Pi Zero Drone: Part 9

Quad Post #9: Calibrating ESC Manually
Each of the ESCs (Electronic Speed Controllers) on the quadcopter has to be calibrated to the RC transmitter signal so that the full range can be read on the quadcopter. To start the calibration procedure first ensure the following steps are done prior to ESC calibration
  •     Bind RC transmitter to receiver and calibrate the transmitter using the steps from blog post #8.
  •       Disconnect ALL propellers from the quadcopter.
  •       Disconnect any external power source i.e. USB cables from computer and Li-Po battery.

    This method of ESC calibration is referred as the “All At Once” method. Here are the steps:
  1.       Turn transmitter on, and set the throttle to maximum.
  2.       Plug in Li-Po battery.
  3.      Wait for the 3 signal LEDs on the PXFmini to start light up in sequence
  4.      Once the LEDs illuminate, disconnect the battery. (Keep throttle at maximum!)     
                     The ESCs are now in calibration mode


            5 .       Reconnect the battery and wait for LEDs to light up in sequence and buzzer will sound.
            6.   Once buzzer and LEDs light up in sequence, move the throttle to minimum
            7.   You now have full control of the motors with new ESC calibration. 

Pi Zero Drone: Part 8

Pi Zero Drone Post 8: On board Instrument Calibration
From the on screen response to tuning PID controllers, APM Planner 2.0 software can give a newbie quad user (myself) a deep understanding of what it takes to control this quadcopter. The software is capable of display real time response of the IMU, RC controller and other instruments you wish to fly with. This post aims to give a brief introduction and how to calibrate your instruments while I wait for new ESC to come in from Amazon.
The APM Planner 2.0 has a very interactive GUI that can be used for multiple vehicle types. A full in depth tutorial of the software can be found on ardupilot.org/planner2/. First step is to download the software onto your computer and follow the instructions on the screen. To connect to the quadcopter, first plug in the battery and wait for the buzzer to sound. This time you can connect to your quadcopter using the 5 GHz Wi-Fi signal.

Steps to configure the Wi-Fi signal can be found in blog post # 6.

After you have connected, the on screen altitude indicator (Very similar to one on an airplane) should start responding to the onboard IMU. Now the calibration can begin. In the APM Mission 2.0 tool bar select ‘Initial Setup’ >> Mandatory Setup. From here you can pick the frame type, this drone uses the first X-style frame type.



Compass
Select Compass to start the calibration. The Compass will need a live calibration by rotation the quadcopter in the air slowly around all axis (front, back, left, right, top and bottom). This will be done for 1 minute. Calibration offset will be created and stored for future missions.  

Accelerometer
Selected Accel Calibration and the calibration wizard window will appear. Be sure to have the quadcopter on a flat surface and select Calibrate Accelerometer and follow the prompts on the screen. It is important not to move the copter immediately after pressing continue for each process, this may lead to skewed stored parameters.

Radio Calibration
Remove all propellers and bind the RC transmitter to the receiver.
Select Radio Calibration and start the calibration by pressing the Calibrate button. Move the sticks and switches to the maximum positions. When you feel that all the movements are correct, press End Calibration and the offsets will be stored for that radio controller.
Attach “RC_Calibration.jpg” here
Flight Modes
This setting can be used to assign switches on your transmitter to modes like loiter or stabilize. You can set your own settings based on what you wish your quadcopter to do.



Power Wheels Racer Post-Race Update

So I did not really keep up with posting about all that has happened with the car lately but now that we are back and unpacked from Maker Faire I am going to update everyone on what we did. As far as the race goes in Orlando we had a great time up there and all the cars everyone brought were awesome. Here is a short video of the race so that you guys can see all the cars!



Although our car had some speed we actually were not able to finish the race due to burning our speed controller. Basically what happened was the connector that was between the forward and reverse and the speed controller melted from the motor running as long as it had to during the race. As that plug melted the two copper lead got closer and closer together until the shorted the wires running out on the speed controller this short then caused the speed controller to burn up. Here are some pictures of the damage. 





So after we really didn't think we had any chances of racing again. We had no replacement speed controller and after calling a few places in the Orlando, the internet was really the only place that we were going to find a 36V 1000W motor controller, but we didn't have time to wait for shipping. Then after talking to some of the other racers the guys over at Familab told us they may have a spare speed controller at their makerspace and they would let us borrow it if we came by to check out the lab. We took them up on the offer. They have a really cool setup over there with a lounge and hangout area plus some really awesome machines in the back. They had a manual mill and lathe, a CNC mill, one really big 3D printer, and lots more. I would highly recommend giving their lab a look if you are in the Orlando area.  And they let us snag a Kelly Controller 24V 400A speed controller to try to rig up on Sunday morning. We found a wiring diagram from Kelly controllers website and tried to follow it as best we could. We actually had to drive back to their space to grab a few extra parts. There was a main contactor, a few diodes, and a precharge resistor. After we had all we needed we still weren't quite able to get it back up and running that weekend.

Since Maker Faire, we ordered a new speed controller and decided to add a voltmeter and ammeter to the car so the diver might have some idea of what is going on back there. While I was driving I really had no idea that something was wrong until the motor locked up from the leads touching. Here is a link to the one we got below.

https://www.amazon.com/MICTUNING-6-5-100V-Digital-Voltmeter-Multimeter/dp/B01JOUZELG/ref=pd_lpo_469_tr_t_2?ie=UTF8&psc=1&refRID=V1PWGVGCC9264MT29MXX

That shunt that comes with it is basically just a really tight tolerance, low resistance resistor so that the meter can measure the voltage drop across it to get the current. If you want to know more about that resistor look up manganin resistors, they were actually used in determining the standard for the ohm back in the early 1900's. Also, Bryan made an Arduino circuit with a few temperature sensors so we can display the heat at a few different places in the circuit on the dash too. With the temperature, voltage, and current readings we are going to hopefully be able to diagnose any problems the vehicle has in the future faster and catch them before a controller looks like it got left in the oven too long (see picture above).

Power Wheels Racer: Car Updates Before Maker Faire

So as I talked about in the last post there are a few things that needed to be done to the car before it is ready to race up in Orlando. We read over the power wheels racing rules and found some things that we needed to comply with. First thing was the bumpers. We needed bumpers that covered 75% of the vehicles front and rear and covered a 1" space between 4-6" off the ground. So we ordered some aluminium tubing from Mcmaster Carr that already had holes in it and they even gave us the option to buy hardware to join them. After the tubing came in I took the stuff over to FAU and cut it to size on a circular saw and assembled it.
Front Bumper

Rear Bumper



Then we wanted to make sure that it wasn't going to tear up other people's cars while it was on the race track to we opted to just buy some pool noodles and tie then on the bumpers. In order to keep the colors right though we ended up painting them at maker faire. We also added some stickers to the car to make it fit our theme of hitchhikers guide to the galaxy, with the number 42, since it is the answer to everything.



The next update was the steering. We had a pvc steering wheel on there but there were a couple of things about it that weren't as good as they could be. For one the pvc is not real strong and could break. Also the hole through the steering column is not the same as any size screw you can buy at 0.3340" so no matter what bolt went through it there was some play in it. And finally, the steering wheel design made it tough to sit in the car because it took up a lot of space. Really the only way to get your feet inside of that thing is to pull your knees to your chest and if the steering wheel is there there would be no room for your knees. We decided on just making a joystick for the steering. We bought some aluminum rod, a smaller stainless steel rod, and a 10-24 Wingnut then I took the stock to the lathe. Here is what we were left with.

So the small stainless part was machined so that it would barely fit through the hole with about 20 thousands of an inch clearance. Then I 3D printed some couplings that would fit tightly around the shaft of the car and would house the aluminium rod and the wingnut. After we printed those the joystick was assembled and we wrapped it with some bike tape for grip. But when we put the throttle on there the aluminum rod was a little too big and tore up the hall effect sensor in the throttle so we had to order a spare and print a coupling from the shaft to the throttle. This kinda worked out for the better because the longer the joystick the easier it should be to turn. 



In the picture above you can see the forward neutral and reverse switch that Bryan added. He found one that used to be on a golf cart for pretty cheap and we wired it up so that the motor leads from the speed controller ran into it and then to the motor. It basically works as a mechanical switch so that the voltage applied to the motor is either forward or reverse or the leads aren't touching.

The last two things that we added were the kill switch and the fog machine. We 3D printed a coupling to go from the kill switch we bought so that it would be secure on the back of the car. The way it works is when the red key is in there is connection and the vehicle will operate. When the key is removed the speed controller will not allow any voltage to the motor. The fog machine is a dry ice container under the 42 on the front of the car and has a cap where you can add hot water to get the smoke going. Then the pipes run out to the back where the smoke would come out. I do not have any video of the fog machine yet but I will get some to post soon.



Power Wheels Racer

So we are nearing race time for the power wheels racer and trying to get it race ready so that we can get a driver some practice time around the shop before we race it next week. It got a new paint job today and a motor cover but there are some more fixes to be done before it is ready to go.

Updates to be done to be race ready

  • Front and rear bumbers
  • Kill Switch
  • Paint Steering Wheel
  • Cut and mount extra PVC piece on steering wheel
  • Reverse
  • Obtain 1 1/2" wrench to tighten back wheels
  • Driver foot rest
  • Chain Guard

Pi Zero Drone Post 7: Modular Body

Quadcopter Post 7: Modular Design

To the reader who has a smaller print bed and wishes to print this part I have designed a modular frame that will hold all the same components as before. Each print will not exceed a 114mm x 114 mm

I have attached the STL files with this post 

Here is a link to the STL Files:


Nano Blade QX2 Repairs


We have two of the Nano Blade QX2 drones at the Boca bearings workshop and I have to say they are a really great option for someone who wants to get into FPV flying/racing. Horizon Hobby sells them for $159.99 and they are sold as a BNF (Bind and Fly package). Only thing about them being bind and fly is that they have to be controlled by a Spectrum transmitter or any other transmitter that has DSMX or DSM2 protocol. However it seems to be mostly Spectrum radios that have it. And we have been using the Skyzone and Fat Shark Teleporter headsets to fly the drones.

We actually only had one of them at first though. But per usual we ran it into something and the drone broke. But that is part of flying drones. I wish I had a picture right after the crash but basically what happened was the camera case on the front of the drone broke and the camera split in half. We began looking around online and we really could not find anywhere that had the replacement parts.

Here are a few of the websites we found:
https://www.hobbytown.com/blade-nano-qx2-fpv-replacement-parts-racing/c10178
https://www.amainhobbies.com/blade-nano-qx2-fpv-replacement-parts-racing/c10178

The only thing missing from the replacement parts list above was the camera and camera case. But we did manage to find that online too.

Camera Case:
http://www.fpv-direct.com/blade-camera-mount-inductrix-200.html?utm_source=google_shopping&gclid=COqG8qCZoc8CFYdZhgodrysPgA

Camera:
https://www.amazon.com/gp/product/B01EZLYT24/ref=oh_aui_detailpage_o04_s00?ie=UTF8&psc=1

But instead of ordering the camera case I figured it wouldn't be too much to draw something up in CAD to replace the case instead of ordering it. And I wanted to get more familiar with Onshape. Although I am more fluent in Solidworks from using it at school Onshape is a pretty awesome program because it is all cloud based. That means that you can pretty much use it on any computer you have as long as you have pretty good internet connection. I even have it on my phone, which totally blows my mind. Because you could have a $500 dollar laptop with a 2.0 Ghz processor 4GB of RAM and still be struggling to run Solidworks. And Onshape makes it pretty easy to collaborate with other people on the same project. Here are some screenshots of the parts I made in Onshape.


Now the real tricky part of repairing the cameras on these drones is that the wires connecting everything are super super tiny. And when I say that I mean 4 stands of braided wire tiny. Which means while you are soldering them you have to be real careful not to pull them out (which I did multiple times). Here are some pictures of the iron and microscope I used to solder it back together.
 Had to scrape the adhesive off to reveal the pad I was going to solder it to.







So after that was done I 3D printed the new camera holder out of ABS and decided to use acetone to "weld" the two pieces together. That is one of the cool parts of using ABS not only is it stronger than PLA but it dissolves in acetone so you can glue two pieces together by melting the two pieces a little bit and holding them together until the acetone dries which happens pretty fast. The bottle below has the acetone is it and I use the sponge on top to wipe it onto the plastic and melt it.



And here is a picture of both of the drones. Now that we have two of everything we will have to get some races going soon. Hopefully we can make that happen and I will have a video to upload.



Luke Ridley

Pi Zero Drone 6: After First Flight and Mods

Quadcopter Blog Post #6: After first flight and mods


After the failed test, instant modifications to the design must be made. The most importantly was to protect the hardware so that no more time and money is lost waiting for parts. I have added a 3-D printed ABS cover that will protect the PXFmini connections and also add some style to the quadcopter. The design of the cover can be seen in the picture below:


CAD MODEL OF ABS COVER


In addition to protecting the quadcopter brains, the controller of the quadcopter has not been as responsive. Some research lead to a possible solution, the APM software that I am connecting to via Wi-Fi was interfering with the signal of the RC controller. The theory is this the Wi-Fi of the PXFmini operates at its own frequency band and the same goes for the RC controller. If the PXFmini comes close to the same band the PXFmini signal will be distorted and must be changed. In the US, a certain frequency band must be used for operations set by the Federal Communication Commission (FCC). Without getting too much in depth, you must choose the channel bands if you are operating in 2.4 GHz (Channel 9) or 5 GHz (Channel 40). Check out the chart below:


CHARTS OF FREQUENCY BANDS (2.4GHz Bands & 5 GHz Bands image)  



To check if the PXFmini is operating in the right channel use this command:
WIFI HOTSPOT COMMAND

Check the channel # in the following file:
WIFI HOTSOT SIGNAL FILE
Make sure that the channel # is 40.
Now check the signal strength using this command:
WIFI SIGNAL CHECK

After it is changed, connect to APM and turn on the controller. The interference has now been fixed.

Pi Zero Drone: Test

Quad Post 5: Test

Check out the video to see how to arm the quadcopter.

Arming Video:  https://www.youtube.com/watch?v=9BmgpX3vKM0

Quad copter test 1….. FAIL

Crash Video: https://www.youtube.com/watch?v=GUAmSo_e-a0

Crash Picture


After this test the PXFmini GIPO pin connection sheared from the PCB board and became useless. But from the video evidence I suspect that the trim on the remote may have been to an extreme right and also a bad ESC calibration. I am waiting for a new PXFmini to come in and I will retry.

There have been some modifications of the design on the quad:

  • Shorter legs with Lipo Battery slot made from ABS plastic.
  • Roll cage that will save PXFmini from future crashes.

Fitbit Cheat-O-Matic 2: 3D Printed Mechanical Simplicity (9-8-16 update)




As a preview to what this thing is shaping up to be, here is the gear motor and the new cam design. That new cam is what the bearings will be riding on to make the fitness device move up and down to fake it out. :)

Inline image 1


The cam follower will snap through the inner hold in the bearing with something just like this:


Inline image 2

It may take a few prints to get the tolerances right, but that's the concept. In this version of the Cheat-O-Matic, the follower will have a light spring damping the vertical motion so it doesn't hop up and down as harshly as the one in the video on my site.

Still a lot of CAD to do. This thing has a toggle switch for power and a port for a 12V DC plug. Hopefully the prints will come out nice and within tolerance and I won't have too much nylon fiddling to do to get the prototype working.

Talk to ya later!


-Andy

Fitbit Cheat-O-Matic 2: 3D Printed Mechanical Simplicity (Intro)

 I've been getting more into 3D printing, lately. One of the many things I felt could benefit from 3D printing technology was the Fitbit Cheat-O-Matic (version 1) I wrote about a while ago. It's very funny (as in, I built it as a joke; see previous linked article) and it even got some press at BBC.com. The crazy thing is I've gotten a number of emails from visitors to my site asking if I could build them one. They all were willing to pay for it, as well. That's cool. I don't judge. ;)

The first design was kinda simple and never really built to run indefinitely; on the contrary, it was meant to last long enough for the office Fitbit competition at the time (about a week I think). It was enough to get me through a competition, anyway. Certainly not a design I'd wish to build multiples of and then sell. They would have all failed soon after their users started them up. Turns out RC servos, even the high-quality all-metal-gear ones. I killed three different kinds of servos, one of them being VERY expensive.

I was watching old engineering videos from 1939 on YouTube one day and saw a cool demonstration of a cam and follower in a machine of unknown purpose. I realized then that I could make a very simple contraption using 3D printing and a geared DC motor to shake a Fitbit up and down very easily. It could be super-simple to assemble, cheap to print all the parts but the motor and probably run off a simple 12V wall wart power supply.

This is the CAD design of the new Fitbit Cheat-O-Matic, done in Onshape cloud-based CAD:



Fitbit Cheat-O-Matic 2 3D CAD Rendering in Onshape

The yellow piece is the cam. The dark grey piece with the cross thing on it is the follower and also the holder for the Fitbit. The light blue shaft sticking through the cam is the shaft from the geared down part of the 12V DC motor. The big round cylinder in the back is the motor. The lighter gray and dark blue stuff is the base. The dark blue piece is the part of the base the gear motor is attached to using two screws. The base fits onto the bottom of the faceplate with a simple tab and the two sets of side supports. Here is the exploded view of the CAD rendering:



Exploded view of Fitbit Cheat-O-Matic 2

I 3D printed the parts at Xometry (http://xometry.com). The tolerances on this first attempt were a little tight, so I'm changing up the CAD a bit to save me some plastic shaving and trimming. That's mostly because I'm still kinda new to 3D printing. The total cost for the four parts and shipping was about $50. If I got around to buying a 3D printer, that would be much cheaper, of course.

The 12V DC gear motor came from Robot Shop (http://www.robotshop.com) and cost about $25 with shipping. The gear motor is design to run at about 58 RPM. Since the cam has two high sides, it pops the Fitbit up and down about twice a second which is a pretty brisk walk, if not a run.

The screws I already had on hand. The power comes from a cheap 12V wall wart power adapter from Radio Shack. Anything that can provide about 60 mA of power will suffice. That's the draw I measured while it was running with a Fitbit attached. 1 amp adapters are pretty easy to find.

Here's the Fitbit Cheat-O-Matic 2 in action:



Originally, I'd designed a gap between the cam and the holder for a 1/8" steel ball, thinking that would allow smoother operation. Turns out that caused a tiny bit of binding which made the device louder and made the motor work harder. Without the bearing, the plastic-on-plastic is quite smooth and doesn't appear to be wearing at all.

As far as efficiency, as I write this it is 8:38 PM. The Fitbit app says the machine has racked up 116,000 steps since midnight. I think it's on its way to about 135,000 for a 24-hour period. I was told there'd be no math.



Screenshot from iPhone for Cheat-O-Matic stats for today

I will compete separately from the Fitbit Cheat-O-Matic 2 in future office competitions. I'm a healthy person with a healthy sense of humor, so don't go off on a "that guy's lazy" rant. I am lazy in that I prefer to find faster ways to do things, but I'm certainly not a slug or a lump. Let's keep that straight. ;)

I like making things even if they have no real useful purpose. THIS device certainly qualifies. If you're interested in purchasing one, drop me a message. I'll work on getting them up on my store on this site. Haven't worked out a price since everything has been purchased at one-off pricing.

Andy is using Boca Bearings and has agreed to allow us to present his latest project as part of our workshop. Andy's original post can be found here

Power Wheels Racer Part 12: Updates and fixes

Part 12:

After the old brake handle broke off and the weld attaching the steering column came undone, we had to makes some updates to the car. First we decided to design and print a new brake handle. The two main properties we wanted the new one to have was an ergonomic grip and durability. The grip has four slots to fit the drivers fingers for a comfortable grip. To achieve the durability desired the grip was printed out of ABS instead of PLA and was given a 50% infill on the print. Once I put the handle I pulled back on it as hard as I possibly could and it didn't budge. I was even able to lift the front of the car off the ground by pulling back on the handle. Here is a couple images of the handle on the hydraulic break.
                            
However as I designed it the first time the handle did have a couple of flaws. The front end needed to be have a chamfer to clear the hydraulic break when it was resting. Also a small slot needed to be cut out of the inside also. The spacer that goes on the inside of the handle also needed to be shaved down a bit because I forgot to add tolerances to the width of it. I was able to do these with a dremel and angle grinder but since have went back and edited the solidworks file so that the next print will have all the features.

                           
The other thing that is different about this handle is the method in which it fastens to the break. I used 1/4 - 20 low profile screws and heat inserted nuts. The heat inserted nuts can be found on Mcmaster Carr and are inserted into pre-made holes by using a soldering iron. the plastic melts and hardens around the insert so it will stay in place. This is really nice for 3D printed parts because the nut is permanently fixed inside the part. The soldering iron wasn't quite long enough to insert the first heat insert through the hole and I melted a little bit of plastic around the outside. Next time I will have to use a longer skinnier iron. The inserts are pictured below.

                                       
Another change that was made was the throttle set up. Instead of twisting your hand around the handle now the throttle is activated by the driver's thumb. This allows the driver steer a little more comfortably while engaging the throttle. Bryan is also going to use two more 90 degree PVC joints and another shaft to make the steering wheel a rectangle rather than a U shape soon.

In order to fix the weld that broke on the steering column. We are going to fabricate a small 90 degree aluminium bracket to hold in place I also had to angle grind the frame so that it sits flat.