The Manta Ray – Part 2

This time we’ll wrap up the build of this flexible robot platform by finalizing the build to a point where we can experiment with some code and build off from that. In part 1 we attached several accessories to show some of the many options that are already available. In this part we’ll make some adjustments to round out the build.

Round Two! Build!

One of the first things we have to do is decide how we’re going to drive the motors. In Part 1 we showed how you could use the HB-25 from Parallax. And if you’re fairly new to robotics this is an ideal solution since the HB-25 allows you to control a brushed DC motor with the ease of a continuous rotation servo. This is ideal for the BASIC Stamp 2, Arduino or other microcontrollers. However I have chosen to use the Parallax Propeller chip and so the HB-25 is not needed since the Propeller can generate its own PWM to drive the motors. All we really need is a driver or dual H-Bridge to handle the voltage and current the motors require.

For this we’ll use the Pololu Dual Motor Driver available from Parallax as #28820. This board is a dual H-Bridge using the Freescale MC33926 IC. It can supply up to 3A of continuous current per channel from 5-28VDC and PWM frequencies up to 20kHz, though we’ll be using 10kHz. We can also sense current draw via a feedback pin on the board which provides a voltage proportional to current draw. This means a simple ADC can be used for this purpose. This allows a measure of safety as we can provide software stall protection as well.

In order to connect the motors to this driver board we’ll need to make the motors more accessible and easier to connect. Currently the motor wires terminate to a 6-pin SIP socket as shown in Figure 1. In many applications this connector won’t be useful as is the case here since the motor connections (red and black) need to connect to the motor driver board on the lower deck while the other 4 wires need to go to the upper deck to our controller. In this case a Propeller Project Board. There is also some heat-shrink tubing holding the wires together which prevents us from separating the bundle.

First we’ll very carefully remove the heat shrink tubing using a hobby knife, and then we’ll remove all 6 wires from the plastic shroud. Notice the arrow pointing to the connector housing. Each connector inside is held in place by a small tab (Figure 2). Lifting this tab very carefully allows you to pull the crimped connector from the housing. I used a surplus dental pick I have in my tool drawer. A very small precision flathead screwdriver, like those used on glasses could also be used. Don’t put tension on the wire until the tab is up to make it easier.

Once the heat-shrink has been removed and the wires are free from the connector housing you can run them to their respective places (Figure 3). The Blue, Green, White and Yellow wires will all run up through the hole in the upper deck. These wires include the power for the encoders as well as the A & B outputs from the encoder.

Looking at Figure 4a, the red and black wires will go to the driver board, but first we need to make a small adjustment. The connectors on these wires won’t fit into the terminal blocks on the driver board.

I cut the connector off after the crimp but before the female connector itself, essentially turning the connector into a tiny spade connector (see Figure 4b).

Now the motor wires can be inserted into the screw terminal blocks on the motor driver board as shown in Figure 5.

Other Changes

One other change I wanted to make was to isolate the motor power and the control board power. This is desirable if there exists the potential for the motors to draw enough current to cause a sag in the supply bad enough to cause the controller or sensors to fail or to introduce noise in the system.

So a second Li-ion power pack was added and I tried to add both to the lower deck as shown in Figure 6. This fit, but unfortunately it blocked the hole for the wires to come up through the upper deck. This means that for now I will move the second supply to the upper deck, however it is now a possibility that I will move the hole for the wires on my board to allow me to still do this. As you can see the design grows and changes as we build something. Each of us should see the changes we wish to make and feel free to make them to customize things to our own tastes and needs.

For now we’ll mount the power pack in the following manner (Figure 7). We can always change it later.

Sensor Wiring

The PING))) sensors remain as they’re somewhat an integral part of this robot design. The only thing I may change at some point is that the current PING))) Protector Stands are aluminum, but the chassis is a smoke translucent acrylic. There are acrylic PING))) stands available from Parallax as #725-32008 for just the basic stand and #725-28998 which adds an IR sensor. I haven’t yet checked height of these stands but they’re now going to become future possible options for my robot.

Okay, lets get the current PING))) sensors wired up (Figure 8). For this I used three (3) 10″ 3-pin F-F cables. Be sure of the polarity of the connections on the pins of the PING))) sensor. When looking at them from behind the pins are in the order signal (white), power (red) and ground (black). It can be easy to get them confused when turned around, especially when mounted in the protector stands because you can no longer see the pin designators.

At this point we can connect the power from the screw terminals on the power pack into the motor driver board. Be sure of the polarity. At this point no option to switch power for the motors has been implemented. If you connect a 2.1 mm barrel plug into the charging jack it will cut power. So this is an additional consideration for revising the design.

We can now run the PING))) sensor cables up through the hole in the upper deck as well as the control wires for the motor driver board (Figure 9). Once all the signal wires are run we can reconnect the upper and lower decks.

On the Propeller Project Board several 3-pin headers will need to be installed to facilitate connecting the sensors. In Figure 10 you can see the connections for the wheel encoders as well as the motor driver signals. The PING))) sensors operate at 5V and therefore will need additional headers that provide 5V.

Finally we need to add the second power pack. This is the one that will power the control board. The Propeller Project Board does not have a DC barrel jack, so we’ll actually connect wires from the screw terminals to the V_IN tabs on the project board. This is another point where a power switch is needed.

In Figure 11 the second power pack is shown upside down (and reversed) to minimize the overall profile of the robot.

In Figure 12 it is upright and the terminal blocks are forward toward the power connections on the project board. I’m still not sure if I like these options so the next revision may see both supplies moved to the lower deck. For now this will work.

Final Thoughts

There are some issues in the final build on this version of the Manta Ray DXF file. No provisions were made for a power switch. We also have two isolated power supplies that should be switched on at the same time. I will look for a DPST or DPDT switch that can mount easily to the Manta Ray upper deck and have a mounting hole added to the DXF file. The motor driver board has extremely small mounting holes and I surprisingly did not have mounting hardware for it. You could use double-sided sticky-tape or Velcro to attach it to the lower deck but I will be actively seeking hardware and add proper mounting holes for that board as well. In this build I secured it to one of the rear standoffs.

THIS ARTICLE IS NOT COMPLETELY RESTORED! Why?

2-1-2024 – Although I was able to recover much of the data / images from the original published article, I was not able to recover the source code or updated DXF file. All is not lost though. This article was very important to me and what I intend to do is to recreate the robot from scratch. The caveat being that there will be many changes from the original because since this article was originally published, Parallax discontinued pretty much everything that was used in it. So I will need to source new parts. When I do I will create a link an update article. Please check back later.

Resources

The Manta Ray – Part 1

Discuss this project on Savage///Chats

This project was published in the November 2016 issue of Servo Magazine