eevRANT: Lily Drones Folds After $30M In Pre-Sales

[EEVblog]

The Lily Drone company has folded, admitted defeat, and will not deliver after 3 years, $30M in pre-sales, $15M in VC funding.

[Muttley Snickers]

Slightly off topic alert, anyone got tips for reducing wind noise should I ever get around to doing a few outside videos, this one sounded like an aircraft was trying to land or something.   

[PartialDischarge]

I hope the Sprite drone eventually succeeds, as their portable design is interesting

[jpc]

Slightly off topic alert, anyone got tips for reducing wind noise should I ever get around to doing a few outside videos, this one sounded like an aircraft was trying to land or something.   

Animal fur wrapped around the mic is good.

[MrAureliusR]

Hey Dave, can you post the email here so we can read it?
I'm curious to see what they sent out as there is NO information on their website, twitter or facebook..

[Muttley Snickers]

Dave has already posted it in another thread, linked below.

https://www.eevblog.com/forum/crowd-funded-projects/lily-drone-folds-after-funding-dries-up/

[jpc]

The UK's Gadget Show programme on Channel 5 recently did a small piece casually trialling three 'follow me drones' that are available now, one uses a wristband, one your mobiles GPS and one uses image recognition and object avoidance. The latter did the best but all ultimately failed. The best appeared relatively OK in open ground but failed as soon as it came to avoiding things like trees on the side of the path taken by the person being followed. I can't remember the makes of the ones being tried out now as I was only half watching and only remembered it on seeing this latest vlog but if anybody is interested it is on YouTube. Search for Follow me drones comparison.

[borjam]

Animal fur wrapped around the mic is good.

R0de (an Australian brand of microphones) sells a very reasonable priced zeppelin and dead wombat set called "blimp". In case you want to use a long shotgun microphone they even sell an extension.

[Barny]

I normally don't back hardware projects.
The risk of unexpected (expensive) problems are way to high.

Even at software projects, I take a good look before I give my hard earned money.
I only backed RadicalFishGames (Cross Code) and other small game - developer at the moment.
And all running projects looking good at the moment.

[Edit]
The only hardware project I support at the moment is the "Pyra".
(A modular, Linux based handheld console / pocket netbook)
And it looks like it will be ready to ship this year.
And well, I didn't had do pay the whole money in advance.
[/Edit]

[FrankBuss]

This refund thing happened to other projects as well, like CastAR. Money making idea: Create a cool Kickstarter project, get lots of money, invest for a year, get interest rate, refund

[Fred27]

Why bother with a refund? Taking the money and running seems to be a perfectly OK business model with no comeback. I was stung with the agent smart watch (by Chris Walker - the guy who created the Netduino). He walked away with $1m.

[wraper]

This refund thing happened to other projects as well, like CastAR. Money making idea: Create a cool Kickstarter project, get lots of money, invest for a year, get interest rate, refund

The difference with CastAR is that they will still ship 3D glasses to those who got a refund.

[FrankBuss]

The difference with CastAR is that they will still ship 3D glasses to those who got a refund.

Well, I backed it and got my refund, but still no CastAR. Last Kickstarter update was more than a year ago, with the refund message, before this they wrote updates at least every month. Same as with the Lily Drone could happen to CastAR: No more VC funding, maybe because of Microsoft's HoloLens, and then that's it. Initial estimated delivery was September 2014. Incidentally Playground Global funded $15M into CastAR, same VC money Lily Drones got. Is this some standard VC amount?

[jolshefsky]

So $45,000,000 profit? Sounds like they won capitalism to me.

[wraper]

The difference with CastAR is that they will still ship 3D glasses to those who got a refund.

Well, I backed it and got my refund, but still no CastAR. Last Kickstarter update was more than a year ago, with the refund message, before this they wrote updates at least every month. Same as with the Lily Drone could happen to CastAR: No more VC funding, maybe because of Microsoft's HoloLens, and then that's it. Initial estimated delivery was September 2014. Incidentally Playground Global funded $15M into CastAR, same VC money Lily Drones got. Is this some standard VC amount?

