Everything I have from 3DR sucks...
- Thread starter SolidSalad
- Start date
R_Lefebvre
Arducopter Developer
To be clear, much of these issues are actually hardware problems, not software. I'm not sure what "not long ago" is in this environment, but AC2.9, which is what brought out inertial alt-hold and started the vibration problems, is about 2 years old. And it's not even correct that to say that Arducopter never had vibration problems before that. You may recall the term "the leans", which was something that happened under heavy vibrations, and was a problem all the way back to when I started 3 or is it 4 years ago?
The root cause of this is that 3DR never understood the need for hardware vibration damping, or wasn't able to implement it. This left users to figure this out all on their own. The changes to the code which made the need for damping more critical did not cause the problems, they just exposed the fact that 1) The APM needed vibration damping, and no solution was provided by the hardware maker. And 2) many people had copters that are built really poorly, and vibrate really really badly.
The problems with the baro, are again, hardware problems. The chip OEM promised the chip is insensitive to light, but obviously that was not the case. This necessitated the system be designed to shield the baro from light with a proper case. Again, not a software problem, but a hardware problem. And software did not fix the problem, other than that it now ignores the baro if it thinks it's gone really bad. That is not a great solution, rather more like a band-aid on a shot-gun wound.
Even longer than that in fact. I would agree that it didn't really get good until about 2 years ago. Not coincidentally, that is about the same time that Randy and Tridge started working on it full-time. That's when it turned a corner. They didn't do all the work, but have been able to check the work of others to make sure that it does not bring in new bugs, and make sure that all the bits that people are working on play nicely with the bits that other people are working on.
However, it has now gotten to the point that I believe the project has outgrown it's current organization. It's too big for two people to manage. Arducopter in particular is moving very slow, something like 6 months (or is it approaching a year?) between stable releases. I don't know what the solution is, but something has to change. There's recently been an announcement about this "Dronecode" thing, which is intended to move Ardupilot out from under 3DR's umbrella, and make it it's own entity. It should help to attract funding from more than just one source (3DR). But I have no idea what is really happening with it. There's been almost zero information about this, even amongst the developer group. From what I can see, it's only created even more work for Randy and Tridge, as now they have to manage this organization, while also managing Ardupilot. With no extra resources at all, at least from what I can see. I hope that this can be sorted out soon. When the Dronecode announcement was made on Monday, it was claimed that there are 1200 developers. This is a gross overestimation IMO. I can't think of more than 12. Actually, I can only name 6 people seriously working on the main body of code, with me being one of them. There are maybe 50 or so doing various things, including porting the code to run on other hardware, etc.
I recently had to make the very difficult decision that I'm going to have to quit the group. I've been heavily involved for 3 years now, and over the past 12 months devoted every minute I could to the project, every evening, weekend, and even my vacation time. My family and home have suffered from neglect. I've been trying to figure out how to get in a position to leave my day job so I can work on it full time, but I just haven't been able to put it together. The disappointment includes $4000 in unpaid work by one company. And I can't just keep doing this forever, meanwhile Richard Branson and whoever else gets to profit off the project that I've helped create. It really sucks as I feel like I bet the farm, and now am folding. But I can't just keep doing this forever, my kids are complaining about daddy never being around... while I watch multi-million dollar companies selling hardware with a program I helped build.
Last edited by a moderator:
Must be a difficult decision indeed. May just be you’ve just been a bit too much ahead of your times, economically speaking. Hard to make a living out of this full time ... But this field is so much in its infancy, even as we watch it grow exponentially. Amazon, DHL, research labs, universities, education, multi-media companies, chip and computer manufacturers, etc ... these guys are starting and will be hiring big time, the revolution is just starting ... (In spite of the FAA who sure hasn’t helped).
I am convinced demand for experts who know a bit more than the average week-end Phantom flyer will grow explosively. This is not an easy field technically speaking, as it requires an unsual depth and breadth of knowledge ... EE, ME, Aerospace engineering, mechatronics, computer science, etc ... Hard to get, so demand will soar and, I’d venture, command high prices, even with the widespread availability of good hobbyists.
