the finished CAD and finihed Full project

so i actually just finished everything

the overall

the CAD include great design with a lot of great features like:

• the amortization spring
• the calculated segment leghts
• detailed ready for printing design

the spec

the robot deisgned to be fast stable and smooth so thats why

• Tibia - 150mm
• femur - 130mm
• coxa -50mm
• full size 600mm
• height 175mm

these values allows the LX-16A servos to work correctly and control the robot

3d printing

the robot consists of several parts for 3d printing overall weight of the 3d printing parts is around 900g calculated by bambu studio
the robot includes :

• cental case
• central case cap
• battery holder
• coxa servo mount
• coxa-femur joint
• femur arm consist of 4 seperate parts
• the tibia arm consists of 4 different parts
• the amortizaion spring and foot

im using PETG over PLA bexause it can handle bad situations

THE NEW ROBOT DESIGN

NEW PCB Design

decision

this pcb was created because previous PCB consisted the 2 buck converters which is really complicated and worst desision ever so i decided to remove them and power the servos directly with the serial bus servo motor LX-16A which is the cheapest bus servo i found with UART which can handle the 8.4V from battery and wont burn itself

The PCB SChematics

Power Stage

the LX-16A servos can reduce up to 2A of stal current which means around 36A for all them but only part of them are gonna work in Peak at the same time which makes the basic current around 9A for whole robot and 30A fuse is making it the max current in the robot thats why all of the next components are gonna be able to handle more than 30A

• XT60
• IRF4905S
• TVS DIODE
• and 4 big capacitor
• INA226(for current monitoring)
• resistor divider (for voltage)

Converters

so to power the logic i need 2 buck converters from battery to 5v and from 5v to 3.3v i used same buck converter for the powering and choose the value for resistors using the formula from datasheet

• 2x TPS563201 buck

Servo bus

so the lx-16a servo motor uses UART protocol and have BUS logic signal the combine of TX RX because MCU uses RX TX but we need BUS i used half duplex trasistor to make it work and also added second half duplex because 18motor for 1 transistor can be alot and dividing it into 2 parts makes the signal a way better and clean

• 2N2222
• 6 ports - because the UART works like a chain so you can power and control all 18 servos with 1 port but we have to use 6 ports 1 per leg because otherwise the wires will stuck and wont be connected correctly

MCU

the mcu is esp32-s3 because it basic and best microcontrolleer with wifi so i can control the robot with wifi which makes it easier

• ESP32-S3
• USB-C
• buttons
  *LED for feedback

IMU

i used BNO055 because while reaserching about IMU i found out that the mpu6050 and other costs the same as the BNO055 but the BNO is much stronger and the best part is that it has it own computing power which helps the MCU which is already have too much to compute

the Layout design

the PCB include 4 layers

• 1layer power zones and logic signals
• 2layer GND
• 3layer Logic Power and signals
• 4layer Power zones and logic signals

The placement

i placed the components so the power components will be seperated from Logic components to make it efficent and work correctly

The second update of schematics

The FOC driver Schematics

currently finished parts

• the Power stage
• the gate driver
• the 6MOSFET
• the MCU

im going to finish the schematics in next session where im gonna work on usb tramsimiters encoder and etc

The Power stage

Overall- the power stage include 24V income ,reverse polarity protection on p-mosfet , and the TVS protection from power surges

24V - im using 24V as the income because the QDD module is gonna be powered 24V so the 5010 bldc can perform up to 6.5Nm by calculation 0.4Nm x 20x reduction x 80% efficenty where 20x reduction can perform to 400rpm which is perfect

The gate driver

DRV8323RS - is used because it is perfect standard of FOC drivers and has its own buck converter which gives us the VCC line 3.3V for logic powering as the reference i used Texas instremnts DRV8323 Datasheet schematics

The 6 MOSFET

CSD18540QB5 - i used a bit different way than in the refence using 6 Rs for each of the Mosfet

the MCU

STM32G431CBT6 - just basic MCU i also added the CAN RX/TX which is gonna wotrk with transmittors work ESP32-S3 for multiple QDD modules

Conclusion : i actually think everuthing is going well because im using datasheets and references from right sources and for now the schematics looks perfect im just gonna finish other parts of schmetics in next session and then start working on PCB which is definetely gonna have 4layers and maybe even 2 sides of components because i need compact but performing well pcb for the QDD module

THE ACTUATOR PCB

This is first day of working on the pcb schematics today as you can see i covered the power line part of the pcb schematics

The main components:

• XT30 - the connector which gets 24V the xt30 pretty small and can easily handle high power income

• SMAJ40A - is a TVS diod which we need to secure the power income from jumps and etc

• AOD4185- is high power P-MOSFET for handling the power

• buck converters ml2596s and ap2112k-3.3- is the step down buck converters we need to power the logical components

Features of the power -

So actually im choosing this components because they can secure and give the perfomance i need in my pcb powering to prevent it from fireing itself and etc

power perfomance for components

• the 40v+ because the income voltage is 24v we need at least 40v to make it safe from peak jumps

• easy to buy im buying all the components from aliexpress and there is a lot of problems to find right components that why it has to be easy to find anywhere so me and other people can easily find the right components

• the qdd module is gonna perform around 20A so i need at least 30A for safe

almoooost finissh CADDDD

The Design

So actuallyy the design is look exactly what i wanted to see before and im so so happy with this abiut overall construction we have today :

• coxa mounting is fixed and changed : about it i decided to start iver and make the coxa mounting better use less plactic and make each mounting seperate which allowed me to make the rangw of motion bigger because of angle of rotation of the coxa which is around 70 and im pretty sure it is enough

• added a but longer femur for bigger range of angle rotation

• and also added long curved Tibia part which is also really great

the mounting

all the parts usuallly mount with just screw and nuts but i also added holes for insert where i need 100% stability and where i have place to place the insert because it is ussually around 5-6mm hole for M3 screw and nut and around 10mm for insert nounting

the overall

im really proud of this design and think that this is gonna be great and now i just have to fixe some small parts in CAD and make BOM and etc

New design

Yeah i finally made new desing of robot

the Genrative design problem

after last attempts i actually got really good results with pretty great design but i think that they are pretty weak thats wwhy i spent all day waiting when it will generate a tonn of design with different constarins and loads and come up WITH result that im gonna make it myself and refuse the idea with light and futuristic frame

the Current design

it is pretty good i think i added the placement for bearings which will hold the axis from the oposite side of the servo shaft mirror it all and now have beatiful design of coxa

the size

about the size of the leg now im thinking to make the fibia 2 times longer than the coxa and the tibia 2.5 longer than the coxa which means the leg should be right propotional

the leg Propotions : 1:2.0:2.5

Failures again

generative design

so i have no idea now why is generative design is doing evertyhing wrong because i used everything i know and all my creativity but still this is not what i want and i think i have to just quit the idea with generative and Use FEA and a lot of a lot of time improving

the main problem

the main problem with the generative is that is always change everything in preverse and i thought if it failured when i 1. change the objective 2. constrains it is just trying to decrese anount of plastic in preverse what i dont want it to do i think

the ways i have left

• i can try again and i think i will make it sucsses next time
• i can make it myself even if it is now as good as the generated design

probably gonna make it myselg for now using generative only for small easy parts

the Finished Ultimate PCB!!!!!!!!!!

i’m actually so happy to say that this pcb is finished because now i can finally work on my CAD of robot

so here is what i did last 5 hours :

• firstly check all components of the schematics work on improving it add some extra capcitors replaced all CP 10uf to 10uf C
• added all Footprints
• placed everything and route everything

Problems

Of course i had problems while working

• The Footprint problems : i had no idea what footprints should go to usb and my buttons which is bi problem because if you choose wrong one you are definetely gonna regret it because of mounting
• The routing problem: i think it is the problem where ispent most of the time so the main reason is the placing which is really compact and makes it really hard to route and because of that while routing the 32 PWM ways connectors everything been wrong if i tried to solve one problem it causes another that why i spent a lot of time changing everyghuing all the time

Overall

so i can say that now i have ready PCB with 4 layers and compact size which takes the project to higher level

yeahhh and now i have to go study exams….

the Schematics of ultimate PCB for hexapod

include:

• 2 screw terminals for main servo power
• 1 screw terminal for 5v batt+ and gnd
• 32 sockets for 32 servos( it can handle 32 servos but 18 servos is enough for hexapod added just in case if some one want more servo )
• 2 PCA9685 for controling servos
• BNO055 for understanding the angle ( i replaced the mpu6500/9250 to bno055 because bno055 has its own brain and can calcute it while mpu cant and the price is the same sometimes mpu even more expensive than the bno055
• 2 resistors for batt sense
• esp32-s3 as main brain with ams1117 3.3 and smd buttons and usb -c port
• some LEDs for feedback

overall of schematics

i think this version is much more better than previous and will be working much much better now and gonna be more compact

next step ideas

now im thinking about finishing the pcb as soon as possible and learn to use generaive gesign properly so i can use it to make light and strong parts using 3d printed PETG because 50-60cm robot is gonna be pretty big and making it as light as possible so DS3235 servos can handle the robot easily is really important

The new ultimatee PCB

why did i start New PCB

Actully while making the CAD i was just thiinking how to make it beatiful moveable and just great and i started searching and found a lot of beatiful hexapods and found out that i really need to make my centre part thicker thats why im making the pcb with all parts in it and also to use 2 PCAs as i know i have to change the adress but i didnt really wanted to make it in ready ones

added to PCB

• 2 INA226 for collecting more data from robot
• 2 PCA9685
• ams1117 3.3
• esp32-s3 wroom1
• usb-c-16pins
  and also replaced the xl4016 8A to 300W 20A buck because 1 servo can require up to 3-4A and 18 is gonna eat around 50-60A if include the fact that im gonna use 2 buck converters it is gonna handle up to 40A and it is enough because only 6 servos gonna work on peak and only half of servos gonna work at one time which means that the 40A is gonna be a lot

Overall

im gonna finish the schematics and the PCb next session and use this small PCB in the robot and also add use FEA everywhere to decrease the weigth of robot for better performance

So im proud of myself so much because now the hexapod looks pretty good i think im done with the coxa servos mounting

the failure of Generative design

So today i forgot to record the time when i worked on my generative design stuff i wanted to make the robot really light and strong but the generation from the fusion is really bad so i had to make everything my self of course it is not as good looking as the Ai generated super robot but anyway i think it looks great for now

i aactually tried to desrease the weight of the robot as much as posiible by adding holes and etc which also works as the wire managment ways

So this session i finished mounting the electronics and made coxe servos mounting place

about some problems i am really confused how am i gonna design the other parts of legs stable and with great performance and also about thw overall look of the robot because i didnt design fully custom pcb and the Xl4016 are really massive the centre of the robot is gonna be pretty heavy and ugly

new idea

sooooooooo now im thinking to learn and use generative degisn to make it a waaaay lighter and look like ALIEN warrior spider robot

its session 5 if im not wrong and today worked on making the mounting of pcb correct and also worked on the CAD of main centre section and placement of electrical components

PCB Changes

• firstly i fixed the mounting holes place on the pcb my matching the pcb holes with pca and mpu holes
• also moved the left PCA a bit to the centre so the fisrt PCA will not interfere with the second PCA terminal block and all wires can be placed easily

CAD

• choose 14 side polygon because it is the best matching amount of sides for the good placement of servos

so overall now i have finished PCB i guess and started working on the CAD for the Hexapod robot

so this is session 4 of working on pcb and i finally finished second desgin of pcb which looks amazing

Added

• LED lights for better feedback
• a bunch of mountong holes for better placment of pcbs
• bigger capcitors
• better come and out of V servos actully seperate them into a servo line b servo line one for first xl and second for second xl

overall

i wanna say that this pcb maybe not the last one but at least great one it contains esp32-s3 mpu6500 and 2 PCa 9685 on the board and 2 XL4016 with 1 small buck down stepper also some capcitors for smooth robot performance

So last session after finishing the pcb i opened the fusion autodesk and it was blank after spending 2 days on fixing the issue i finally did it but understand that the PCB is wrong so firstly it was really ugly the PCAs mounting is really bad and i decided to make new one

updates

• i decided to add second Xl4016 because i figured out that the one xl can handle only 8A but 18 servos require a lot more so i put second one 1 Xl4016 for 1 PCA9685
• added some capcitors for smoother performance
• added LEDs
• added DIOD for protection
• i also added mounting holes for whole pcb and PCAs with MPU just in not this photo of schematics

so overall i think this design of PCB is gonna be a way better at least because i reasearched and figured out my problems

PCB V1

so this is session 2 and i made pcb
so today i had changed some things :

• i have changed the schematics a little bit by adding testpoints
• add the footprints for each component
• and made the rounting with 3 different width of lines
• added 2mm super fat lines to make it possible to work because 18 servo motors is gonna eat a lot a lot ampers and 1mm 0.5mm is just gonna fire

Todays Problem

so actually while making the pcb i got some problems fisrtly in my head the battery voltage goes directly to the xl4016 which means its not gonna get to pcb but i needded the battery voltage for that 2 resistors to understand the current voltage and battery so i decided to solder second line of wire from main wire which goes to xl4016 and head the secondary wires of battery voltage to 2 pads 2.5x2.5mm on the pcb

hi its first session working on my hexapod robot named MADHEX (name come from MAD is first 3 letters of my name and HEX is from first 3 letters of Hexapod )

So in Schematics we have :

• all of the components is ready modules and pcb is just placing one to make it clean from wires and etc
• Esp32-s3 i use it because i have it at home and i wanted to find the place for this microcontroller anyway its really great option for the controller and im pretty sure its gonna run inverse kinematics easily
• Pca9685 x2 so because the hexapod require 18 servos it need also 18 pwm signals but 1 PCA can provide only 16 what why we use 2 of them
• the step down converters i use 2 of them XL4016 for stable voltage for the servos because its big and great and mini 360 for 5v logic
• Capcitors and Resistors so as you can see in schematics i use C1 and C2 for stabilizing the voltage jumps and just making the robot smoother and resistors for Battery sense to make robot understand the current battery
• MPU 6500 its just basics for every robot and for hexapod it is the must have sensor but i choose the 6500 not the 6050 because 6500 is newer

so this is the schmeatics and everything about it it is great and powerful and the main idea is that if something is gonna brake i will just replace 1 component and it will cost cents and im not gonna replace the whole pcb