Close
0%
0%

Choose your own adventure bot

Ultra low cost 3D printed Open Educational Resource Walking Robotic Platform

Public Chat
Similar projects worth following
We are starting a makerspace in Dayton and we will have make and take classes. There is also drive to teach real skills like Mechanical Cad, E-Cad and programming. Whether materials are part of the cost of a class or provided by the makerspace, the budget is going to be limited to
sub $10 per class. How can robotics be taught with this limited budget? By doing things differently!

1) Make a cheap motor a smart motor. Use cheap and familiar sensor modules with tutorials parts all over the web.
2) Maximize usage of inputs and outputs on powerful $5 ESP32 processor.
3) Through DFx! Design for assembly, reliability, expandability, and accessibility.

Classes will start with a build to get people intrigued then allow them to mod their creation with open source tools at home. With good documentation this may become an Open Educational Resource (OER) for the community at large.

Top 3 struggles I have with robots and how this project solved them.


1) Cost too much. Mostly because of expensive servos and high end processors.

-  Designed a cheap driver/sensor board to read the absolute angle of the motor joint and run the motor.   Designed a 3D printed connection that use the motors as structural members and use the 3D printed ramps to change the distance from the optical sensor with rotation. Used the worlds most ubiquitous yellow gear motor found in inexpensive projects everywhere. Using extremely common FFC cables to link the legs and the boards for one robot cost 15 dollars and can arrive on your doorstep in 8 days! The boards also have what would normally be the guts of a servo motor, but now are more accessible to the processor which can access the position of the joint at any time.

2) Processor creep. Robots eat IO pins for breakfast.

- Used a double demux  in which  the PCA9685 board is also controlling the analog demux. In this way I can run 6 motor joints and read their position with only two I2C pins and one analog input. This leaves ~20 other IO pins to do range finding, balance, audio out, I2S audio in and many other functions yet to be added. This can all be run with a $5 ESP32. Lastly, it is expandable to allow 3 additional pairs of arms, legs or wings, whatever you may fancy. This is achieved with just adding a daughter board  with the PCA9685 expander and the 74HC demux chip. 

3) Wires and batteries make a robot ugly and unstable.

- The last advancement is a unique setup for how to connect the motors and batteries. The legs are  connected to the brain board through a double row header so they can easily be exchanged. First hip board with the header has two FFC cables that  run down to the Femur and Tibula boards which are common and can connect to the battery. The cables are elegantly threaded though the joints with service loops to allow the legs the full range of motion without risk of the cables being snagged.  The batteries, one in each foot to lower the center of gravity, have a separate set of traces that link them in parallel through a BMS in the head board. This allows the motor to be run at 7.4-8.4V and be charged through a 9V adapter plugged into the headboard. 

Purpose 

Cya was designed to encourage makerspace members to learn about robotics and the tools used to make this Robot. Here are some reasons Cya serves this purpose well.

Designed with maximum availability in mind. 

All the electronic components and motors are available on Amazon, mostly with Amazon prime.

Board files are available and 10 bare headboards and 10 sets of populated leg boards can be ordered from JLCPCB for $150 and delivered in around a week. With some better availability on alternate motor drivers and sensors I expect to be able to cut this in half. 

There are eight 3D printed components that make up a leg and two legs can be printed on a$259 Sovol S01 printer in one print job.  The filament for the one print job costs less than a $2 and finished in less than a day. Another option might but using a 3D printer at one of the main libraries that now offer them. 

The rest of the parts are standard hardware like screws and headers.

Everything is attached with through hole soldering or connectors. The only tools needed are a soldering iron and a hex driver.

Everything can be run with Arduino to start with. Micro Python modules are being built out and there is a plan for native C in Doxygen using Expressif libraries. There are example sketches and tutorials for most of the modules in Arduino so we stand on the shoulders of giants and this project leverages  the open source efforts of the community. A special shout out to Adafruit and the PCA9685 board that Cya is designed around. https://www.adafruit.com/product/815 Their great documentation spurred the idea and got me going....

Read more »

x-zip-compressed - 369.32 kB - 10/02/2021 at 22:07

Download

Comma-Separated Values - 1.03 kB - 10/02/2021 at 13:00

Download

Comma-Separated Values - 1.48 kB - 10/02/2021 at 12:57

Download

x-zip-compressed - 24.09 kB - 10/02/2021 at 12:57

Download

Pinouts for each board.pdf

PDF that shows the connections between the boards and the value for each through hole pin header.