As far as that DroneCode foundation, seems to me it’s barely at the “blue print/draft ” level now, more an idea than anything else, and a bit too 3DR centric at this point imho. But I think the idea is right, although there no doubt will be growing pains. Making a living on open source is not easy, but history has shown there definitely is a viable economic model, Linux in general is a good example, so is the Linux Foundation ...
I am convinced demand for experts who know a bit more than the average week-end Phantom flyer will grow explosively. This is not an easy field technically speaking, as it requires an unsual depth and breadth of knowledge ... EE, ME, Aerospace engineering, mechatronics, computer science, etc ... Hard to get, so demand will soar and, I’d venture, command high prices, even with the widespread availability of good hobbyists.
As far as that DroneCode foundation, seems to me it’s barely at the “blue print/draft ” level now, more an idea than anything else, and a bit too 3DR centric at this point imho. But I think the idea is right, although there no doubt will be growing pains. Making a living on open source is not easy, but history has shown there definitely is a viable economic model, Linux in general is a good example, so is the Linux Foundation ...
R_Lefebvre
Arducopter Developer
Yes, I think you're right. After 3DR got it's first round of VC and most of the developers quit, I decided to stay on because I did see the corollary to Linux. The problem as I see it, is there are way too many people using, benefiting, and profiting from the program, and almost nobody is paying back into it. Will that change? I don't know. But at this point, the developer avenue seems to be a dead end for me.
3DR could jack up prices, say to $300-400 or more, still a bargain given functionality and the competition, *and* exclusively redistribute the delta to developpers. And innovate, innovate, innovate, on the hardware side, constantly, to always be ahead of the cloners, sort of the Sparkfun model. Everyone to some degree would benefit. Hobbyists tight on money would still be ok with availability of very capable low priced clones or reduced capability 3DR FCs. Although I am sure there'd be massive cries of fowl playing, they'd just have to forgo the latest and greatest FC with integrated px4flow or lidar or what not super advanced hardware functionality, if not able or wanting to pay full price.
Issue it seems is that 3DR appears inadequate on the hardware side at the moment. Be it board development or full copter offerings. Especially as opposed to, say, DJI. I mean, Iris is ok but, meh ... Not much innovation there ... Not much innovation with Pixhawk either, innovation came from ETH. Also still behind the Phantom as far as ease of use is concerned, and even though it can do much more, that may be too marginal from a business point of view. And nothing on the higher end of things, e.g. S1000 or Zenmuse competition, or fully integrated, high end, dead simple push a button and fly RTF filming platforms. Part of the reason imho why 3DR is far from the goto place for many commercial flying applications that don't require some of the advanced functionality of AC. With the sort of VC funding they got, I do not understand why they can't seem to even be able to offer this sort of product... Or move through acquisitions ... one man shops included ... Maybe too much DIY legacy.
Issue it seems is that 3DR appears inadequate on the hardware side at the moment. Be it board development or full copter offerings. Especially as opposed to, say, DJI. I mean, Iris is ok but, meh ... Not much innovation there ... Not much innovation with Pixhawk either, innovation came from ETH. Also still behind the Phantom as far as ease of use is concerned, and even though it can do much more, that may be too marginal from a business point of view. And nothing on the higher end of things, e.g. S1000 or Zenmuse competition, or fully integrated, high end, dead simple push a button and fly RTF filming platforms. Part of the reason imho why 3DR is far from the goto place for many commercial flying applications that don't require some of the advanced functionality of AC. With the sort of VC funding they got, I do not understand why they can't seem to even be able to offer this sort of product... Or move through acquisitions ... one man shops included ... Maybe too much DIY legacy.
Last edited by a moderator:
R_Lefebvre
Arducopter Developer
Yeah, I don't know. I definitely think the Pixhawk is worth $3-400, or even more. But it's strength is also it's weakness. Since it's open source, anybody can copy the hardware, and the software. So they have to remain competitive against Hobby King, RCTimer, and whoever else is selling the hardware, without investing anything in hardware or software development. On "the other website", it seems the majority are using clones. Many of them express that they think it's "unfair" that 3DR should mark up the price of the Pixhawk even $50. It's all really very strange to me, I don't understand a lot of these people.
You're right that the business model should be to drive innovation and stay ahead of the cloners. But it does not seem to work out that way. It does seem to take a long time to get new hardware designs out. But that might be due to the complexity involved. Sparkfun is not building autopilots. However, I agree, the 6S PM situation is just not good. And you're also right that they should have much better multirotor offerings. I have offered to help them with that, but they weren't interested.
I think the problem there is that they are not interested in the machines. They are interested in solutions. They're right, that people want results, not toys. People want crop data, not a drone. So they focus on getting the data. But IMO it's getting to the point where the hardware is holding back their ability to go get the data. As you stated, 14 minutes for a mapping multirotor is not very good. It wasn't even very good 1 year ago.
You're right that the business model should be to drive innovation and stay ahead of the cloners. But it does not seem to work out that way. It does seem to take a long time to get new hardware designs out. But that might be due to the complexity involved. Sparkfun is not building autopilots. However, I agree, the 6S PM situation is just not good. And you're also right that they should have much better multirotor offerings. I have offered to help them with that, but they weren't interested.
I think the problem there is that they are not interested in the machines. They are interested in solutions. They're right, that people want results, not toys. People want crop data, not a drone. So they focus on getting the data. But IMO it's getting to the point where the hardware is holding back their ability to go get the data. As you stated, 14 minutes for a mapping multirotor is not very good. It wasn't even very good 1 year ago.
Old Man
Active Member
Rob,
I've had issues with cloned hardware and software for a very long time. Even open source stuff should have some level of copy protection to prevent those not involved in the development from turning a profit on the work of others that receive nothing for their efforts. In RC, EVERY gas engine being sold by the Chinese has been copied from developers/manufacturers that put a lot of time, effort, and money into developing and producing a high quality, durable product only to see a one or more copies of the product hit the market within months of the original release. Those copies sold for up to 50% less than the original, generally were lower in performance and quality, and significantly cut into the designer's market share and profitability. In some cases the original makers have gone out of business because of the copy infringements. Might want to call that a violation instead of an infringement because the practice has been so blatant. In some cases parts inside the copies were found to be OEM parts stolen from the company that was hired to make the parts for the OEM.
This is a very serious problem and prohibits a lot of people from putting quality parts on the market. Why make something you know to be superior and highly beneficial when you know that you'll spend all the money to produce just to see someone steal your design and undercut you with a a copied product of lower quality? Worse is the fact that since the practice of cloning RC products began there has been no warranty support from the manufacturers. The cloners push for dealers world wide, who in turn make most of the sales but when things break, and many of them do because of lower quality levels, the dealers perform work on the products sold because the clones stated there was a warranty. However, the dealers end up doing all the work out of pocket because the makers never reimburse the dealer for parts or labor. Pretty soon the dealer quits selling the product and the cloner start advertising for a new group of dealers to be next on the list of suckers.
I've been in and around the RC hobby for a very long time and the stuff that's happening with multirotors is a 100% copy of everything that has been going on for the last 20+ years with the RC hobby manufacturers. Quite a few of the people turning our MR cloned parts have been part of the group cloning RC long before MR's got started. The problem won't be fixed, and the quality levels will not rise, until a means is found that prevent people from cloning the software that drives the hardware. The cloners want to take the easy road and let someone else bear the cost of design and development. Once it gets hard to copy they'll move on to some other easy to copy product. We won't see better products until this situation is corrected.
I've had issues with cloned hardware and software for a very long time. Even open source stuff should have some level of copy protection to prevent those not involved in the development from turning a profit on the work of others that receive nothing for their efforts. In RC, EVERY gas engine being sold by the Chinese has been copied from developers/manufacturers that put a lot of time, effort, and money into developing and producing a high quality, durable product only to see a one or more copies of the product hit the market within months of the original release. Those copies sold for up to 50% less than the original, generally were lower in performance and quality, and significantly cut into the designer's market share and profitability. In some cases the original makers have gone out of business because of the copy infringements. Might want to call that a violation instead of an infringement because the practice has been so blatant. In some cases parts inside the copies were found to be OEM parts stolen from the company that was hired to make the parts for the OEM.
