side ports. I really don't want my board on stilts to fit a Pololu under it. I put my Sparky 2 on my hex as the hex top frame plate is a PDB that supports a Pololu so I don't have to raise the board. The hex frame has much more available real estate so I have lots of room for a minim or dal OSD.

That said, a Sparky clone with a Minim OSD costs less than a Brain. For me it really just comes down to preference and the frame and power setup.

 

[jhitesma]

Best bang for your buck is probably a Sparky1, tough to beat at <$25:
http://www.aliexpress.com/item/Spar..._1&btsid=4a6cf3a7-e199-4fd8-af40-f79add4c22cb

The only two "issues" with it are it's shape and that it's a F3 and not an F4.

The shape is basically a cosmetic issue, it's actually the same 30.5mm mounting pattern as Naze32/CC3D/SP/whatever but with one corner missing. It just kind of bugs those of us who enjoy symmetry

F3 instead of F4 is a plus in many people's minds...but the F4 is a bit more future proof due to the extra memory and higher clock rate. There's currently nothing available that the F4 boards can do that F3 can't. That may change with Peabody's LQG stuff...but I suspect it will end up working on F3 as well and it's not like he's done more than teased that work so far.

CC3D is "fully supported" but the F1 chip really holds it back as it just isn't powerful enough so you do loose some functionality. If you don't want to do GPS navigation stuff then it may be acceptable. But it's days are numbered. It's a better option than the Naze32 but only barely.


The new Brain RE1 is looking REALLY cool. But no word just when it will be available and original brains are out of stock and not going to be made again so unfortunately right now Brain isn't an option even though I absolutely love mine.

Sparky2 and Revolution are both fully supported F4 boards with built in radios. The radio is optional on Sparky2 but I'm not aware of any manufacturers currently making them without the radio. Which is a bummer because leave off the radio and the (useless to non developers) debugging port on the bottom and it can have a bare bottom which would alleviate Cranial's issues and lower the cost a bit.

So basically I'd suggest a Sparky1 if you want something cheap, but if you can spend the extra Sparky2 and Revo are a bit more future proof. I personally also like the radio in my Sparky2, I use it with my OpenLRS TX module for control so I don't need a separate RX, but want to pick up an oplink one of these days so I can use it off my notebook for GPS navigation/telemetry.

 

[Balu]

The new Brain RE1 is looking REALLY cool. But no word just when it will be available and original brains are out of stock and not going to be made again so unfortunately right now Brain isn't an option even though I absolutely love mine.

I've not seen anything about the Brain RE1 but the announcement that they are working on something new. Do you have a link perhaps?

 

[Twitchity]

Sparky2 and Revolution are both fully supported F4 boards with built in radios. The radio is optional on Sparky2 but I'm not aware of any manufacturers currently making them without the radio. Which is a bummer because leave off the radio and the (useless to non developers) debugging port on the bottom and it can have a bare bottom which would alleviate Cranial's issues and lower the cost a bit.

I'm strongly considering this route with the Sparky2 as it will get me a purple board, and I can build it to omit the radio portion on the bottom so it's nice and flat. The only issue is buying the boards and parts is more expensive than getting one from a Chinese seller. Still sounds like a fun project though!

https://oshpark.com/shared_projects/1ATjBell