Adobe Portable Document Format - 665.70 kB - 10/02/2021 at 11:58

Preview
Download

View all 8 files

View all 13 components

  • Assembly

    shane.snipe3 days ago 0 comments

    Pins on top. Note Tibia and Femur locations. Silver traces down, blue spacer up.

    Solder wires to motor.

    Sandwich motor, sensor board and hip joint with 40 mm screws.

    Sandwich from the front.

    Temporarily attach with screws and solder on the wires.

    Sandwich motor and board again. Slight three handed operation.

    Back up one and attaché the FFC first.

    Finish motor sandwich.

    Put motor adapters in hip. Using one skinny and one fat.

    Put top of hip on and thread the FFC through the middle.

    Attach the motor wires to the Tibia board and the FFC. Screw it in the boot.

    Attach motor to Tibia sensor and put the top back on. Screw it down. I do not have the right FFC here so it is not routed yet.

    Boots on backward!

    Turned the boot around. Should be able to test the sensors as I wait for the longer FFCs?✌️

    From the back.

  • Hip Repair

    shane.snipe10/09/2021 at 11:15 0 comments

    In the last log, I detailed the issues with the Femur joint sensor. The face was not in the linear range and the last half of the travel was not linear. There was a step where the gap is between the parts and the last half was not sensitive at all.

    So I fixed three things in the model below.

    1) Added 1.5mm to the thickness and pushed the reflected surface out by 0.5mm.  This should get it more into the linear range and reduce any weird effects from the outside being too thin causing it to reflect a different color.

    2) Added a tongue and groove feature to prevent a big gap from showing and causing a discontinuity in the sensor output.

    3) Changed the hard stop location to increase the range of motion. This may allow Cya to be able to sit up. Presently limited to about 120 degrees total travel.

    Lastly I made them symmetric because they were starting to look ridiculous.

    I will get these printing and see how they work out. I will post some pictures were they are done. I made them pretty much line to line so we will see how it works out.

    I also want to make full 1 pcs TPU feet. I am going to try that next.

    Lastly, the hip board I made last weekend just shipped. Pretty good for a holiday week. Last week most Chinese companies were 100% shut down for the national holiday.

  • Deep dive on the sensor

    shane.snipe10/03/2021 at 14:55 0 comments

    I had some basic questions that needed answering. Firstly, how linear is the sensor. Secondly, what is the optimum distance between the sensor and the reflector ramp.

    For background, I am reading full scale 0-3.3V with 12 bit resolution or 4096 counts. 

    You can see for yellow, I have a roughly linear window from 2.66mm to 4.33. This is measured against a flat surface with cardstock as 0.33mm increments. 

    I am slightly embarrassed to admit took this data today, after working on the project for 9 months.  When I started the project, I played with some different resistors but because the sensors were not mounted to anything it was hard to measure the distance accurately. i saw  that the voltage changed a measurable amount with the resistor combination I chose and I moved forward to make some boards.

    Angle measurement 


    Femur measurement

    Unfortunately, my first design had the white PLA too close. I played with grayscale shading on paper and had good results so I tried grey. They reflectance was terrible there was almost no change. I then switch to yellow and saw much larger changes with angle but often on the last half of the rotation. I reduced the slope of the ramp and made it a bit further away and got good results. Here is a graph showing the counts versus angles for these 4 versions.

    After seeing the good response on this profile reprinted the ramp for the to the hip joint and got this response.

    Now if the other axis was similar, I could call it a day but I am getting some weird responses. 

    There are 5 things that are different with this axis.

    1) The reflective wall is printed vertically so it does not change in defined steps like the horizontal surfaces. The horizontal surfaces look like a staircase because of the layer thickness of a print which actually work quite well.

    2) There is support printed in the channel and remnants may be causing strange reflections.

    3) The outer wall is so thin, it looks darker from the inside as you can see the tape on the outside.

    4) The outer wall is thinner than the other two parts. 

    5) It is made of 2 parts so there is a seam that lets light in and it reflects differently.

    Looking at it more deeply, I have noticed there is a large axial play in the shaft. This axis is not trapped very tightly. This was because it was trapped tightly and I had to almost deflect the hip to put it on. I reduced the model by a mm off the adapter on each side and now there 1-2mm of play. Looking up the page at the first graph  we know that we only are working with 1.67mm of liner range and  being two close does not do us any favors.  By shifting the shaft back and forth axially,  I can change the output by 500 ticks. No wonder I had such a hard time to measure it. When I measured the left, it was biased away from the sensor, when I measured the right, it was biased toward the sensor, reducing the gap.

    I also see the sensor board is on an angle due to proud solder joints. This also makes it hit the side of the wall instead of the back and bouncing off. Combine that with a little support and that might explain why the end is wonky. 

    I will make the shaft adapters taller again so that it is line to line. The new boards with sit flat so proud solder joints will not be a problem. I will make the part 1.5mm thicker and push out the ramp by 0.5mm. The response does not have to be completely linear but it has to be continuous. If the slope change to negative my control algorithms will not work.

    Too much play


    Hip joint

    Part progression

  • Lots to do

    shane.snipe10/02/2021 at 09:28 0 comments

    When its 4 AM on Saturday and your dreams are you punch list, there nothing to do but get up and live your dreams! I have some interesting challenges to work through this weekend.

    1) Remake the hip boards with the correct pinout on the header. Add better silk screen. Check for the stock of the less expensive motor driver. Order the assembled hip boards. Order the longer FFCs from Aliexpress.  Fix the connection on the brain board  to the 2.1mm plug. The ground and Vin are reverse which would not make for a good day. I should make the files available for the boards incase anyone feels ambitious.

    Well I used the auto router for the first time in Easy EDA. The last auto router I used was on Eagle and it was not great.  I then changed to Kicad and it does not have one native to it. So I was shocked to see the following really nice layout. 

    The only thing I will change is the bottom trace that goes close to the mouse bites.

    I confirmed I need a B type FFC with pins on exposed pins on the opposite sides.

    The pin layouts are updated here.

    Files I used to order the boards are also in the file list.

    I also tried publishing it on the Open Source Hardware Lab.

    The integration is super smooth between Easy EDA and JLCPCB. It is not so important for bare boards but for PCBA assembly, I do not think I will go back to Kicad. If you have a self made footprint. I am not sure how you would upload it. They need a footprint defined in the BOM. I am not sponsored by JLCPCB but I will shill for them anyways as it is so cool to see the minute by minute progress updated live. I have ordered populated PCBAs now 4 times and 8 days later I have been getting them.  This order is about $3 each on a quantity of 20 and half of that cost is components.

    For the Jack output I swapped the power and ground that were mistaken.  When I created the symbol I made the pin numbers differently than in the footprint. Power should be at the back of the Jack.

    I put the archive file in the files section for the Brain Board. I am still struggling with libraries although I did archive the libraries before I archived the project. I will have someone try to download it and see if they can do it without getting library errors.

    2) In depth study on the encoders. Answer the following questions.

    a) Graphically compare the output from before and after I moved the ramp out by 2mm.

    b) Look into the femur joint which does not have a linear second half. 

    c) Investigate if tape color will change the output.  The tape blocks the outer light to make it more repeatable  but  may also affect the reflection.

    d) For counts over 1000 should I pull the ramp back a bit? Remodel, print and retry.

    e) Make sure the Onshape model is updated with the latest parts.

    3) Debug the issue with the batteries, BMS and regulators. Presently the regulators drop the voltage to 1-2V when used with batteries in series and the BMS. DC power supply voltage is fine. I expect high frequency effects are tripping the regulators. Maybe when Doug comes back from vacation he can look at it on a scope. For now I will breadboard it and reproduce it to ensure it is not a layout problem.  I will then add some capacitors to see if I can filter it out. Last resort is to swap out regulators or the BMS until I find a combination that works.

    4) Solder up a new flex to replace the one I put a screw through last time. The goal is to get a robot working with each leg operational in each direction. First thing I want to try is to make it stand up. To make it pop up from full splits, just using the hips. Next would be to get to full splits from lying on its back. I can hand manipulated it to do it but I am not sure if the motors have enough power to do it. Additionally I would like to see if I can use the 1:48 motors to do...

    Read more »

  • Lessons Learned

    shane.snipe09/27/2021 at 10:17 0 comments

    I have a friend that has been telling me that is not what I do that is important but it is what I learn. Well I thought I would explain the three things I learned this weekend in 20 hours of robotting.

    It is not always the things that you worry about that bite you.

    My boards came in as planned on Friday night and I was ready to assemble. 

    I was worried about 3 things. First that the FFC would not be tight enough in the connector and they would fall out as the legs moved. - Absolutely no problem on this point. The connectors were great and I could use the FFC as a handle to pick up the robot. -  No problems here.