This is a very serious problem and prohibits a lot of people from putting quality parts on the market. Why make something you know to be superior and highly beneficial when you know that you'll spend all the money to produce just to see someone steal your design and undercut you with a a copied product of lower quality? Worse is the fact that since the practice of cloning RC products began there has been no warranty support from the manufacturers. The cloners push for dealers world wide, who in turn make most of the sales but when things break, and many of them do because of lower quality levels, the dealers perform work on the products sold because the clones stated there was a warranty. However, the dealers end up doing all the work out of pocket because the makers never reimburse the dealer for parts or labor. Pretty soon the dealer quits selling the product and the cloner start advertising for a new group of dealers to be next on the list of suckers.
I've been in and around the RC hobby for a very long time and the stuff that's happening with multirotors is a 100% copy of everything that has been going on for the last 20+ years with the RC hobby manufacturers. Quite a few of the people turning our MR cloned parts have been part of the group cloning RC long before MR's got started. The problem won't be fixed, and the quality levels will not rise, until a means is found that prevent people from cloning the software that drives the hardware. The cloners want to take the easy road and let someone else bear the cost of design and development. Once it gets hard to copy they'll move on to some other easy to copy product. We won't see better products until this situation is corrected.
Old Man, I agree with you that when it comes to cloned hardware and software there are several issues. But I think there’s a significant difference here between the 3DR, Ardupilot and APM/Pixhawk and cloning situation, and the RC Hobby manufacturers getting cloned you mention.
Pixhawk was not funded and developped by 3DR. All the hard design and implementation work was done by Lorenz Meier and his team at ETH. It was open-source from the get go, “non-profit”, and if anything funding came from the taxpayers of Switzerland, or European grants, or whatever public sources or grants the project and university was funded with. Sure 3DR and Chris Anderson have done a great job championing it and commercializing it, along with working on a few final product issues, but it is nothing compared to the enormous amount of original work. Likewise, Ardupilot, at least until Randy, Tridge and others started being funded/sponsored by 3DR, did not come to be through the cold hard cash of for profit 3DR, or any other business for that matter. Development was the work of many unpaid volunteers with no affiliation to 3DR. Again Chris A. and now 3DR deserve many kudos for championing, facilitating, and being part of the movement, but that’s a far cry from “spending all the money to produce just to see someone steal your design ...”. 3DR got the open source hardware design of Pixhawk essentially for free, same with Ardupilot. That’s a heck of a sweet deal for a business.
So in this case it has been a two way street. 3DR got a lot of work for free, (originally pretty much all of it), and still is, that it can now use in a package where it can make money. Given this, one cannot expect 3DR to have it’s cake and eat it too, use something free and now change/protect it/close it for it’s own profit. And just as I wish that the cloners could give back more to the community, I also wish 3DR also gave back more to developpers. In a sense, there are many instances where the very criticisms aimed at cloners apply to 3DR too.
The cloners are not stealing anything, because whatever it is they are taking was not only free from the get go, it was actually designed and developped in a framework that actually encourages free use and copying. That’s FOSS. And to his credit, Chris Anderson actually welcomes clones and copies, and has said so publically many times.
(That said, this framework also encourages improvements, and legally requires attribution and changes to be published. That’s where the buck stops for me, and there’s no question that there’s been some serious issues with this and cases where attribution and/or changes were not published back. But that’s another story ... )
Rob’s work is actually a good example of these rather new economic dynamics. 3DR is essentially getting TradHeli for free, minus some small integration overhead for Randy, Tridge and others sponsored by 3DR. Quite a sweet deal for 3DR, who can now add Heli support to “its line”, and ride and profit from that competitive advantage. But not exactly so for Rob unfortunately. (Although I am convinced his time investment will pay off, sooner or later). Could 3DR actually complain if TradHeli was cloned? Would’nt make sense to me. Could 3DR decide to implement some sort of copy protection mechanism? Seems to me it would not only be unethical, but plain illegal given the existing FOSS license ... And completely out of line with the original intent and principles of both hardware and software open source that encourages redistribution.
