Falcon NoseCam Build Thread

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Finicky

Well-Known Member
TRF Supporter
Joined
Nov 2, 2017
Messages
157
Reaction score
377
Location
Upstate NY
A few people have asked for a build thread describing the NoseCam system I have been developing over the last few years, so here it is ! I’ll be describing the most recent NoseCam I have built, which also has deployable winglets for vertical orientation control (VOC). I have built six different versions so far, fitting body tube diameters from 2.6” to 4”. The 3” body tube version I will describe here is the most advanced design.

For those not familiar with NoseCam and the VOC system, you can refer to this thread:

https://www.rocketryforum.com/threads/a-new-spin-on-vertical-orientation-control.174525/
This build thread is mostly to show what is “under the hood” w.r.t hardware, and will not go into too much detail about the software or control approach algorithms. Hopefully within the next year I’ll post all the code, etc. on GitHub and then adventurous folks can try building their own. This is by far the most complicated rocket gizmo I have ever built, and I can say that building one is not for the faint of heart !

Here’s what the overall system looks like. It’s a nosecone with a LOT of stuff inside. A nice feature is that it can be moved from rocket to rocket. I have two different 3” body tube rockets that I swap this particular nose between.

1661713575447.png


1661713610935.png 1661713622381.png

1661713639012.png




 
Last edited:
For describing the hardware, I will split the NoseCam into two sections – the lower “Power Head” section and the upper “Rotating Nose Shell” section. The power head contains most of the electronics, the batteries and the main rotary drive motor. It is built around a FW 3” coupler tube (Wildman), and is intended to be fairly rugged so that it can handle rough landings.

The rotating nose shell is entirely 3D printed parts, and contains the camera (RunCam Split 3 Lite) and the winglet servos. I wouldn’t describe the nose shell as delicate, but it’s nothing like a commercial FW nose. The top of the nose (shown in red) is removeable so that it can be easily replaced if it gets crunched during a hard landing. So far the nose has survived landing just fine, but if it gets to be an issue I can reinforce the tip.

1661713832731.png
The guts of the power head are shown below.

1661713858686.png

I’ll provide more detail about the individual components as I describe the assembly steps. More to come …………………………
 
I’ll start by describing the construction of the power head. The photos below show the different sides of the coupler flange (left) and the motor mount (right). These are 3D printed. The black material is carbon fiber filled nylon, and the blue material is PETG. The CF/nylon is very strong, but prints a bit rough. In the case of the motor mount, I glued together parts made from both materials to get the combination of strength and surface finish I wanted.

1661810453457.png

1661810489315.png

Next I bond 4-40 nuts to the back side of the coupler flange.

1661810565419.png

Then the flange gets bonded into the coupler tube.

1661810669429.png

It’s very important that the centerline of the coupler tube assembly is well aligned to the centerline of the motor mount and motor, particularly w.r.t. angle. Even a small angle will cause a wobble in the flight video when the rocket body spins.

One of the things I like about the CF/nylon is that it is much more machinable (not great, but OK) compared to most 3D printed plastics. After the flange mount is bonded to the coupler tube, I face the end on a mini-lathe to square-up that surface. I made a mandrel of sorts to hold the assembly on the lathe. Anyone who owns a mini-lathe always wishes they had a bigger one LOL.

1661810715431.png

A trick that I use to improve the dimensional accuracy of some 3D printed parts is to coat a surface with JB weld and then machine most of it off. Below I’ve coated the backside of the motor mount with JB weld. This is the area that will eventually get mated to the coupler flange face.

1661810744236.png

The photo below is a bit confusing because I’ve got the motor mount piece temporarily mounted upside down on the coupler tube assembly so I can face off the JB weld area.

1661810773589.png

I then flip the motor mount to its correct orientation for test fitting to the coupler tube assembly.

1661810802403.png 1661810820629.png
 
The motor is an iPower GM4108H-120T brushless gimbal motor.

A quick historical note – The first several NoseCams used brushed DC gearmotors for the main drive motor. These were simple to control and readily available it many sizes and gear ratios, but they did not offer smooth rotary motion at very low RPM’s. The backlash in the gearboxes was also not ideal for getting smooth video. BLDC gimbal motors were much more challenging to wrap a position servo loop around (for me anyway), but they offer better rotary control with zero bachlash. And the hollow center of the motor allows for easier pass-thru of wiring to the top of the nose.

I bond a piece of thin-walled 5/16” OD stainless tube to the inside of the upper part (rotating part) of the BLDC motor. This acts as a support shaft for the rotary encoder and a pass-thru conduit for wiring coming off of the electrical slip ring.

1661811023749.png 1661811042058.png

The motor then gets attached to the motor mount.

1661811143689.png 1661811162338.png

The rotary encoder is then attached to the shaft. The encoder is a CUI Devices AMT10E3-V, configured for a resolution of 2560 pulses per rev [PPR]. That translates into 4*2560 = 10,240 counts per rev [CPR], or a rotational resolution of 0.0352 degrees. Probably a bit of overkill, but the microcontroller can handle the hardware interrupt burden OK from all those pulses, so why not ?

1661811194270.png

An additional cover piece is then attached which will later act as the mount for the electrical slip ring.

1661811231032.png 1661811266467.png

At the other end of the coupler tube, 3D printed parts are bonded inside that will act as the supports for the circuit board stack, and also the mounting points for the nose bulkhead fitting. This design uses four 4-40 stainless anchors that will eventually hold the bulkhead and eyebolt in place.

1661811298886.png 1661811316294.png
 
To finish the motor assembly an electric slip ring is added. This allows electrical connections to be provided to the upper (spinning) part of the NoseCam from the lower (stationary) part. For this build, I provide ~8V power to the camera and servos, and also the PWM control lines to the two winglet servos. The slip ring I chose for it’s small size had six conductors. Since I only needed four conductors, I doubled up on the power lines.

1661979555036.png

This is what the motor / encoder / slip ring assembly looks like when everything is wired.

1661979585399.png 1661979602500.png
 
The electronics stack is composed of four boards using the Arduino Uno form factor. Initially I was hoping to use a form factor a bit smaller and more modern than the Uno’s, but the BLDC drive amplifier I ended up using was only available in this form factor, so that made the decision for me.

The four boards (starting from the bottom of the stack) are:

1) The microcontroller (MC) board – Adafruit Metro M4 Express
2) A custom PCB that I made for mounting sensor and Bluetooth modules
3) The BLDC drive amplifier – SimpleFOC V2.0.3
4) An empty protoboard for adding unexpected stuff

1661979711232.png

Next I’ll show a few photos from stuffing the custom PCB. (Sorry – I know the surface mount guys reading this are cringing right now at the thru-hole PCB) The primary 6DOF sensor is an Adafruit ISM330DHCX. The PCB has room for a second sensor module, which in this case is a 6DOF Adafruit ICM20649. Having a second module slot allows me to evaluate modules against one another. Both modules communicate over the SPI bus because it is significantly faster than the I2C bus. The Bluetooth module is good ol’ classic Bluetooth, which I still prefer to Bluetooth LE for its simplicity.

1661979773452.png

1661979792580.png

1661979815784.png

1661979835826.png

The third board on the stack is the BLDC driver, which is followed by the empty protoboard. For this build I added a buck voltage convertor to drop the ~25V motor supply from the batteries to ~8V to power the rest of the stack, the servos for the winglets, and the camera module. The FET on this board is used to remotely turn the power to the servos and camera ON and OFF via Bluetooth. This way I can wait to turn the camera on until the very last moment to save battery power. The camera is the biggest power hog of the entire system.

1661979881300.png

Once the stack is complete the various wire bundles for the motor, encoder, and slip ring can be completed.

1661979915909.png

Here’s what it looks like before test fitting everything.

1661979938042.png

Putting the stack into the coupler tube for the first time is always a bit of an adventure. This is definitely a case of fitting 10 pounds of s*!t into a 5 pound bag !

1661979957832.png
 
Final assembly of the power head involves adding the motor to nose shell coupler. This piece is mostly 3D printed CF/nylon for strength, with a PETG skirt bonded to the bottom. Similar to other parts of the rotary system, I used the mini-lathe to clean up the outside surfaces.

1662077460167.png 1662077474564.png

Here is the finished power head with a split band bonded to the coupler tube. As I mentioned at the start of this build thread, the power head is the foundational, rugged part of the NoseCam. I’ve used a single power head with several different upper nose sections when trying different things out.

1662077506633.png

As a bit of a side note, below is a photo of the bench setup I used while learning for to commutate the BLDC motor / encoder using the SimpleFOC software and drive amplifier. I spent many, many weeks learning how to correctly tune this system. For folks who might be thinking of building a NoseCam, being able to successfully build and operate just the motor system I’ve shown below is necessary before attempting to build a complete NoseCam.

1662077526189.png
 
At this point the complicated bits are done. Just a few more sections to go. Next is the nose camera module. It’s comprised of several 3D printed PETG parts, the RunCam 3 Split Lite, and a heat sink.

A critical feature about the RunCam 3 Split is that it starts a new recording as soon as power is applied. You don’t need to hit any buttons on the control board. This means that all the components can get buried inside the nose without concern for how to access them before launch. If anyone knows of other cameras that have this “instant video ON” feature I’d love to know about them.

The camera head itself is a bit of a weird shape, so I attached an adapter to make it easier to mount.

1662158030229.png

1662158046205.png


After I install the camera head into the nose shell I secure it with a strip of Gorilla tape. (The tape isn’t shown in the photos)

1662158091359.png 1662158116596.png


This is what the camera shell looks like on top of the power head.

1662158158550.png


Other 3D printed parts are bonded inside the camera shell to create a support platform for the camera electronics. A ½” diameter aluminum rod is used as a heat sink for the camera control board chip. These chips get really hot, and can shut down if they overheat. The heat unfortunately still gets trapped inside the nose since there is no “ventilation”, but the heat sink does prevent the chip itself from getting super cooked.

Other pass-thru wiring for the winglet servos also gets routed at this point.

1662158208026.png 1662158221361.png
 
The servos and winglets are the last section of the NoseCam to build. The body parts are 3D printed PETG. The winglets are cut from fiberglass sheet. A custom bracket is first used to hold the servos in place, and then the bracket is mounted into the nose section.

1662227775009.png


1662227786717.png 1662227797695.png


This is what the sections look like before final assembly.

1662227832610.png


Here are a few photos of the Falcon NoseCam mounted to a rocket.

1662227858167.png

1662227870692.png

1662227882890.png

1662227896301.png

That completes the construction overview. I hope to accomplish several more test flights this season. Hopefully this winter I will clean up the code and then post that on GitHub. Cheers !

Steve G
 
So it took a while, but I've finally posted the info on how to build a NoseCam onto GitHub!

https://github.com/sjgregorski/NoseCam
For those hardware junkies interested in NoseCam it's a bit of good news / bad(?) news. The good news is that it's finally posted. However, I "only" posted the info on how to build a NoseCam without VOC. If you download the distribution, you will see it's still quite a bit of info. I'm going to take a wait-and-see approach to adding the design info on the VOC module to GitHub. If and when folks start posting NoseCam videos of their own then I'll revisit the topic. As I state in the distribution READ ME files, what I posted on GitHub is 90% of a VOC system. Everything from the camera module down is exactly the same hardware.

P.S. - I have no idea if I posted things on GitHub "correctly". This is totally new to me. I think (hope) anyone should be able to download a single ZIP file with everything.
 
continues to be very cool, far above what I ever expect to achieve. The system of course adds some weight and drag. I am wondering however how much altitude is lost by the average high altitude junkies’ rocket due to almost all rockets have some degree oF off vertical tilt. Just shaving that off may significantly increase absolute altitude achieved. Still, just the improved quality of video alone is a huge advancement in the hobby.
 
continues to be very cool, far above what I ever expect to achieve. The system of course adds some weight and drag. I am wondering however how much altitude is lost by the average high altitude junkies’ rocket due to almost all rockets have some degree oF off vertical tilt. Just shaving that off may significantly increase absolute altitude achieved. Still, just the improved quality of video alone is a huge advancement in the hobby.
Thanks. I think a while ago I mentioned in one of the NoseCam threads that you will never need to add weight to the nose again :) The full VOC NoseCam weighs just over 2 pounds, which is pretty heavy compared to a "normal" 3 inch FWFG nose plus a video camera strapped somewhere on the rocket. I've worked on making the camera bump as small as possible to reduce drag. With the latest VOC system there should be no additional drag when the winglets are fully retracted. I'm very hopeful for the first test flights !
 
Anticipated max acceleration, speed, or temperature constraints?
I don't think I can go much over Mach 1.5 (total guess) with the current system because it has 3D printed plastic everywhere. I do have plans to remake the top section using a piece of FWFG nose as the starting point. That will require some interesting machining ..............

This season I will push everything more aggressively to see what happens. Last season I had one flight hit 42 G's (for less than a second) and the system was fine throughout. The challenge I have is that flying in upstate NY means if you go that fast you may very well end up in a tree :p
 
Will cold temperatures at extremely high altitudes be an issue?
I really don't know. Certainly any plastic that embrittles would need to go. Everything has thermal mass so it would take a few minutes for the internal components to cool off. I imagine the gyros might see a big change in null offset once they get cold, which will mess up the calculations. Another issue with extreme altitude is gyro integration error buildup. I think the system is reasonably accurate for the first ~60 seconds after launch, but after that I have no idea. My flights never take longer than 40 seconds to get to apogee. I started writing the code to use a 3-axis magnetometer to help error correct the gyros, but I found that using magnetometers was a big pain and wasn't making things better for the kind of flights I was doing.
 
Might be that be the time the cold sets in, the air is so thin the rocket is essentially ballistic anyway . If I read the Mesos thread right, the rocket was pointing just about anywhere it wanted for the terminal part of the upward flight.

Well, now that I think about it that could be a problem for breakage rather than flight. At altitude where fins stop working , the rocket is gonna wobble and your correcting winglets are gonna be flipping in and out like mad trying unsuccessfully to correct. So if anything jams from cold shrinkage, could be a problem.

I imagine however for a project of that magnitude they will have thought this through and built it out of heat and cold resistant materials. Just daydreaming.
 
Back
Top