Canada
Asked

Affordable Modular ARC Robotics Kit

Is it time for a new ARC compatible Robotics Kit? Most of us came across ARC back in the days of EZ-Robot.  These are great robotic kits for schools and summer camps to teach kids robotics.  The problem is the cost and availability.  ARC now works with ESP32-CAM and PCA9685 controllers. Together these provide a lot of the functions of an EZ-B for around $5.  So now using low cost servos like MG90S or other 9g servos you can build a bipod robot that works with ARC for under $50.  Using a modular design the pieces could also be used to build other robots like Hexapods, Robotic Dogs, Rovers etc.   ARC has all the functions built in to support a new robotics kit all we need to do is design the connectable servo casings, servo horns. body and other components for 3D printing.   We could then provide the free 3D printable models and even make kits available for sale.  ARC makes it easy to program, build and configure the robot and this would make robotics a low cost barrier of entry for students and still teach all the skills needed to learn basic robotics with all the features that now come with AI integration.

There are a bunch of opensource ESP / PCA based  robots available that you can make work with ARC today,  but I think a modular design that allows you to build a range of different robots that have pre programmed projects, scripts and build instructions available for ARC would be the best approach.

Anyone interested in working on something like this ? 

User-inserted image


ARC Pro

Upgrade to ARC Pro

Take control of your robot's destiny by subscribing to Synthiam ARC Pro, and watch it evolve into a versatile and responsive machine.

#89   — Edited

Sorry about the long post that follows so I'll add this up front: If you guys end up building another level of EZB I think it should have a couple changes that would make it more appealing and easy to use for the common robot builder.

  1. Less expensive
  2. Have a better way to power it up other then the little plug that now sits under the unit and plugs into a power base. Get rid of the powerbase and add some kind of power jack on the side of the board (like the one the V3 EZB had and the current IoTiny has.
  3. Add a jack to output audio rather then having to use that small onboard speaker. If that jack was used, the little speaker would need to be disabled or removed to keep the robot from having an echo effect.

Other thought on this subject: Well, I'm excited that there is a conservation about finding a comparable alternative to the current EZB. Don't get me wrong, I love the EZ Robot's EZB and have used and suggested nothing else. I'm no expert but on a consumer level I believe there is nothing better available that offers as much power and function in one package.

I will admit that the EZB probably overbuilt for the common hobby robot builder because of all the stuff it can do smoothly and quickly and the protection it offers. I think it's more suited for advanced robotics, industrial and learning environments as it sits. Endeavors like what @fxrtst, WIll, is undertaking with his robot classes is like a gateway drug for robot builders who want to make their robots do a lot more then roll around the floor and need to know what is needed to make that dream happen.

However the price point for the EZB unit and software (the only way to get it to work) is above what many hobbyist what to pay. There are so many builders just in my club alone (The B9 Robot Builders CLub with almost 2k members) that will not use the EZB/Synthiam combo because of the costs (both being the one time purchase of the EZB with camera and the Synthiam's $100+ a year subscription) . Even using a flashed microcontroller they refuse to pay the Synthiam subscription or the high priced EZB when a mini costs under $20 and the programing is free. Again, to me, the price of the EZB and the ongoing synthiam subscription is worth it but I feel I have a more advanced robot then most are building in my club. I understand that the combo of EZ-Robot and ARC has allowed me to get to that level but most builders don't want or need my level of animation.

Even in my build I could benefit from a cheaper, stable and powerful controller to control some of the less critical functions. More so when I mess something up and burn out a $100+ EZB.

I have a friend in the club that uses many Arduino's programed with his own scripts to get his wanted level of control and animation in his robot. When he's done he will have over a dozen Arduino's running different functions both autonomously and with a game controller. That sounds overwhelming to me but that's what he's experienced with and he knows how to get it all working together. In his opinion, his budget wont allow him to go the EZB/ARC route. I point out that he would probably spend as much on a couple EZB's and Synthiam's subscription (with ARC in Read Only mode). He tried it and junked the idea. He's been a C++ programer for 20 years, with a limited budget and choose to go go with what he knows and feels he can afford over the life of the project. In truth as it turns out for a lot of builders, building a B9 project is a lifelong build.  So there is that.

Author Avatar
PRO
USA
#90  

Quote:

I started to try and collect functional and non functional requirements earlier in this thread and as DJ stated maybe we should continue on that process.
Get back to it fellas.

I ordered more ESP32s and I have an SSC32. DJ exposed the UART on the ESP32 in the latest firmware. Going to test later today. That would make a MEGA_EZB_EPS32_SSC32 with control for upto 32 servos AND camera all on wifi! That will fit into my GAR hexapod and cover all the servos (there is alot).

Both Jer (thru EZ Robot) and DJ have their businesses to run or participate in, and making money with their products is how they keep the lights on. They are not in the business to save the (robot hobbiest) money. If they were the EZB and camera would be $50 and ARC would be $2.99 a month or free. Its been said here on these forums 1000 times in 18 years...the money is in education. The hobbiest just had to take whatever pricing was set for education market. BUT what is different now with the split from EZ Robot is DJ now supports other controller options. He is actively supporting those boards with updates. Not a polished hardware product but a DIY option. The solution for affordability is already here and continues to grow. EZ Robot on the other hand has most likely saturated the educational market has become stagnant due to lack of new cash flow. Too bad Ricky Ma's robot head (Mark 1) didn't produce an additional revenue stream.

@dave brings up a great point though with ARC and subscriptions. We old timers remembered when it was free so it pains us to look back. New hobbyists are looking at the subscription costs and saying no way and looking for other solutions. Its something I had to think about with my students and ARC Subs. I could have a problem with my students feeling forced into subbing to ARC to explore more robots functions, (although they do have other options for software/hardware like the SSC-32 and servo Sequencer or the Maestro Control Center for servos ) but they will not have access to higher level tools in ARC like object recognition, etc. Unless something comes along and competes with EZB/ARC combo I think it will be status quo (as it has been), and back to point 1, they have businesses to run and profits to make. I cannot imagine the overheads.

@Nink. That's a good idea with using a serial camera connected to UART of the ESP32.

Author Avatar
PRO
Canada
#91  

This is what I think I am going to use for now.  Turns out the original ESP32 did have a DAC but they removed it on the ESP32 S3 so we have audio if I use old ESP32 classic and since we offloaded the camera we will have more than enough free GPIO ports.   I ordered the mini ESP32 S3 + OV5640 for the seperate camera but in meantime I will an AI thinker and either serial connect it to the main ESP32 or Wi-Fi connect to ARC.

I will also add 1 PCA9685 for 16 PWM ports and a Buck converter to power the servos off it  I can create the other 8 PWM ports from the GPIO pins on the ESP32 for a total of 24 Servos.

If that all works I will 3D print a new case to mount all the components.

User-inserted image

Author Avatar
PRO
Canada
#92   — Edited

Looks like the PCA only runs at 6v so I am going to need a better buck converter. 

User-inserted image

Current one LM2596 is about 2 to 3 amps.  Can power esp32 Cam and sensors but useless for 16 servos. 

User-inserted image

I just ordered one of these.  Will see if I have something else kicking around in the meantime.  EZ-Robot servos like to run high voltage so let’s see how they perform at 6 volts.

User-inserted image

edit  interesting Adafruit says you can run 12v if you only use the V+ or header block for servo power and don’t throw higher voltage on VCC pin.  You would think that was obvious so why do they write 6v for V+ Max on board.   (Already smoked 1 PCA board during earlier tests)

User-inserted image

Author Avatar
PRO
Synthiam
#93   — Edited

They write that because the traces on that board are so tiny, it can't handle much current. that's why the power header is so small because it's not rated for enough current to power 16 regular servos. maybe it's good for those tiny micro plastic servos.

just be careful and keep an eye on stuff. the reason u can blow up these cheaper boards is there's also no gpio protection. the ezbv4 has isolators on all i/o so they're less prone to being damaged by static or accidentally shorting or pulling too much current etc...

Author Avatar
PRO
USA
#94  

Huh that's weird (6v -12v) on the PCA9685. I didn't notice. I'm going to be offloading the servo power, to a power distribution block. So only a signal and ground will goto the PCA9685. But means sacrificing space.