Pixhawk was not funded and developped by 3DR. All the hard design and implementation work was done by Lorenz Meier and his team at ETH. It was open-source from the get go, “non-profit”, and if anything funding came from the taxpayers of Switzerland, or European grants, or whatever public sources or grants the project and university was funded with. Sure 3DR and Chris Anderson have done a great job championing it and commercializing it, along with working on a few final product issues, but it is nothing compared to the enormous amount of original work. Likewise, Ardupilot, at least until Randy, Tridge and others started being funded/sponsored by 3DR, did not come to be through the cold hard cash of for profit 3DR, or any other business for that matter. Development was the work of many unpaid volunteers with no affiliation to 3DR. Again Chris A. and now 3DR deserve many kudos for championing, facilitating, and being part of the movement, but that’s a far cry from “spending all the money to produce just to see someone steal your design ...”. 3DR got the open source hardware design of Pixhawk essentially for free, same with Ardupilot. That’s a heck of a sweet deal for a business.
So in this case it has been a two way street. 3DR got a lot of work for free, (originally pretty much all of it), and still is, that it can now use in a package where it can make money. Given this, one cannot expect 3DR to have it’s cake and eat it too, use something free and now change/protect it/close it for it’s own profit. And just as I wish that the cloners could give back more to the community, I also wish 3DR also gave back more to developpers. In a sense, there are many instances where the very criticisms aimed at cloners apply to 3DR too.
The cloners are not stealing anything, because whatever it is they are taking was not only free from the get go, it was actually designed and developped in a framework that actually encourages free use and copying. That’s FOSS. And to his credit, Chris Anderson actually welcomes clones and copies, and has said so publically many times.
(That said, this framework also encourages improvements, and legally requires attribution and changes to be published. That’s where the buck stops for me, and there’s no question that there’s been some serious issues with this and cases where attribution and/or changes were not published back. But that’s another story ... )
Rob’s work is actually a good example of these rather new economic dynamics. 3DR is essentially getting TradHeli for free, minus some small integration overhead for Randy, Tridge and others sponsored by 3DR. Quite a sweet deal for 3DR, who can now add Heli support to “its line”, and ride and profit from that competitive advantage. But not exactly so for Rob unfortunately. (Although I am convinced his time investment will pay off, sooner or later). Could 3DR actually complain if TradHeli was cloned? Would’nt make sense to me. Could 3DR decide to implement some sort of copy protection mechanism? Seems to me it would not only be unethical, but plain illegal given the existing FOSS license ... And completely out of line with the original intent and principles of both hardware and software open source that encourages redistribution.
Last edited by a moderator:
R_Lefebvre
Arducopter Developer
Generally right John, except for a few things.
First, 3DR didn't really get the Pixhawk design for free. There might well have been some money flowing between 3DR and ETH. I don't honestly know. But I strongly suspect 3DR bore the cost of prototype production, which probably wasn't inconsiderable. So while they might have gotten the basic design for free, they would have had considerable cost putting it into production.
And I do receive some money from 3DR. Not a lot, basically enough to keep me in equipment. I actually quit last year because, well, I was broke. I just couldn't do it anymore. The hobby was costing me way too much. So they have been helping out with that. At this point I have a hobby that pays for itself, which is nice, but I still don't think the burden is being shared fairly. 3DR is doing their part, but I'd like to see more companies doing something.
First, 3DR didn't really get the Pixhawk design for free. There might well have been some money flowing between 3DR and ETH. I don't honestly know. But I strongly suspect 3DR bore the cost of prototype production, which probably wasn't inconsiderable. So while they might have gotten the basic design for free, they would have had considerable cost putting it into production.
And I do receive some money from 3DR. Not a lot, basically enough to keep me in equipment. I actually quit last year because, well, I was broke. I just couldn't do it anymore. The hobby was costing me way too much. So they have been helping out with that. At this point I have a hobby that pays for itself, which is nice, but I still don't think the burden is being shared fairly. 3DR is doing their part, but I'd like to see more companies doing something.
Meanwhile ...
ok, I know armchair philosophizing and what-if is all too easy. But how about 3DR using that VC money and working on things like this?
Killer with ac 3.2 I'd bet ...
https://www.youtube.com/watch?v=gJ7eTnmV4A0
ok, I know armchair philosophizing and what-if is all too easy. But how about 3DR using that VC money and working on things like this?