    I was worried about the 0.5mm pitch FFC not having enough current capacity. I connected the wires and got a toasty FFC. Then I realized that I had designed in the B type connections and the pins were reversed. I changed the FFC to the one with pins up on one side and down on the other and this took care of the heating - No problems here either.

    The third thing I was worrying about was that I was regulating 3.3V and 5.0 and putting them both into the ESP as inputs. I usually only put in 5V. It seemed that in the end, it would with both 5V input and 5V and 3.3V inputs so it was wasted cycles thinking about it.

    I hooked up the batteries in series, got my 8.2V at the BMS output and I thought I was off to the races. I soldered in the 7805  5V regulator and the output at the BMS dropped to 2.0V.  I thought I burned out the BMS and I took it off the board and it still worked. I put a new one on and got the same thing. I cut off the regulators and everything worked. I tried the regulators separately and they worked. I tried with just the 3.3V, double checking the pins because it is different from the 5.0V regulator but nothing there. It did not work when it was on and worked as soon as I removed the regulator. I finally removed the BMS and ran voltage from a power supply at it was fine.  I put the BMS back on and it was still fine.  I tried with another different type of BMS and it did not work with batteries but did work with external voltage.  My conclusion is the BMS and the voltage regulator do not play well together. I guess I will just have to find some that will. The big take away is I need to breadboard more. 

    The second big lesson came from the next step. I decided to go forward with just the power supply and to finish the debugging. I then found this issue that ended my run with the new boards. My pin definitions on side of the header were different from the other. On the brainboard I was using Kicad and on the hip board I was using Easy EDA. On footprint was using alternating pins and the other was running ccw so even though the schematics looked the same, the pin definitions we totally different.  My end pins worked but the ones in the middle did not. Game over, time to order again. The big lesson is a I cannot skip the excel sheet showing the pins on each connection. Also, no more ECAD at one in the morning!

    Well I still had some closed loop programming to do and newly printed hips I was dying to try out so I went back to the Flex circuit design. I decided to run with a power supply and I tapped in a 7805 5v regulator to the 5V and  ground line.  I think ran wires down to the flexes from the input and now I had adjust power for the motors only and a stable 5V for the rest of the circuit. I conformal coated all the flex connections with hot glue and got on with the assembly. 

    The new prints with deeper pockets worked OK but there are difference between joints that I can not yet explain. I will do a more detailed log with some plots and consider it more. I put a screw through the left legs flex in assembly and lost movement of the left femur but over all I had something to work with. Putting in...

    Read more »

  • Stand up and walk!

    shane.snipe09/25/2021 at 12:07 0 comments

    Maybe I should first target moving the joints with a repeatable motion.


    To get there here are the logical progressions.

    Confirm the correct voltage is running through the system.

    I made some big changes with regards to voltage management.  The batteries are connected in series to the BMS that is on the main board. Previously I had a battery charger on each leg which pumped the battery voltage to 5V. This round with the batteries in series, 8.4V is regulated down to 5V and 3.3V to run the lower power processor and sensors. The native 6.4 - 8.4 Li-PO battery voltage is applied directly to the motor drivers for minimum losses and maximum power. The concerns I will have to investigate are:

    a) How much does the battery voltage degradation affect the motors?

    b) Is it going to be important to monitor this voltage and how can I do it? One idea might be to connect the switches to a parallel circuit to the main battery voltage and have them normally closed. 

    c) I am  sending regulated 3.3V and regulated 5.0V into the ESP. I have only done one or the other in the past and I am not sure if I am going to blow up the regulator on the board. I read that the regulator on the board does not have enough current capacity to allow the ESP to run WIFI which is why I am running 3.3V as well as the 5.0V but this makes two 3.3V sources to the same power input to the ESP. For now,  I will hook it up and hope for the best. If my ESPs start blowing up, I will remove the 3.3V pin on the module.

    Hardware Debug

    1) Get all the power components in place.

    BMS - Check

    Leg headers on board - Check

    Leg headers on hip board - Check

    FFC to leg boards - Check

    Connect batteries to Tibia board.

    (I just realized I named by lower joint as amalgamation of the Fibula and the Tibia. I called it a Tibula. I will go back to Tibia)

    I need to but more silkscreen labels on the board next time.  To help with understanding, I made a pdf which all the pin definitions for the legs. I remembered one key point. This robot will only work with B  type FFCs where one end has the exposed pins up and the other has the exposed pins down. It essentially transposes the pins in the cable.  I designed for it, forgot about it and rediscovered it in my QC of the boards upon receipt. Now that the world is straight and I know which pins to connect the battery to, I can go about confirming the power layout is going to work.  Here is the layout. Pin 2 is Battery + and Pin 3 is Battery -. 

    2) Once power is measuring good, add the ESP and make it blink.

    3) Add the screen.

    4) Add the PWM board.

    5) Read some sensors.

    6) Move some motors.

  • New connections

    shane.snipe09/24/2021 at 23:27 0 comments

    New boards. Right on time.

    Hip boards

    Hips

    Holes line up.

    FFCs fit great.

    Tibula and Femur Sensor Boards

    Connected

    This time for the win

    Holed all line up

    Screws are a little tight on the pins

    Power jack fits but I was thinking about it on the back. On the back it would have interfered with the motor so can we call it a happy accident?

    Hips on the bed.

    How to blink an led when it is on a robot. Step 1. Solder on the femur chip.

    Solder pins on for the BMS.

    Pins are swimming in the big holes. Bend them if necessary.

    Femur made a nice stand to solder but some space would probably been a better idea.

    Next solder on the headers for the ESP

    Solder headers on the back for easy removal of the legs.

    Solder the headers onto the hip boards.

    Solder the 7805 on the lower position for the 5V regulation and the 1117 above it back to back fir the 3V. Time to find out if my power circuit works.

    Sensor readings

  • Agility

    shane.snipe09/16/2021 at 11:26 0 comments

    I want robotics to be accessible.  I want to to be able to hold a make and take class at our makerspace and have the class be able to assemble the robot in an hour.  Well this week I had to admit the present design was not going to get there. 

    I worked hard on soldering up the PCBs to the flexes and if the wind was blowing the right way, I might be able to do it. More often then not, I tore a pad of the flex and I had to start all over on that leg.  Or I had a trace break when I was folding the flex and now one of the encoders could not get signal.  I figured it could have been improved through better and better flex design but this was hitting me in two places. 1st, the flex circuits were expensive  and 2nd they were hard to assemble. It was really going against the key tenet of the project which was accessibility.  The final straw came when Doug worked really hard on assembling a bot and despite having everything work as a flex, once he screwed the 3D printed parts on and connected it to the body, he broke a bunch of the pads between the flex and the surface mount header and all that work that ended the parallel path of development.  If the founder of a makerspace can not put my robot together, I know it is time for a redesign!

    So we were considering options and he mentioned a ribbon cable. The connectors for ribbon cables are too tall for this project but there are some very low profile FFC connectors! So I was off to redesign the robot to use FFC instead of custom flex circuits.  They are available in any length on Digikey and super cheap on Ali express so that checked the accessibility box. 

    If that is all hard to imagine, here are some pictures to help.

    Here is the board that goes on the hip. It has a 14 pin header that connects to the main board and then two 10 pin ffc connectors the head off to the femur board and the tibula board. 

    I turned the header 90 degrees to make room for the FFC connectors. I am not sure if  I have the angles right on the connectors but I only made 10 sets this time and I will revisit it after I have put some together.

    The headers at the top are just for debugging. The through holes J1 and J2 at the bottom are for motor wires. The double row of headers connect to the head board. There is a rectangular hole in the middle of the board to allow the board to sit better on the motor and accommodate the hook that holds the motor strap.  It still will sit proud by the thickness of the motor strap but I think it will be acceptable. Otherwise I will make a 3D printed spacer again.

    For the femur and tibula boards,  I was able to make them common with each other.  Both have the ability to connect to the battery and either battery connection will run to the main board through the FFCs and the 14 pin header. Here is the smaller board layout. 

    Ignore the headers as they will not be populated. I still have not figured out how to put a connected plated through hole without associating it with a header. If anyone knows, please leave me a note in the comments.

    The battery will connect to J3 and J4. Which brings be to the battery management redesign. I really wanted the robot to be able to pop up but even with the 5V boost circuit, the motors did not have enough power to stand up from a splits position. What's the solution? - MORE POWER! - I considered putting back the battery boards I started with that had a adjustable voltage gain but it felt kludgy and I was not sure there would be enough current capacity. So I decide to connect the batteries in serial through a BMS on the head board. This takes the battery management out of the legs and gives me 7.2V.  It also allows the motors to be charged with one  9V adapter that plugs into the main board.  

    Tomorrow I will describe the main board and all the changes but last night I ordered new head boards...

    Read more »

  • Video proof!

    shane.snipe09/05/2021 at 12:28 0 comments

    Well since I have something that now moves, it is all just software. Twiddle a few bits and Cya is going to dance. There are some many things wrong with that statement. First hardware is never done, it is just  done enough to get by and I do not think I am there.  As for the bit  twiddling, I need to back it up a bit and think about how I want to control Cya.

    But first, a little glimpse of that I have been doing. I had mentioned a trestle which I imagined to be  something to hold the robot up as it tries to walk. Well, two chopsticks, 6 yearbooks and one milk crate later, I had something workable. 

    If you look closely, you will see the encoder positions are read and displayed on the  screen  at the endpoints.  This is one of the main reasons I wanted a screen. It so much easier that relying on  the serial terminal for to understand the encoder limits.

    So how to go from having a hardware platform that can move its joints and be aware of where they are to dancing?  I guess the first point is to think about how to control the positions.  Traditional robotics us inverse kinematics.  This is basically where you know a position in space and then you calculate the angles of the joints to get there using  trig. I think this is a non-starter for any non-roboticist so lets find a short-cut!

    I am going to propose that movement is just stringing together a bunch of discrete positions and if the gap between the positions is small enough, you may not care how the robot moves to get there.  So if the robot has of list of targeted positions and it moves from position to position, some reasonably compelling motion can be created.

     Now, how can I make this easier. What if each joint has positions 1-100 with 1 being the smallest angle and 100 being the largest. With my low resolution encoder and joints with backlash, this is probably an appropriate resolution for the joints.

    So what are the steps to get there.

    1) Map the encoder output to 1-100 and display the joint position on the screen for each joint.

    2) Manually move Cya's joints to the desired positions to make up the choregraphed movement. Record the sequence of positions.

    3) Code the function that moves each joint to the desired positions.

    4)  Call the function to work through the array of points to make one cycle.

    Stretch goal would be:

    Take points from one robot and transfer them to another via ESP Now.

    This seems simple and straight forward but it also is different from what I have seen other projects do. First of all, unless you have super high end servos, you do not have positional feedback available. In other words, you can not move the robot to a joint and record the position. You have to go back to the inverse kinematics. Another way to get at it is to use reinforcement learning to get to a set points. I was working with a team to do this but unless you can ignore inertia of the limbs, the math is too intense for model to work. This is why you see a bunch of stick legged robots. So on my team, the computation guys kept pushing the mechanical team to pull the weight and actuators out of the legs.  However, I think the farther the weight is from the ground, the hard it is to make the robot balance so these are conflicting goals. Sacrificing physics so you can calculate something makes no sense so I went the opposite way on Cya and put the batteries as close to the ground as possible. This should make the robot more stable. We will see how it goes!  


  • Leg assembly

    shane.snipe08/31/2021 at 01:40 0 comments

    Here is my attempt to document an assembly attempt. I was pretty happy with the way this much of it came together. The plastic assembly is supposed to be the easier part.

    Assembled flex


    Backside


    Remove adhesive backing

    Monday morning quarterbacking this, I see although sticking it down makes the assembly a little easier, I is causing way more issues. Next run I will not stick it  down until everything is together.

    Mount on motor

    I put a header under the board and came up through holes in the flex. The two pins on the right were great and it holds it to the flex because there are plated through holes to solder to before the board is put down. This is power to the circuit. The two pins on the left are for charging and are OK but  a little tight into the foot housing. The two in the middle should be removed and replaced with just wires. I soldered to the pin headers I bent over here but then had a hard time putting them in the foot and getting the battery in the foot. 

    Stick down the flex


    Mount in the foot


    Check power.


    Squeeze the battery in


    All assembled

    The one off button has almost no room. It is a momentary switch but it needs to be pressed to get power to the circuit. Need to make the window bigger again. The board is not sitting well because the bent header pins that connect to the wires that go to the battery are wedged under the housing.

    Cut off the tops of the motor

    The motors butt up top to top and the flex is in between. Cutting the tops give more room for the flex.

    Add tibula


    Screwing it together. Electric driver is a life saver.


    Major Fail!!! Ripped the solder pads off the flex. Game over……

    Once the pads come off the flex, it is game over. I think if I go with a double sided flex, they might be more robust. I also might try a thicker flex in general. Lastly, I seem to have the most trouble when I put down the adhesive. There is no longer any give and things start to tear. Also the adhesive backing makes it less likely to tear as well.

    Another broken trace in the corner

    Here is another example of a tear and you can see it cut through the trace. After this I turned on the battery and connected it to the head but no power went up so its back to the soldering iron. 

View all 25 project logs

Enjoy this project?

Share

Discussions

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates