Inmoov - D.cochran Style

 
#21

I just timed it. A reboot right now takes about 10 seconds total. It takes 4 seconds from off to asking the user to log in. I will take that. Once other services are added, the time will get a bit longer, but not much. Very workable in a robot.

#22

So far, I have decided to do without a couple of components, mainly because they aren't needed.

The touch screen listed was giving me some issues, but downloading the drivers from the vendors site made a huge difference. The monitor came with a 90 degree HDMI adapter so that wasn't needed. Because this is working so well, I am stopping my battle with the bluetooth radio on the motherboard. It can win, im good with it. This causes me to not need the mini keyboard, which too works great on everything I try it on except for this computer.

I will be using my robot outside during the summer months. Because of this, I have opted for a very large CPU cooler on this motherboard. If the robot were only going to be used indoors, I would have stuck with the stock cooler. The stock cooler and the huge cooler both came with CPU paste so purchasing the Arctic silver was unnecessary.

The speakers are under powered so I am sure that I will eventually replace them with something with a little more power. I dont know what that is going to be at this time, but my brain is working on it a little and its not that big of a deal to me right now. It will be addressed once I take him outside and the sound from the speakers isn't loud enough.

The 2 EZ-B's have been ordered from EZ-Robot and should be here early next week. Other stuff is on the way from England and the 3D printed parts are underway. I think the base and power are the next things to work on while I await these parts, but I really think I should wait to decide on the base until after I have the upper part of the robot build complete. Power however doesn't need to be attached to the base at this point so I think that is the next on the list to address.

#23

I am thinking about puting 5 EZ-B V4's in my inmoov. I have 5 of them and think it could offer some good benefits. Here is what I am thinking.

1 in each arm. This will reduce the number of wires that will be running to each arm and through the shoulder. The shoulder, elbow, wrist and hands would be driven off of these. Also, I am considering placing the v4 camera in the hand to help with picking up objects or moving the arm and hand to the object to grasp.

1 in the head used to drive the Eyes mainly. I would probably add other sensors and also drive some sort of lighting to try to change the color of the eyes.

1 in the chest used to drive the waist and neck servos, some sort of lighting in the middle of it and possibly interface via serial to an arduino for whatever I can figure out to do with it.

1 in the base. This one would be mainly used to drive the proxy sensors and the sabertooth/kangaroo combo that will handle the mobility. I would probably add more sensors here like gyros and the like.

I want to add a total of 4 or 5 v4 cameras. 2 in the hands, one high in the back of either the head and/or chest, and one in the base. There will also be 2 cameras connected via usb to the eyes and 2 usb cameras in the front of the chest above the screen.

The computer will be onboard and will have a usb GPS, 4 usb cameras and speakers.

Am I just setting up far too much complexity with this layout to save on cables through the joints, or is this something that I should steam forward with? I know the scripting in EZ-Builder would be a bit more complicated, but not that much. Separating things out to various controllers (1 for each major body part) is logical but am I going to run into issues using the controls such as the auto position control? I might just have to script all of the movements I guess. I dont know. What are your thoughts?

#24

d.cochran, I use this method in my B9. At one time I had 3 EZB's but down scaled to 2 for now. I'll probably end up with one or two more. It really cuts down on the wires running between moving parts. They all work so nicely together.

EDIT: It's really not complex at all. You just have to remember what EZB is where, It's number and what it is hooked up to.

#25

Go for the complexity. Since you've got the EZB(4)controllers in hand layout what you can and build some scripts, make some tests and see how far you can go with your hardware. Remember benchmarking can be your friend in this endeavor.

#26

Thanks guys. Dave, I agree. It's easier in a way I guess. 0 in head, 1 in chest, 2 right arm, 3 left arm and 4 base. The more complicated part (I think) is that it pretty much eliminates the possibility of using the auto position control. Other than that. It won't be bad. In a way, it makes identifying what I am doing in scripts easier. D0 for thumb, d1 index, so on with two of the v4's. Very few servos on the other controllers. Knowing or remembering what port goes to what becomes much easier.

#27

Each arm including fingers / wrist / elbow / 3 servos for shoulder, total 10

Each arm will have 2 cables.
One individual power cable going to power supply
One HDMI cable ( 20 available connections) Connected to EZB at the back
One EZB controler should be enough to drive 2 arms

Still there are 10 extra connections on each arm available for sensors or other

The inspiration comes from Markus in Youtube


HDMI connection in Inmoov

User-inserted image

User-inserted image

#28

I was looking at using the HDMI cables but realized that if I could fit a v4 in the arm, this would eliminate the need for the hdmi cable. I love what he has done, but want to take it one step further Smile . I will probably use the hdmi breakout device he has used. I plan on running one power cable down the arm to a block in the arm. From there I would split the cables on the servos out to power them from the block along with powering the V4 that is in the arm from the block. I would take all of the servo signal cables and run them to the digital signal pins on the v4 that is in the arm. This would take the number of wires that are passing through the shoulder down to 2 (one positive and one negative power). This would be easier to hide in the shoulder.

The other option is for me just to power these 10 servos off of the v4. The only issue that I have with that is that the large servos especially could draw quite a few amps on startup.

I had the extra V4's laying here and thought that if I used them for this project, I could reduce the length of the wires which has its benefits in multiple areas. I am trying to see if there is room in the upper arm to house the V4 and the wires that would need to go to it. I realize that this will add some weight but really this is small if placed high enough in the shoulder. I would probably use the run of hdmi cable to pass through the elbow and provide power and signal to the hand servos and the rotational wrist servo.

This solution would make the inmoov more modular. I would also put a power connection at the shoulder so that the arm could be removed as one unit and tested as one unit. I would do the same with all of the other body parts making the inmoov effectively 5 separate robots working together through EZ-Builder, which can use up to 5 EZ-B's in the same project.

All of this does depend on if I can find a location to house the V4. If I need to, I will remove it from its case. The V4's have been modified to not use the mini deens connection for power.

#29

this one might be better for my use.

#30

I am thinking about adding 5 arduino mini's to this project, 1 for each EZ-B. I would use these to do things like control neopixel rings or strips for some cool lighting in different areas. I would also use the ones in the arms to add touch sensors to the fingers or hands of the InMoov. I don't have a lot of experience programming on Arduino but there is no time like the present to learn, right?