Killer with ac 3.2 I'd bet ...
https://www.youtube.com/watch?v=gJ7eTnmV4A0
R_Lefebvre
Arducopter Developer
I think that one is a little bit big. But I do think they should have something bigger than an Iris. 
Av8Chuck
Member
cracks me up, the goal shouldn't be to make as large a MR to fly such a comparatively small payload, it should be to build as small enough of a MR to fly that payload safely. I see people flying GoPro's with S800's, they must be the ones who think it requires an X8 with 18" props to fly a DSLR.
R_Lefebvre
Arducopter Developer
I agree Chuck. But there's a few issues here.
The reason this tends to be done, is for flight time. Everybody wants to fly a DSLR for 30 minutes. Larger rotors is the only way to do that. However, it begs the question "why do you need to fly for 30 minute non-stop?" There aren't really any good answers to that. Some applications, such as really hardcore mapping with specialized lenses would require this. I've only actually seen one person who was doing this. (and they are doing it with an 800 Heli )
For the rest, just bring multiple machines. You can have one in the air on-scene at all times. It's WAY easier to have 2 machines that can fly for 15 minutes, than a single machine that can fly for 30.
Most general mapping can be done with a 200 gram camera though. And you can fly that for 30 minutes on a 2kg quad with 15" props.
The reason this tends to be done, is for flight time. Everybody wants to fly a DSLR for 30 minutes. Larger rotors is the only way to do that. However, it begs the question "why do you need to fly for 30 minute non-stop?" There aren't really any good answers to that. Some applications, such as really hardcore mapping with specialized lenses would require this. I've only actually seen one person who was doing this. (and they are doing it with an 800 Heli
)For the rest, just bring multiple machines. You can have one in the air on-scene at all times. It's WAY easier to have 2 machines that can fly for 15 minutes, than a single machine that can fly for 30.
Most general mapping can be done with a 200 gram camera though. And you can fly that for 30 minutes on a 2kg quad with 15" props.
R_Lefebvre
Arducopter Developer
From what I've seen, mapping long distances with heavy payloads, multirotors are a dead-end. I would love to be proven wrong. But I have not seen it done yet.
200 gram camera for 15 minutes, yes. And maybe a bit more, yes. But you're not going to *fly* a DSLR for 30 minutes on a multirotor.
Oh sure you are. It's just a matter of how big you want it to be, and how much you are willing to spend. In fact that Foxtech Devourer might just be the ticket, don't know at this point, but no matter what it'll do better than anything else.
> For the rest, just bring multiple machines. You can have one in the air on-scene at all times.
Yes. But go tell that to a Red Epic dragon flyer with 60k+ worth of gear. Or (can't remember where I saw that), someone flying *two* Reds filming for 3D. For, like, 5 minutes? Granted this is not exactly your average set-up, but that doesn't mean that doubling flight time (or whatever) wouldn't be a major plus, at least in some circumstances.
R_Lefebvre
Arducopter Developer
I'd like to see it done. Well, let's see what it would theoretically take, to start with. What would you guestimate for the weight of a DLSR, and gimbal?
I was tossing around some numbers about what it would take to lift a 10kg payload for 15 minutes. It would take so much battery, that it would take a 220V 30A circuit to charge the battery in 1 hour. Or a 4000W generator. We're rapidly approaching the point where electric flight just doesn't make sense anymore.
Well, we already know thanks to Ferdinandk that a super lightened and optimized version of a similar X4 design with 28" prop can hover for two hours and nine minutes with no payload. And that that d130 X8 with 7kg payload can fly for 19 minutes on 6S Lipo 32AH, pretty close to your 10kg payload target for 15 minutes ... (https://www.youtube.com/watch?v=Nm0ChGJ-xYU). And 30 minutes with Zenmuse + Gh4 on 44AH (https://www.youtube.com/watch?v=xcPXPJeu7cs). So depending on actual weight of DSLR, may make 30 minutes with additional battery, or may not ... But it seems to me it would with Li-Ion if one were able to go through the trouble of building such a thing. And stand a very good chance on 8s or 10s ... Wether it ends up being practical is definitely arguable, though.
Last edited by a moderator:
