Dragonframe’s DMC-16

Hi Folks. So i have a couple current problems that seem to stem from this Dragonframe DMC-16

SEE UPDATE 12/5/2016 at the end

I figured i would spend some time going over it, what it can do, what it is having problems with, and what I am HOPING to get resolved via Dragonframes support department.

That is the DMC-16.

It provides several critical functions

  • Motion control

    • The moves are all scripted out in the Dragonframe software on the PC(or mac), but the computer needs a way to tell the stepper motors how to move. The DMC provides the required connections to provide Step and Direction commands to the motors.  Stepper motors dont just spin like DC motors when current is applied. Instead, you tell it what position you want the shaft to be in. There are a normally a total of 200 positions for a stepper motor shaft. That means it can move 1.8 degrees with each step.  For these motions to work, there is something called a Stepper Driver which provides power to the motor and gets it to those positions, but something has to tell the Driver what to do. This is where the DMC-16 comes into play. Tell it to move the PAN motor 100 steps clockwise, and the motor will rotate the shaft 180 degrees. Tell it to move the same distance counterclockwise, and it will return the shaft of the motor to the position it started. This sort of logic allows these stepper motors to conduct highly coordinated and very precise moves. In essence this portion of Dragonframe is just a fancy CNC machine. Not much different from the lathes, routers, plasma cutters, used by machinists to cut out very precise shapes.
      With Dragonframe you create Keyframes. You can think of these as targets for the system to move the camera. For Keyframe A you want the camera pointed at the stem. Keyframe B you program to be one of the leaves, Keyframe C you set to be the flower.
      Dragonframe software then figures out the best way to move the camera from A to B to C, and will smooth out the motion to give everything a nice smooth look.
      One of the special things about dragonframe, is once you have that A to B to C figured out, you can have it break that movement down into thousands of individual discreet movemensts. So you dont have to tell the camera where to aim on every single image it takes, rather give it the general directions of what you want and DF will figure out all the stuff in between.
      You can also run test runs and snap screenshots from the live view to assemble test videos to show the motion and make sure it is what you are looking for.
  • Camera Control

    • Well of course you need to control the camera too. Dragonframe is not a CNC machine for cutting out parts, it is a CNC machine for making movies! It has the ability to integrate with various DSLR’s to provide camera controls from the computer via tethering, it also provides a live view of the set to help you compose your scenes andand takes. When running the routine it will even take the images straight off the camera and store on the PC rather than the internal memory card. So no worries of moving the camera on accident when retrieving the mem card.
  • DMX Lighting controls

    • This is something that I am super excited about. Many people in film, music, or other entertainment fields know all about DMX. This is basically aa lighting program. Think of your favorite concerts, all the lighting that happens, or the lights at a club when the DJ is spinning. These are all likely to be DMX lights. You have the ability to turn lights on or off on a pre-recorded program. Some are on off, some can be dimmed to varuious levels, and some of them (like the ones I use) are RGB LED lights where you can shift to any color, red to blue, green to yellow, day to night, suimulate sunrise and sunset. I have not had much time to play with these yet, however I have recently moved all my lights to DMX control and away from the Biolapse Control Module(more on that later)

  • Integration with other equipment.

    • The DMC 16 also has a whole host of connections that provide inputs, outputs, triggers. Everything you need to get the hardware to all integrate together.
    • Laptop to DMC-16 connects via USB
      DMC sends the step/direction signals to the stepper Drivers via RJ-11 interfaces.
    • Input trigger from external source that can be prorgammed to trigger a camera or other various things.
    • External trigger that can connect to an external system for motion control or trigger work lights, or whatever
    • Camera trigger to control the focus and shutter of a camera
    • Limit switches which prevent the hardware you spend all this time working on from accidentally moving too far or wigging out and damaging itself or you.
    • Emergency Stop connection to shut the motion hardware down in case shit hits the fan.


It has a huge amount of capability, but it is also very expensive. about $1800 with shipping. Add in the $300 software and you have a sizeable investment on this system. That does NOT get you any of the hardware, motors, drivers, gears, cameras, cables, laptop, lights, dimmer packs, or anything else that goes along with it.


However when you look at the amount of capability, nothing even comes close. This software and the DMC-16 have been used on some very impressive stop-motion animations that killed it in the Box Office including Boxtrolls, Kubo and the Two Strings, Shaun of the Sheep…. When considering how much you CAN do with it, and the huge amount of incredibly well though out features, even $2100 for this control system is an absolute bargain. This is the standard in Hollywood for stop motion animation.

It Aint Perfect.

Now this is where I stop boasting about it and throw it right the hell under a bus for a minute.

I like this system, I like Dragonframe, but i have been running into some issues. I have some messages in with the company, I am hoping they can help me out.

1) External Trigger.

Mine stopped working. There is a 3 prong DMX output used to trigger the camera. Why they used this I have NO idea. There are NO camera cables I have ever found that terminate to this connection, so I had to make my own. Luckily I have built hundreds of camera trigger cables in the past 5 years. It is not terribly difficult, but a 2.5 or 3.5 TRS connector would make more sense to me.

Biolapse is a time consuming process. Some shoots will take days, some will take weeks, even a month. It is important that I am able to maximise my results by constantly shooting multiple cameras. I have the 6D on the main rig, two Sony A7’s, a Fuji XE-1, Fuji XT-1, and a Panasonic GH4. I like to have 4-5 cameras shooting at a time. Otto is my main rig with 8 degrees of movement on the main stage. another 5 Axis rig that gives linear, focus, pan and tilt and rotary for the 2nd stage. And then there will be a third stage behind the Main with a turntable and a static camera.  That is 3 cameras right there. Then I have the option of setting up a couple static cameras for additional footage, or for some BTS work. All of these cameras have to shoot at the same damn time. They need to be triggered by the SAME source.  To pull this off, i built a 10 channel splitter, each channel is discrete using optocouplers which is sort of like a relay, but has no moving parts to wear out. This keeps the circuitry from all the cameras separated. I have used this Splitter for years now and it has never failed me.  It is a smart splitter using an Arduino for the brains. This means I can program each channel to behave however I want. If i want it to shoot, wait, then shoot a 2nd time, or hold a longer signal, or skip every other shot, or use a delay I have plenty of flexibility.

Until Thanksgiving, the splitter was triggered by the shutter control from the DMC-16. Then, on thanksgiving it stopped triggering the cameras. The USB connection still caused the Canon 6D on Otto to trigger, but all the other cameras stopped.

I have been fairly exhaustive in looking through the manual for anything I can find that would cause it not to work. I followed the instructions for making sure it was enabled, yet nothing I do can get it to work. I would suspect it has a faulty optoisolator on it, but if I reset the DMC-16 it causes all the cameras to trigger.  Also, I had tried reloading the firmware and that caused it to trigger all the cameras quite a few times. This tells me hardware = ok. Something in the software is preventing it from working.  I have spent quite a bit of time designing/building camera control systems, building camera cables, building smart splitters. I know exactly how these systems work, but i cant for the life of me get the DMC-16 to trigger.

I ended up rigging up an alternate fix. I have some Yongnuo wireless flash triggers. I put the trigger on the Canon, and the receiver plugged into the splitter. Now when the Canon opens its shutter the Trigger signals the reciever and that signals the splitter, which gets all the cameras working.

This is fine for a workaround, but I want to know why it is behaving the way it is. Luckily Dragonframe has a pretty slick trouble reporting system built in, and it grabs all the data from your shoot, screenshots of your setup, all the XML information and puts it in a package. So if I have messed up somehow they should be able to identify where/how. TO be honest I dont think it was something I had managed to cause. We will see, they have been good about responding.

2) Internal trigger.

Ok, some more bullshit to deal with. The DMC sometimes does not see the trigger from the BCM. The BCM (biolapse control module) is the timing system I have been using the last few years. It controls the temp, humidity, watering cycles, day/night cyclesfor the grow lights, and will even turn off the grow lights off and the fill lights on when it is ready to take an image, and trigger the cameras.  It has been running for years flawlessly.

I had decided to keep using it as it will make sure the humidifier is disabled 3 minutes before it shoots. This prevents any fog from the humidifier from being present when it shoots.  It also has dual power inputs, one for the logic and the fill lights which would be plugged into an UPS in event of power outage the fill lights would STILL work. The Grow lights and humidifier are on another power rail which would NOT be plugged into the UPS, so if a power outage happens the grow lights and humidifier just shut off rather than sucking power out of the battery backup.

The DMC seems to have issues of reliability with an external trigger. It works great for the most part but i keep having situations where it no longer sees the trigger from the BCM. I dont think it is the BCM that is the problem, because when it happens I dont reset the BCM, i reset the DMC-16 and it starts working again.

I tried running the BCM at a 5 minute interval yesterday, and put Dragonframe into a 5 minute timelapse interval and let them run free. However something happened and they drifted, and after about 50 frames Otto started snapping shots while the grow lights were still on.

Today I pulled the BCM out, and tied the growlights in as a BASH light using a DMX dimmer pack i purchased.

The Bash light is a work light that would normally come on between images so the artists could move the puppets between frames and have good lighting. Then it shuts the BASH light off before it takes an image, and turns it back on afterward. By setting a day/night timer between the dimmer pack and the grow lights, i can still maintain Day/Night sequences and have the grow lights shut off when it takes the images.

This is not ideal. The BCM kept the power rails separate. Now if a power outage hits, the 30min of battery life I maintain during that power outage is going to be reduced to about 25 minutes. I would rather have the full thirty.

This may just be what I have to live with though, not a HUGE deal, but I would much prefer controlling the lighting sequences through the BCM as i can program in the Pre-shot delay on disabling the lights and the Post shot delay on turning them back on. Plus  i have to be more mindful of the humidifier

3) Limit Switches

They dont work for shit to be honest.  The system has i think 8 sets of limit switches. You dont need them for the most part, but I had planned on limit switches on the X Y and Z axis. Here is what I am running into

Y axis, one the far end when the limit switch is triggered (just a whisker switch at the end, when the gantry gets too close to the back it tuns into the whisker and triggers the switch) it stops the rig as expect, albeit not fast enough, it just slows to a stop. In my view it should just stop immediately like my chronos rails would. So this means you have to have extra long switch and make sure it will still stop before reaching the end. this is a design flaw in my view.
Once that switch is hit, it slows to a stop and will not move in that direction anymore. However, I CAN back it out the opposite direction. Makes sense right? That is how it SHOULD work.

When the Y axis nears the closer end, it hits the switch and slows to a stop. I can no longer move it in that direction. Makes sense. However I cant move it in the other direction either! So i have to shut off the goddamn motor (thank god I had the brains to add disable switches on the power box) and twist the ball screw by hand to back it out enough in order to disengage that limit switch. WTF?

I had wired in a single limit switch for the Z axis to prevent the system from raising too high, there is a serious possibility of something being damaged if it starts elevating the Z axis and does not stop.  After all it uses a 425oz in stepper with a 30:1 gear ratio. I cant get that limit switch to work at all.

I spent so much time working on the Y switch that i never got around to wiring the X switches. And the Z does not work period. And yes, I did read the instructions, all non used switches are pinned into the common for each set. All switches are NC. This is not rocket science.


The DMC-16 is a powerful piece of hardware that comes at an expensive tag. So far I have been able to work around most of the issues I have run into, but I would rather NOT spend time troubleshooting and creating workarounds for some of its failures. I want to be FILMING.
Am I glad I got it? Hell yes. Even though I spent some time harping on it, those are sort of nit-picks.  Even with these minor issues I have run into it has giving me some remarkable capability for botanical timelapse. I really hope the dragonframe support team can help me work through the bugs. The Input trigger, shutter trigger, and limit switch issues are the only thing tarnishing what is otherwise one bad ass piece of hardware.

If anybody else has used the DMC-16, I would LOVE to hear your experiences and if you have run into any of these troubles. I do recognize the problems may be that I am just doing things wrong, or messed something up, but I have a pretty good understanding of how input/output triggers work.

If the issues I have run into are user error, I will be sure to provide an update advising I am just an idiot.  If Dragonframe comes through and gets this stuff working, I will also let you all know as well. I may seem like I am being hard on this, mainly because these minor issues irritate the shit out of me. I have failed shoot after failed shoot on these stupid lillies, and spent a lot of time trying to debug the issues and creating workarounds rather than shooting footage. But in the end this really is one badass piece of gear, combined with the excellent dragonframe software, this is going to let me do some incredible stuff.

I have only talked to one other person who has used the DMC, and he also had a hard time with the limit switches, I think in the end he abandoned them as he did not see them as overly important.

UPDATE 12/5/2016

So as of today, with much time spend with Dragonframe support team the external shutter still does not work.  Thier recommendation is to use the Relay output to trigger the splitter instead of the shutter. I got it wired up, and it works like a champ. I would like to have the shutter output working properly, but not at the expense of shipping the unit back to dragonframe, waiting on repair and shipping, delaying my shooting.  Right now it does everything I need so i consider that matter resolved.

The limit switches are buggered up, but they seem to know about this now and said they have a firmware update they are working on. I am looking forward to getting that resolved.

The issue where things would lock up and become unresponsive is directly tied into the shutter issue.  It was a bit tricky to pinpoint the events needed to cause it. If DF is set to use an external shutter trigger (which mine is), the USB connection still preps the camera for the shutter but does not trigger it. If the external trigger does not trigger the camera, it goes into a limbo state, and if you start monkeying around while it is in that state you can totally fubar it up. So if the external trigger works as expected, there are no issues.  Otherwise you can refresh the camera connection which also restores everything. Something that was tripping me up, is TEST SHOT always uses the internal shutter via USB. Add that in the mix and things get a little bit strange, you really have to start paying attention and making notes to bring these issues to light.

So right now, everything is running just fine.

I have Otto the 8 DOF rig running one scene, and a 5 DOF rig running on the 2nd stage with Pan/Tilt/Focus/Linear/Turntable.



Shooting with Otto

Quiet on the set!


I had no illusions that building a system like this would be easy. Nor did I have any illusions that using a system like this would be easy either.

For the past week I have been trying to shoot with Otto and I have had some mixed success. Rather than spending weeks at a time, I have been trying to test it out with fast growing plants and flowers.

The first test went fairly well. From a robotics standpoint everything worked like a charm, but I failed to properly predict the behavior of the plant I was shooting,  Until I get into the hang of things, I plan to keep an extra camera trained on Otto so in case something goes wrong I have some visual data of what is going on.

Its kind of cool to watch I think. What you see above is 5 days of filming. Otto’s motion is fantastic, I could not be happier!

However I misjudged the plant. You will see below.

I still consider this a success, besides the fact that the timelape is nothing exciting to watch. I misjudged the speed that the plant would grow, to be honest im not even sure what it is. I have several packs of seeds of various herbs and not all of them are labeled.  I expected it to get larger in the end which is why the camera pulled out.

If you look at the 9 second mark on either video, you will see the camera jerks down just a tiny bit.  Otto’s routine was set to tilt up and it was going a bit too far.  Between frame intervals on the BCM i had stopped the program long enough to adjust the amount of tilt to reduce it because it was going to put the plant outside the frame. This sort of worked, the plant never leaves the frame but I should have immediately applied a keyframe at its current spot, then adjusted it. Since i did not the last keyframe was in a spot further back, meaning the new position of the camera was tilted up slightly more than the last frame, resulting in that little pop.

In general i was fairly pleased and setup for a new shoot using some lillies. I got everything setup, I have 2 shooting stations going, on the left is the main stage, and on the right is stage 2. There might be a stage 3 behind the main stage soon as well.


Above: Main on left, 2nd stage with the herb on the right


Above, Stage 2, using an eMotimo as a makeshift turntable. This will be a 4 axis setup with linear, pan, tilt, and focus when everything is finished. I am waiting on some connectors, in the meantime a simple rotation always looks great on camera


Above: Otto and the main stage with the Lillies

Thats where everything started to go to shit.

First the camera shut off on me.

Not sure why that happened, I suspect it is the power supply, if it happens again Ill buy one from canon for way too much money.

Started it up again,

Things were going along splendidly, but at one point after a nice thanksgiving meal at my parents house, I was worried I had misjudged the motion of a plant and felt i should try to correct. First things first, I applied keyframes at the current spot to ensure the camera could move back to the exact same location. Everything was working smoothly until Dragonframe stopped seeing the triggers from the BCM. (BCM = biolapse control module, this is the controller I use to coordinate the lighting for shooting, growing, day/night, and control heat/humidity)

After playing around with it I managed to completely botch up the shoot. No big deal, that is what this phase is all about.

Started it up again. This time i noticed that the Canon on Otto was being triggered by USB, and none of the other cameras were triggering from the external shutter control. Even today they still dont work, I had to rewire all the triggers to get around the problem.

And around that point I think I sort of drowned the lillies, the flowers stopped opening.

In the end though, I did not get anything particularly useful. I do have some quick clips though that give an interesting demonstration of the movements that I can get. They are pretty dynamic. Each routine was going to be about 15-20 seconds, and I keep running into issues.

This video is the results of the failed experiments, but with each failure I learned something new. So I’ll chalk it up as a success. I am shooting again right now with some new lillies and feel optimistic.


Project Otto session 9

Otto Works!

Some people have asked about “Otto”, its a nod to the Autopilot (Auto / Otto) on Wall-E, as it drops down from the ceiling and has a camera in the middle.  🙂

Other than running into a few minor issues with the camera power supply, everything has been going together quite well. Otto is about 98% complete and already fully functional for testing purposes.