Why would they make any updates on kickstarter? Everyone who wanted got a refund, game over. Also they explicitly said that with venture funding coming in, they cannot reveal much of the details anymore. Also they said that it won't be released till 2017, and will be a different product from what was originally planned.

[2N3055]

Slightly off topic alert, anyone got tips for reducing wind noise should I ever get around to doing a few outside videos, this one sounded like an aircraft was trying to land or something.   

Windshield for a microphone ?  Foam or a zeppelin type... That is, if you have a separate mike... For mikes integrated in camera, I'we seen people glueing a piece of foam over mike opening... Sometimes it works surprisingly well..

Cheers.

[FrankBuss]

Why would they make any updates on kickstarter? Everyone who wanted got a refund, game over. Also they explicitly said that with venture funding coming in, they cannot reveal much of the details anymore. Also they said that it won't be released till 2017, and will be a different product from what was originally planned.

This sounds strange, why would a VC forbid this for CastAR? Something like "we are on track, got first pre-production units, shipping still expected in 2017" doesn't reveal anything they didn't say already. They even didn't wrote about their new CEO and COO, which joined the company in 2016. I would have expected a bit more information, if they really care about their backers, as they wrote in their last update, because some people (like me) might think about writing games for it, but if the release date is uncertain, might postpone this.

See for example the Lily Drone blog, every few days an update til mid December 2016:

https://www.lily.camera/updates/

Hey Dave, can you post the email here so we can read it?
I'm curious to see what they sent out as there is NO information on their website, twitter or facebook..

https://www.lily.camera/adventure-comes-end/

I really wonder how they can do the refunds if they don't have the money for the production. Shouldn't it be less money for production than what you got from the buyers?

[Fungus]

Why bother with a refund? Taking the money and running seems to be a perfectly OK business model with no comeback. I was stung with the agent smart watch (by Chris Walker - the guy who created the Netduino). He walked away with $1m.

Yes, I suspect that what really happened is that they stopped all R&D when they suddenly had $45 million of party-party money in their bank accounts. Those new sports cars won't find their handling limits all by themselves. 

The drone doesn't look that difficult to make, it's just two GPS units, wireless communications, a bit of math. Three years and $15 million should easily have covered it.

[BMF]

It seems they could have at least delivered the basic hardware. If it was a camera drone by remote and was good they would still have a company. If they made the tracking work later they could release an update. If they can't offer a refund or hardware I hope someone audits the company and makes it public. These kinds of fails make it tougher for others to get funding.

[FrankBuss]

The drone doesn't look that difficult to make, it's just two GPS units, wireless communications, a bit of math. Three years and $15 million should easily have covered it.

There are lots of interesting problems. Even the big competitor DJI with its Phantoms didn't get it 100% right, like if there are some leaves in the way when the drone follows you, the props will try to do some garden work on it. And sometimes it loses you. But works pretty good in an open environment with no trees or other obstacles. There are some videos of it following a car.

It is a bit simpler for the Lily drone to follow you, because of the tracking device, but still complicated enough. And as I read on the webpage, no mechanical camera gimbal. This won't work for high quality videos.

[Fungus]

It is a bit simpler for the Lily drone to follow you, because of the tracking device, but still complicated enough.

Every third-person video game in the last 15 years has had camera tracking code in it.

Trees in the way? Just put "only for use in large open spaces" on the box.

[FrankBuss]

Every third-person video game in the last 15 years has had camera tracking code in it.

Trees in the way? Just put "only for use in large open spaces" on the box.

Even for games it is not that easy. Search for "game camera bug" on Youtube, 120,000 results

[Don Hills]

Windshield for a microphone ?  Foam or a zeppelin type... That is, if you have a separate mike... For mikes integrated in camera, I'we seen people glueing a piece of foam over mike opening... Sometimes it works surprisingly well..

Cheers.

Go to Youtube, search "video camera microphone fur".

Get a snippet of faux fur cloth from a fabric shop (ask Mrs EEVBlog). Cut it a bit larger than the microphone area and stick it on around the edges. See the Youtube videos. Might be a bit tricky with very small action type cameras with just a tiny hole in the front for a microphone.

