Path Forward for the Arlobot Robot Chassis

[thomasco...@gmail.com]

Hi Chris,

When do you expect to move forward with replacing the discontinued DHB-10 Motor Controller?

I have responded to your comments on open issue # 176.

Regards,

TCIII

[thomasco...@gmail.com]

To anyone who wants to build an Arlo knockoff robot chassis.

If you want to build a motorized base similar to Arlo, I have included a motor/encoder/mounting hardware/wheels/rims parts list that can all be purchased from Servo City.

The total for the two motor/wheel assemblies is around $150 plus shipping.

A 16 in dia  black Acrylic base is available from here for $24.95 plus shipping.

Cutouts will have to be made in the base to clear the wheels like on the Arlo base unless you want to mount the wheels totally under the base.

Additionally front/rear casters will have to be purchased depending on how the motor/wheel assemblies are attached to the robot chassis.

Regards,

TCIII

[Christen Lofland]

Step one is to get my Raspberry Pi to work consistently for the existing setup. It is just shutting off randomly. My best guess at the moment is that I have a bad Raspberry Pi board.

I have many Pi boards around the house in operation, some for many years, and I've never had one behave this way, so I'm guessing it is a hardware issue. This is both hopeful and sad. Hopeful because swapping it out is an easy fix, but sad because I'm not sure if/how to get this board replaced.

Once I can get my existing Arlobot to operate consistently with a Raspberry Pi board, then I can start thinking about building a new generation of bots with a Roboclaw controller.

It sounds like you have found a few hardware options that work. I'm interested in the Servo City motor that you mentioned on GitHub (https://www.servocity.com/5202-series-yellow-jacket-planetary-gear-motor-99-5-1-ratio-60-rpm-3-3-5v-encoder/) and it also sounds like, from a code perspective, that either motor should work the same other than possibly settable parameters.

So once I get past my Pi issues, I'll start looking at slapping together a minimal bot.

The biggest obstacles at this point I see are:

1. My time. Life is so busy lately. I am a father of 3 kids (1 in College, 2 in High School) and I have a full time job.

2. Money for parts:

 - I have some motors and I can slap together a chassis, but I either have to decommission my trusty Arlobot to scavenge its RPLIDAR or I need to buy a second RPLIDAR for the new test bot.

 - I'll need another PI as well (possibly 2 more if my existing Pi on this Arlobot is bad.

  - I'm not a poor person by any means, it is just a matter of priorities and timing. I might need to wait a month or two before I'm flush with cash for things I don't need. :)

This has been a very much back burner idea because I didn't think anybody would use it, but with you testing things, and the possibility of designing a fully "open source" robot that anyone could build and run this software on with off of the shelf parts instead of using one company's dying product, it seems like a fun idea again.

[thomasco...@gmail.com]

Hi Chris,

You might want to check the power supply that you are using with the Rpi.

If you are running a Rpi 4B then you really need a power supply that can source around 3-4 A.

I assume that the red LED power light is going out or does it stay on and the Rpi quit responding?

Regards,

TCIII

[Christen Lofland]

I had the Pi running off of the official CanaKit power supply and it was shutting down several times per day.

I pulled the Pi off of the robot, and now it has been running all day on my desk, so it is going to be a fight to sort out what the problem is.

Fortunately, I was able to pull another Pi 4 out of another project, and I'm installing the Arlobot code on it now, so at least now I can see if they two behave differently.

This provides a lot of testing for the setup procedure anyway. :-D

I'm already discovering that this setup script isn't really ideal for a Pi, which makes sense, as I never got as far as testing it for Pi on purpose. It is nice that it works as well as it does, but I'm going to improve it.

---

I definitely need to look into the alternative boards you mentioned, although since I play around with so many things, it is nice to just have one consistent hardware platform.

For instance, when one of my boards gives me trouble, it is nice that I have others around. :)

---

I will keep you posted as I go along. Once I feel good about my Pi setup and the Pi on my Arlobot, I'll see about rigging up a Roboclaw to one of them. Fortunately I do have a few of those already. I don't technically need to build a full robot to start toying with the ROS Roboclaw nodes. I just need to slap the hardware together on my desk.

I realized I also have some old "lidar" scanners in a box that I can use for testing. They aren't great, but all they have to do is prove the concept while I sort out ROS and Roboclaw.

I agree that the A3 is a bit expensive. Because I have already purchased 3 cheap lidar units before I discovered RPLIDAR I was determined not to cheap out a gain, so I spent some months saving up to get the A3. In hindsight, the A2 seems fine. At some point I hope to pick up a couple of the A2 units when I find myself flush with cash again, as they seem like a good compromise, although I should evaluate how well Slam Toolbox works with some of the other even cheaper units too.

So many ideas. At some point I need to post some information about the other "robotic" project I'm working on, but it is technically more "animatronic" than robot. My goal though is to merge that project with this one to create a sort of easy setup framework for making your own robot/remote control device. Although the donkey-car project looks like it may already cover a lot of what I had in mind, so we'll see.

[thomasco...@gmail.com]

Hi Chris,

Thanks for the update and your expected path forward, much appreciated.

A couple of questions concerning adding an Arduino for NeoPixel control and an Adafruit USB/Serial 16x2 LCD display:

If I want to use an Arduino with Arlobot ROS, it appears that I need to determine the "arduinoUniqueString" : "?" and "arduinoStringLocation": "ID_SERIAL" of the Arduino I am using.

I believe that this Arduino program will give me the "ID_SERIAL" of the Arduino:

//

// ArduinoUniqueID.ino

//

// Example shows the UniqueID on the Serial Monitor.

//

#include <ArduinoUniqueID.h>

void setup()

{

Serial.begin(115200);

UniqueIDdump(Serial);

Serial.print("UniqueID: ");

for (size_t i = 0; i < UniqueIDsize; i++)

{

if (UniqueID[i] < 0x10)

Serial.print("0");

Serial.print(UniqueID[i], HEX);

Serial.print(" ");

}

Serial.println();

}

void loop()

{

}

But I have no idea how to determine "arduinoUniqueString": or should I leave it as "Arduino__www.arduino.cc__0043_8533434373835150C141"?

Also, I have several of the Adafruit USB/Serial 16x2 LCDs and once again how do I determine "LCDString": or leave it at ""Adafruit_Industries" and how do I determine "LCDLocation": "ID_MODEL"?

Can I use "udevadm info /dev/sdX" to determine the LCD's "ID_MODEL"?

Regards,

Tom C

[thomasco...@gmail.com]

Hi Chris,

Bump?

Regards,

Tom C

[Christen Lofland]

All of those strings are obtained by running:

udevadm info -n /dev/ttyXXX#

Where XXX## would be like USB1 or ACM0

You will see a lot of info for a given device, and hopefully you can see that the options there are saying, "Which line should I look in?" and "In that line, what string should I look for?"

I presume that YOUR Arduino will have a different number, but I could be wrong, they might all be the same.

Hope that helps.

[Christen Lofland]

Good news is that both of my Raspberry Pi 4's have been running for over 24 hours now. Not sure why the one was freezing up, but reimaging it from scratch seemed to help.

The one on the robot seems happy too.

I was able to do an entire build from scratch and it all works, except for one issue with the RPLIDAR, but you haven't gotten that far yet.

Issue #187 seems to be some sort of out of sync issue between your setup and what the setup attempts to create. Perhaps from adding and removing slam_toolbox a few times?

Hopefully we can sort through that in GitHub.

[Christen Lofland]

Random update: My 8GB Pi crashed again. Good news is that CanaKit is replacing it no questions asked!