I am still waiting on the heatsinks so I am limited to running the system about an hour at a time then shutting it off to cool down. Once the heatsinks get in that should clear that up. In the meantime I can just knock the power down a bit to the motors and that should keep them under 80c.

Here is my list of stuff left to do.

  1. Adjust the set screws on the geared motors
  2. Adjust for backlash with dragonframe
  3. Add heat sinks for the motors
  4. Replace the crappy motor on the 3d slider
  5. Run some electrical outlets along the top of the framework for lights and such.
  6. Attach the supports beneath the cable tracks
  7. Finish the limit switches
  8. Wire up the Emergency stop
  9. Fabricate the final brackets for the X axis.

I should be able to get most of this done tomorrow morning, except the heat sinks which I am still waiting for.  The new stepper should arrive today. So tomorrow all 8 degrees of movement should be functioning

Stuff I am considering…

  1. Add a second ballscrew and motor for the Y axis. I am not sure this is really needed yet though. It works fairly well as it is
  2. Getting that stupid spool machined out. The temp one seems to be working just fine though

So lets get past all this build talk nonsense and get into some testing!

This is the first test, and I am pretty excited about it.  The first couple seconds are what excite me the most. I moved in real close and nailed the focus dead on. Keep in mind, 60mm, F/16 that depth of field at close focus is less than a quarter of an inch. I absolutely LOVE how the footage looks too.

As the camera moves left up the leaf to the stalk then to the other leave it has a very gritty “Stop motion” feel to it. I need to play with the “shoot while move” capability of Dragonframe.

Towards the end of the video the last pitcher spends some time out of focus, I was going to try to correct this… Maybe i should explain a bit.

Dragonframe has some incredibly awesome features. Right now my rig is not moving too well with live speed. it CAN move at those speeds, but it starts making some bucking movements that reach the camera at some speeds. I have not had a chance to delve too deep into it, I assume it is due to the X axis brackets are still the temporary ones and I have not replaced those with thick 1/2 inch aluminum brackets. Also this may be the reason i add another ball screw on the Y axis.  At first I was a bit worried about this, because with live speeds you can preview the movement in real time without having to rap off 600 images off your camera.

For those who are not really that into this hobby who may for some reason be reading, shutter count on a camera is like the odometer off the car.  The ability to live preview the motion allows you to see the moves and focus targets without having to run it through the sequence and rap off a bunch of shutter actuations.  Without live view I was worried i would be putting 3-4x the number of shutter clicks as I proof moves before I start a 3 week timelapse.

Luckily Dragonframe also has the ability to take snapshots off the live view. So the shutter stays open, dragonframe will move each frame, pause for a moment then snag a image off the live view. Not only that, but you can scrub through the results and play it back instantly. The video above was gathered in about 15 minutes using this method. So only one shutter actuation was used to do the whole thing. This lets you preview the motion, then you can go in and move the camera to any spot in the timeline that concerns you and you can adjust the motion accordingly.  So while the live speeds dont work, I can have a full preview in about 10 minutes, and it looks fantastic! It is not full HD, but close to 720p I believe. Once the motion is exactly the way I want it, i can run the actual timelapse.

So getting back to the focus, after reviewing this footage I was going to correct the focus at the end, rerun the move, and throw this footage out. However, as I am still learning I ended up sending Otto direct back to the start position on accident and knocked the plant over, so that was that. It would have been very difficult to get the plant back in its original spot so I moved on.

I hate using this term, because I think it is thrown around way too much, but the fact is Dragonframe IS a GAME CHANGER for the way I do timelapse.  Combine it with a rig like this, and the possibilities are amazing. I feel completely recharged and inspired just off this first test.

When running this, i put my Panasonic GH4 on timelapse mode and filmed the move.

What I really like about this is the feel of ease and weightlessness the camera rig demonstrates.  Considering there is 40lbs of gear being moved around, this is impressive. You can see the focus ring shifting back and forth quite a bit 🙂

I wished i had managed to get some video of dragonframe itself and some of the features. There are so many more techniques to learn, so many more options and features. I played around with the DMX lighting a bit as well and it worked flawlessly. This will mean i can change and adjust light sources and color throughout a multi week timelapse.

I can also edit and change the move on the fly. If i am 1 week into filming and see something awesome about to happen, i can change the rest of the motion sequence on the fly to take the camera in a better direction to capture something of interest.

This stuff is amazing.

I am going to try to get as much stuff done as possible this weekend. Once the heat issue is taken care of it will be time to start filming non-stop.

This weekend I will do some more testing, I will try to put together a VLOG that shows more of this amazing software and its flexibility.