[timb]

The drone doesn't look that difficult to make, it's just two GPS units, wireless communications, a bit of math. Three years and $15 million should easily have covered it.

There are lots of interesting problems. Even the big competitor DJI with its Phantoms didn't get it 100% right, like if there are some leaves in the way when the drone follows you, the props will try to do some garden work on it. And sometimes it loses you. But works pretty good in an open environment with no trees or other obstacles. There are some videos of it following a car.

It is a bit simpler for the Lily drone to follow you, because of the tracking device, but still complicated enough. And as I read on the webpage, no mechanical camera gimbal. This won't work for high quality videos.

You're right, the quality wouldn't be 4K Ultra Mega Maxi Super Turbo HD like a Phantom with a gimbal and a GoPro, but using using a FishEye lens and doing pan/tilt in software can be surprisingly effective. I've got a Parrot Bebop and the video quality is alright.

The other advantage to a software gimbal is no moving parts, which makes the drone lighter and much more robust. Also a lot cheaper, too.

I've crashed mine a number of times (owing to inexplicable disconnects which I think I've finally fixed), one of the crashes involved it automatically "returning home" after a disconnect. Unfortunate home turned out to be some pine trees on the edge of a field, where I had taken off from. It gracefully flew back, but because the GPS didn't have a full 3D fix when I took off, it overshot by about 10m, putting it right over said trees where it started to descend, then hit every branch on the way down. The only thing that broke was one of the plastic feet. I'm glad the return to home feature worked, because when I initially lost control of it, it was hovering 200ft over a pond!

[gmb42]

The other advantage to a software gimbal is no moving parts, which makes the drone lighter and much more robust. Also a lot cheaper, too.

I've crashed mine a number of times (owing to inexplicable disconnects which I think I've finally fixed), one of the crashes involved it automatically "returning home" after a disconnect. Unfortunate home turned out to be some pine trees on the edge of a field, where I had taken off from. It gracefully flew back, but because the GPS didn't have a full 3D fix when I took off, it overshot by about 10m, putting it right over said trees where it started to descend, then hit every branch on the way down. The only thing that broke was one of the plastic feet. I'm glad the return to home feature worked, because when I initially lost control of it, it was hovering 200ft over a pond!

/rant

And as an RC flyer this makes me very afraid that folks are "flying" such devices in public.  If a flying club member had a model or transmitter behaving in such a manner they would NOT be allowed to fly their model(s) until they had demonstrated they had fixed it, usually by means of long range on-ground testing.

Manufacturers by creating such ill-behaved devices, and users (not necessarily you timb) that then don't take responsible steps (i.e. by ground testing and having insurance) to ascertain device safety are going to kill unlicensed personal drone use along with airborne RC flying.

/end rant

[timb]

The other advantage to a software gimbal is no moving parts, which makes the drone lighter and much more robust. Also a lot cheaper, too.

I've crashed mine a number of times (owing to inexplicable disconnects which I think I've finally fixed), one of the crashes involved it automatically "returning home" after a disconnect. Unfortunate home turned out to be some pine trees on the edge of a field, where I had taken off from. It gracefully flew back, but because the GPS didn't have a full 3D fix when I took off, it overshot by about 10m, putting it right over said trees where it started to descend, then hit every branch on the way down. The only thing that broke was one of the plastic feet. I'm glad the return to home feature worked, because when I initially lost control of it, it was hovering 200ft over a pond!

/rant

And as an RC flyer this makes me very afraid that folks are "flying" such devices in public.  If a flying club member had a model or transmitter behaving in such a manner they would NOT be allowed to fly their model(s) until they had demonstrated they had fixed it, usually by means of long range on-ground testing.

Manufacturers by creating such ill-behaved devices, and users (not necessarily you timb) that then don't take responsible steps (i.e. by ground testing and having insurance) to ascertain device safety are going to kill unlicensed personal drone use along with airborne RC flying.

/end rant

I know this wasn't directed at me, but I'll mention it anyway: I was flying on private property, in the middle of nowhere, with no buildings, cars or people around.

I do agree there are *a lot* of reckless drone operators out there who think they basically fly themselves and don't require any skill to operate. There's hundreds of videos on YouTube proving this. (There's one of a guy who took his Phantom straight up (800+) feet, above the clouds, in a residential area, wherein it stopped responding... He finally managed to get it responding after switching off the GPS position lock, then it started drifting as he was bringing it down; he got it to 400ft before the batteries gave out. Luckily it was over a small park, he could have killed someone!

That said, some issues can't always be reliably ground tested. I once watched as a 20+ year RC vet took his plane up and then proceeded to lose contact with it. This was after spending weeks tracking down a problem with (and bench testing) the receiver. The plane kept going, never to be heard from again.

In my case, I'd flown the drone successfully about a dozen times, no problems, then one day it just stopped responding and flew into a tree. The next day I tested it at very low altitude (ran two batteries through it) and everything seemed fine, so I popped a new battery in and started taking some footage, about 8 minutes in it disconnected again and crashes as I said in my last post.

Nearest I can figure, something's wrong with that particular battery, so when it gets below 30% there's a brief voltage sag or something which causes the WiFi module to brownout.

I haven't had a chance to test it yet, but my plan is to run it back and forth across the field with that battery and see what happens.

Some issues can only be tested in flight, because they only occur during specific conditions.

[EEVblog]

The plot thickens:
https://techcrunch.com/2017/01/12/sf-district-attorney-lawsuit-against-lily-may-have-prompted-refund/

[FrankBuss]

The plot thickens:
https://techcrunch.com/2017/01/12/sf-district-attorney-lawsuit-against-lily-may-have-prompted-refund/

From the article:

Part of the suit has to do with the initial pitch video, watched by millions of people, showing off what appeared to be a Lily drone following users and shooting video. The drone responsible for all that fancy aerial work and video was not in fact a Lily, but a DJI Inspire, something the creators failed to mention.

For those who don't know it, the DJI Inspire is the professional model series from DJI, starting at EUR 2,299 for a basic Inspire 1, and up to EUR 7,098 for the Inspire 2 Premium Combo, used to record movies in cinema quality.

[langwadt]

The plot thickens:
https://techcrunch.com/2017/01/12/sf-district-attorney-lawsuit-against-lily-may-have-prompted-refund/

From the article:

Part of the suit has to do with the initial pitch video, watched by millions of people, showing off what appeared to be a Lily drone following users and shooting video. The drone responsible for all that fancy aerial work and video was not in fact a Lily, but a DJI Inspire, something the creators failed to mention.

For those who don't know it, the DJI Inspire is the professional model series from DJI, starting at EUR 2,299 for a basic Inspire 1, and up to EUR 7,098 for the Inspire 2 Premium Combo, used to record movies in cinema quality.

http://dilbert.com/strip/1996-07-09

[Gary350z]

Slightly off topic alert, anyone got tips for reducing wind noise should I ever get around to doing a few outside videos, this one sounded like an aircraft was trying to land or something.   

Watch this video:

Reducing wind noise when recording video

[Gary350z]

The drone doesn't look that difficult to make, it's just two GPS units, wireless communications, a bit of math. Three years and $15 million should easily have covered it.

My Hubsan H501S drone will follow me around, has an HD camera, and FPV, and only costs $230. 

[SpaceCow]

I too an very interested to find out what went wrong with this. The fact that they used pre-orders on their website instead of a kickstatrter-type system seems to point to a non-nefarious operation, as does the fact that they are returning everyone's money. On the other hand, the fact that they were unable to even come close to producing a product that does what lots of current products already do at a much lower cost than Lily leads one to believe that there must have been some kind of foul play. The most interesting result would be if no one actually stole the money and they somehow honestly tried and failed with $15 million. I'd love to see what went wrong.

[EEVblog]

It's looking worse by the day:
https://www.suasnews.com/2017/01/the-people-vs-lily-robotics/

Here is an extract from the DA’s complaint against Lily Robotics:

“In May 2015, Lily Robotics used a false and misleading promotional video for a camera drone to obtain over $34 million in “preorder” sales from customers all over the world.
A “preorder” is when a customer pays a seller for a product with the understanding that the product is not yet available to be shipped but that it will be shipped by some stated future date.
In addition, Lily Robotics did not refund those customers their money when it delayed for almost a year the shipping dates for its product, as the law required. Instead, it told its customers that the preorder money: “remain[ed] untouched,” was in “cold storage,” and was not being used “to run the company.” This, too, was false and misleading. According to a witness, Lily Robotics had actually used that preorder money as collateral to obtain a $4 million load so it could continue its operations.
It is now January 2017. Lily Robotics still has not shipped a single product to any of its customers, and, according to a witness, it is running out of non-preorder money.

[jonovid]

my idea of sub $300 technical improvements for any compact quadcopter used for aerial photography.
IF I was selling aerial photography compact quadcopters.  by compact is to say it will fit in a backpack.

I would IMO add High Visibility colors to the quadcopter & a onboard piezo siren that can be activated by the pilot. to help find a lost quadcopter.
add an infrared LED collision avoidance system,  simple  idea is to add 8 to 16 infrared LED's and infrared sensors to the quadcopter.
to stop quadcopters from colliding with walls floors & ceilings indoors. but with the outdoors its from colliding with trees & buildings.
pulsed infrared sensors that detects objects that the quadcopter is about to collid with. and pulls an automatic split second turnaround.
 by embedded software. job for programmers.
also add an anti-wind drift system by useing an optical mouse sensor on the bottom of the quadcopter looking down at the ground.
that will keep a quadcopter in a fixed hover. alternative to GPS without constant joystick input to hold a fixed position. even in a wind.
a optical mouse sensor detects if the ground is moving and embedded software removes the quadcopters tilting from the calculation
leaving pilot input and drift. so this is an autopilot that stops non piloted quadcopter drift.
also an integrated quadcopter video system, that gives Standard definition as FPV to the pilot, without a time lag.
but also gives video to an onboard High Definition Digital Video Recorder for aerial photography.
also the FPV video to the pilot . Heads Up Display needs to show countdown timer from a calculation of the battery voltage.
also 2.4Ghz up link signal strength. vr the down link. with a beep for low! must return to base.

also need a signal test preflight checklist for the video so you get a coordinated 5.8ghz channels frequency set
FPV Channel video TX frequencies used by the quadcopter. available frequencies for the down-link.
if you have two or more FPV quadcopters in the same RC radio zone.
 end of rant

[EEVblog]

Haven't seen my refund yet...

[timb]

my idea of sub $300 technical improvements for any compact quadcopter used for aerial photography.
IF I was selling aerial photography compact quadcopters.  by compact is to say it will fit in a backpack.

I would IMO add High Visibility colors to the quadcopter & a onboard piezo siren that can be activated by the pilot. to help find a lost quadcopter.
add an infrared LED collision avoidance system,  simple  idea is to add 8 to 16 infrared LED's and infrared sensors to the quadcopter.
to stop quadcopters from colliding with walls floors & ceilings indoors. but with the outdoors its from colliding with trees & buildings.
pulsed infrared sensors that detects objects that the quadcopter is about to collid with. and pulls an automatic split second turnaround.
 by embedded software. job for programmers.
also add an anti-wind drift system by useing an optical mouse sensor on the bottom of the quadcopter looking down at the ground.
that will keep a quadcopter in a fixed hover. alternative to GPS without constant joystick input to hold a fixed position. even in a wind.
a optical mouse sensor detects if the ground is moving and embedded software removes the quadcopters tilting from the calculation
leaving pilot input and drift. so this is an autopilot that stops non piloted quadcopter drift.
also an integrated quadcopter video system, that gives Standard definition as FPV to the pilot, without a time lag.
but also gives video to an onboard High Definition Digital Video Recorder for aerial photography.
also the FPV video to the pilot . Heads Up Display needs to show countdown timer from a calculation of the battery voltage.
also 2.4Ghz up link signal strength. vr the down link. with a beep for low! must return to base.

also need a signal test preflight checklist for the video so you get a coordinated 5.8ghz channels frequency set
FPV Channel video TX frequencies used by the quadcopter. available frequencies for the down-link.
if you have two or more FPV quadcopters in the same RC radio zone.
 end of rant

A lot of this already exists. For instance, the Parrot Bebop has a 6MP OVA CMOS camera on the bottom and uses the video stream to keep the unit from drifting when hovering.

It also has a loud piezo sensor that activates if the unit crashes or is forced to land due to a low battery. Plus it constantly transmits GPS coordinates to the controlling phone/tablet/SkyController, which allows you to pull up a map on your device and see exactly where it is.

IR would be useless for collision avoidance outdoors. Sonar or radar would be much better, however it can also be done with the main front facing camera and software processing. That's the cheapest option (but the hardest to get working reliably.)

[cdev]

There are (multiple input capable) SDR front ends suitable for RTK GPS which I suspect could give you a precise positional solution in real time in all three dimensions- you can use several GPSs with the antennas as far apart as possible to give you the UAV's attitude.  Thats likely what pros use for photo-mapping applications.

[EEVblog]

Raided!
http://www.forbes.com/sites/ryanmac/2017/01/31/failed-drone-startup-lily-robotics-raided-for-possible-criminal-investigation-says-source/#315d56c13835

[Brumby]

Technically, just rumours and speculation ... but a troubling report nonetheless.


This also does not feel encouraging...

“I am worried that a lens geek could study our images up close and detect the unique Gopro lens footprint,” Lily CEO Antoine Balaresque wrote in an email to the filmmaker hired to make the film. “But I am just speculating here: I don’t know much about lenses but I think we should be extremely careful if we decide to lie publicly.”

[EEVblog]

They finally declared bankruptcy
https://cases.primeclerk.com/lilyrobotics/
http://www.eevblog.com/files/LilyBankruptcy.pdf

[EEVblog]

And you thought this one was over!

This morning I got an email from Youtube saying there was a Trademark claim against my Lily rant video, from non other than the new owner of the Lily IP, Mota (www.mota.com) another drone company. Likely paid pennies on the dollar it.

Yes, a Trademark claim, not a DMCA copyright claim.
Absolutely laughable, as my Trademark Attorney agrees.
I'm fully protected under the Nominative Fair Use (a.k.a Trademark Fair Use) (1992) U.S. Court of Appeals for the Ninth Circuit
I emailed them to sort it out and their response was even more laughable.
They have about 24 hours left to remove the claim before it automatically goes to a Youtube dispute of which of course I will appeal, along with unleashing the internet flood gates on this.

And I wasn't the only one hit.

Stay tuned...

[jpc]

When will these companies learn that it is much better to just ignore things like Dave's piece, in which case the odds are it will eventually slip down the google results page and out of sight. But instead they have to poke the hornet's nest and drag it all back to the surface again. Haven't they heard of the Streisand effect?  Time to get the popcorn out?

[EEVblog]

Haven't they heard of the Streisand effect?  Time to get the popcorn out?

I coincidentally have an unreleased video already uploaded where I mention that exact same thing.
Was going to release this morning, but decided to cut them some slack and see what they came back with, which as I said was laughable.

[jpc]

Cool, I shall look forward to your video 😈

[EEVblog]

Mota have very wisely avoided corporate suicide and decided to withdraw the Trademark request.

[jpc]

Wise move on their part. But it means I'll have to put away the popcorn 😕. Well at least until until the next time some company tries it on 😀.

[EEVblog]

They are back!
Clearly they had this drone ready to go, and just paid $750k for the Lily Trademark and slapped the brand on it.

https://petapixel.com/2017/08/31/lily-drone-back-grave-ready-try-another-takeoff/

[EEVblog]

Wise move on their part. But it means I'll have to put away the popcorn 😕. Well at least until until the next time some company tries it on 😀.

Free unreleased popcorn fodder: