Cinewhoop // The Shendrone Squirt

The journey for my involvement with the squirt is a little bit of a weird process. To be honest, I’m not sure where it all started or why, but I have some ideas and thoughts that I want to pursue with you. First of all, I wanted to call out all of the inspirational material listed below. Before I ever started down this journey, these clips and videos were inspirations to me, and they are the true reason that I’ve been playing around with this thought. Please give them a read and see why each of them has impacted me in some way.

Okay, now that we’ve gotten that out of the way, we’ll go back to the beginning. I believe that my pursuit of this all started when I saw the “Onnanocos” video a few days before DRL found it. I saw the video and watched it a few times. While this was not the first of its kind that I had seen, I remember that is stuck with me better than anything else. The choreography, taking time to think things through ahead of time, and going places that cameras could never go made this really special to me.

A day or two after I initially saw the video, DRL marketing reached out to me and asked me if I would be interested in trying to create a similar kind of shot at an upcoming DRL race. At the time, I was helping lead the charge with DRL’s “specialty shooting” team, where we would film little projects within the show that would be used for social marketing at a later date. I didn’t have a drone that would work, so I reached out to Shendrones to see if they had something that fit the bill. Turns out, they did, and it was a project that had just slightly stalled out. That being said, we were excited to work together to flesh out the project a little further.

Unfortunately, the gig with DRL ended up falling through, and we were unable to shoot the project that we wanted to while at the event. But, since Andy (Shen) and I had already started down this journey, we aimed to see it all the way through.

Eventually, I got my hands on a prototype version. Beautifully cut and printed by Shen, I jumped in and built it up right away for testing.

I built the machine, and started tuning right away. When I first got it assembled, I found that there was a lot of slop due to the weight of the machine, so I tuned P values up by around 10-15 points, and brought the D term up by 5 points to match from default 3.3.1 (which was current at the time). I also increased the I term by about 5 points to help fight against the error introduced by wind, breezes, or prop wash when low to the ground.

I flew it very cautiously for the first handful of flights and tests that I did with the machine, almost never doing any kind of hard maneuvers or major elevation changes. While I don’t think that this machine will ever be used for any intense kind of acro or anything like that, I DO want to be able to count on the machine to go in and out of dives, as well as change elevation with confidence, and then have a little bit extra left in the tank for “oh s*** moments.” When I started doing more and more aggressive maneuvers and flying at higher speeds, I found that the quad would have some major issues, including, but not limited to big hard yaw-pulls to the right. Essentially, the whole quad would torque, go to full throttle to try to get out of the yaw torque, and then continue climbing and being annoying. These sorts of results I couldn’t count on, and I almost lost the drone attempting a flight in Colorado, where it would have been very tricky to recover.

I pushed back on Shen, stating that I couldn’t get my drone to fly right, and I was wondering what special sauce he was doing to his to get it to work so well. Turns out.. Nothing. He was running weaker motors on older electronics and firmwares with a stock tune — didn’t change anything out of the box. Meanwhile, to get rid of the issues I was having I tried:

      • New motors
      • New FC
      • New ESC
      • Tight tune
      • Loose Tune
      • Moving the battery around to rebalance the whole machine
    • Different kinds and blades of props

After nearly two months of mucking about with it, one difference that Andy and I had not yet resolved together was that he was using custom-cut props. The reason for this is that, since the Squirt is a ducted quad, it needs very very tight tolerances between the propellers and the edge of the ducts. Otherwise, the thrust benefit that you get from having ducts is practically useless, and then at that point it’s just dead weight. If you read through the article on Shendrones, Andy states that “At full throttle it was 1953g with ducts, 1710 without” – and that in a different test “I got 616g with ducts on, 554g with ducts off. That 1mm gap [caused by having non-perfect props] was bothering me, so I broke out a fresh set of props and slowly sanded them down until they just started to spin freely. With the improved tolerance thrust went up to 648g.”

Andy found that there was a non-insignificant difference between running perfectly shaped props and not.

Thus, as a last-ditch effort, Andy and I tried two last things. The first was that he sent me a set of props that he custom-cut. I believe that these are originally HQ 4x3x3 props, sanded down to be just barely in contact with the edge of the ducts. Running the motors would self-sand the props down the rest of the way so that the tolerance is as low as possible. The second thing that I changed was battery placement. I had consistently been moving the battery to the very rear of the drone so that the weight was evenly distributed across the entire machine, rather than tending to be very nose-heavy. However, after thinking harder on it, I’ve decided that this was probably also contributing to the negative flight characteristics I was experiencing. By having the weight spread out across the yaw axis, rather than more centralized, it was harder for the weakest fulcrum on weight distribution to keep up, causing many of the bad flight characteristics I was experiencing. Thus, despite that I’m causing the quad to be unbalanced by having the battery sitting across the middle of the top plate, the drone flies better than fully distributing weight across the entire yaw plane.

With these two items buttoned up, the drone flies MUCH better and smoother, and I couldn’t be happier with it. I’ve been using it on a number of different shoots with great success, and it’s been doing its job perfectly.

Inspirations

Japanese Band: Onnanocos

While this was not the first of its kind that I had seen, I remember that is stuck with me better than anything else. The choreography, taking time to think things through ahead of time, and going places that cameras could never go made this really special to me.

Muscle Up

Made by the creator of Reel Steady, Robert Mcintosh I think that this is the video that most of us have been inspired by primarily – it has it all. ReelSteady, gapping, highs, lows, and a bit of a story. It’s single-take movie making nirvana.

Pretty Sweet

Another Mcintosh special, this intro video for “Pretty Sweet” is definitely the first shot that I’ve ever seen that did a single take drone shot. By today’s standards it’s fairly tame, but look at the date it was recorded: 2012. For reference, the DJI Phantom wasn’t released until January 2013. These guys were and are so bleeding edge, it’s hard to even comprehend.

Hooza Good Boy

This is just a perfect example for me, and while it’s not mind blowing or earth shattering in any way, it’s just a video that has impacted how I think about making shots along the way.

Baby Driver: Coffee Scene

Edgar Wright’s “Baby Driver” is by far my favorite movie. The filmmaking is incredible, and if you haven’t watched videos about why it’s incredible you need to. But, this scene in particular has inspired the way that I think about framing and building a story even in just a small, pointless section of a film like this title sequence. Also, yes, I know that this wasn’t filmed with a drone, but it IS an incredible example of storytelling with this choreographed, single take style.

Everything from Airblastr

Airblaster in my mind is the king of Cinematic mini drone antics. The guy can fly, frame, think, and edit better than the best out there, and if you haven’t seen his stuff, you need to go check it out on instagram

Gab707

Another pioneer in the reelsteady/storytelling approach with mini drones, Gab707 first inspired me with this awesome shot running up a mountain, but what really stood out to me was his use of sound. You’ll find that his videos always overlay waterfalls or birds or wind on top of the footage to make the whole experience more immersive. That has always stood out to me and inspired me.

Why the Squirt

    • Ducted, in ideal conditions, adds ~200g of thrust, making 3” motors more capable of carrying the heavier weight of the hero7 without having to remove the casing of the GoPro itself
    • Why not 250g? In order to achieve a 250g weight so that the drone counts outside of drone registration laws and qualifies as a toy, you have to do damage to the GoPro and build a machine that is not safe. If you go over 250g, allowing yourself to not disassemble gopros and make small drones, you have a much safer, more durable quad that can be counted on for many shoots.
      • Fly near soft targets. While the Squirt is not perfectly safe, and it could still cause a good deal of damage to a person, it’s much safer than any other drone capable of carrying the same amount of weight, so there is no sacrifice in image quality or safety.
    • Go places other cameras can’t. By being able to go high, low, in, out, through, and over, this camera platform can outperform any other method of camera movement I’ve ever seen. While it cannot carry a big cinema camera, you can use a GoPro Hero 7 combined with reelsteady to make incredible images. Large brands like Kohls and RedBull have even started using these techniques, and I’m sure that more will follow in their ranks.

What the Squirt isn’t:

    • A racing drone
      • It’s not fast — it’s actually quite slow and quite overweight for what it is. It will never be something that you can get performance out of
      • A freestyle drone
        • The drone can be used to fly around proximity and be happy with the way that it flies, but it’s not built to be used for freestyle. Remember: it’s another arrow in the filmmaking quiver, not a drone to just go out and have fun with
    • A beginner drone
      • Due to the weight of this machine and the complication of getting the props just right, it’s not a beginner drone really. With some weird flight characteristics and the complexity of the ducts, it can be cumbersome. There are beginner drones that have bumpers for the props, but this is not one of those.

Questions and Answers

On social media, I asked what questions people would have about the Squirt so I could pull together a few answers, so I’ve grabbed a selection to talk about here.

    • How do ducts affect flight performance and handling?
      • So for type of flying that I’m aiming to do, which is, for the most part, straight and level, I haven’t really had a problem with ducts affecting flight performance except when there’s strong winds. Strong winds make this thing pretty hard to use, that’s for sure. In the little bit of acrobatic flying that I HAVE done with the machine (again, remember that it’s not really designed to be flown like a miniquad as much as a flying camera), it does okay. I’ve found that at the extreme ends of some maneuvers, it will get squirrely, but for the most part behaves like you would expect — just a big, heavy quad. In fact, I would argue that most of the weird performance issues that it has might be more the result of the lack of thrust/weight more than anything else.
    • Curious about ur PIDs setup
      • I’ll post a full dump of all of my PIDs and rates attached to this post. The short version for PIDs for this is that they were too “relaxed.” I believe that I bumped everything up by a good margin to lock it in more.
    • How much does the weight of the GoPro affect performance?
      • So the way that I look at it like this is that the GoPro is part of the quad. This is not designed to be flown without it, because it’s a flying camera, not an acrobatic drone. Thus, asking how it affects performance is like asking “how does the windshield affect your car’s performance.” It doesn’t matter – it just is haha.
    • What is the Camera and gopro angle? Kv? And how much weight? It looks pretty cool
        • I run about 15 degrees on the FPV camera and 10 degrees on the gopro.
          • The KV of the motors is 3600 kv.
        • The weight is JUST shy of 500g (493.8) with a gopro hero 7 and an 850 mah tattu 4s.
    • Recommended flight modes for use with cinematic shots.
      • So the whole beauty of this machine is that it’s a small, full manual drone, that can shoot extremely high quality video. Therefore, I’m exclusively using acro mode to film with this thing because that’s the way it was designed to set itself apart. I don’t think that angle or horizon mode would work for this in any way since they “bounce” back to level, which is not a very cinematic quality.
    • Ducts effect on performance and durability? Delamination?
      • I’ve already addressed the performance aspect, but I did want to mention durability. That’s actually a really good question. I have crashed this thing a number of times from 20 feet up and had some good blips into the ground from the testing phase, but I haven’t broken anything yet. And, even if the ducts do break, I’m sure it’s not expensive to get another set printed by Andy.
      • Is it safe indoor?
        • I wouldn’t call this thing safe. Unlike tiny-whoops or small small brushless micro drones, this is a big, heavy quad. It weighs upwards of 500g, so there’s a lot of mass behind it, and it still is running 3” brushless motors inside of the props, so if you accidentally get a finger or hair in there, you will be quite sad.
    • Will you post some sample video before stabilization?
      • Yes.

Gallery

Build List

10 thoughts on “Cinewhoop // The Shendrone Squirt

Leave a Reply

Your email address will not be published. Required fields are marked *