PDA

View Full Version : [Project] My next hexapod project, codename BMX



Zenta
12-27-2014, 03:18 PM
Hi,

I hope you all are having a nice Christmas!

For a long time I've been thinking about some new projects to work on. Since 2006 I've made a few more and some less successful robots, mostly hexapods but also an octopod, a biped and a quad. For me this hobby has been incredible fun and I've learned a lot from people like Kurt Eckhardt and Jeroen Janssen. Until now all my custom made robots has been RC servo based, mostly using the popular SSC-32 servo controller combined with the Basic Atom Pro 28 (BAP28) mcu.


I think I was pushing the limits (payload capacity and complexity) when using RC-servos on my last project MorpHex. Using some sort of serial bus controlled servos would certainly make the wire work a lot easier.


From now on I've decided to go dark.. Lately I've invested in several MX-106T and MX-64T servos.

5745

At first I'm going to make a rather simple 3 DOF hexapod. I'm thinking more "Back to the basics". I want a good platform for experimenting on new code, terrain adaptation and learning new stuff, especially C++ and the Teensy.. My plan is to make a rather light weight hexapod with large leg segments and a relative small body. Since the cost is so high I'll probably reuse the servos for several future projects.

When I'm done with the hardware I'll study what others like KevinO and KurtE has done. I probably need to pick up some books about kinematics again.

Until now I've not used any form of CAD when designing my robots. So I'm a total noob when it comes to that part. So far I've peeked at almost every free CAD solutions. At first I ended up with Autodesk 123D, but I find it a bit limiting and a bit frustrating to work on. I've also tried freeCAD but found it a bit overwhelming at first, but I might give it a second try. Any other suggestions for a good low cost CAD solution is appreciated.

Like I said I'm going for a simple design at first. I decided to go for the MX-106T for both femur and tibia. The MX-106T has a much better weight/torque ratio than the MX-64T. For only 26 gram additional weight (158 gram vs. 132 gram) you get a lot by upgrading from the 64's to the 106's. Here is a picture of how I plan to design the leg:

5746

The femur are going to be made of 4,76mm Delrin and the length is 18cm (servo center to servo center) and the tibia (3,15mm Delrin) is about 27cm (servo center to tars/foot). I've not completely designed the tars (foot) section yet. But it will be relative simple at first, I'll later replace it with a foot with sensors/switches. A while ago Tyberius (Andrew) was very kind to laser cut some test parts for me so that I could get the feeling of the Delrin plastic. My first attempt on the femur got very solid and rather heavy. After the test I removed more material and made it slightly slimmer. A Picture of my first attempt:
5747

My progress is very slow, due to life.. Little free time. But today I'm home alone, lol. Giving me time to finally make a post on the forum and making a simple and relative small body for the hexapod. I'm not finished with the body section yet. I'll probably add a smaller top section for covering electronics. The 5000 mAh 3S LiPo takes most of the space inside the body.

Some pics of the hexapod:
5748
5749
5750

In this picture I've just lifted the battery for illustrating the battery size vs body size:
5751

Again, I find Autodesk 123D terrible to work with when the model gets more complex.
My plan is to bend the tibia sections slightly closer at the foot section. I've no idea how to bend them in Autodesk 123D though, lol.

Let me know what you think. Any suggestions is highly appreciated!

(Btw, my blog is down. Hopefully I'll find a solution next year.)

KurtEck
12-27-2014, 04:49 PM
Looking great. Will write more later.

As for Blog. I believe my wife has one on wordpress.com, which I believe are free. You might take a look at their site.

Kurt

KevinO
12-28-2014, 12:59 AM
Can't wait to see the evolution of your project. You will get spoiled with the MX-64's and MX-106's. I thought the AX's were great but I was amazed with the MX line once I switched. You might want to try a lower deck for the battery. On my 64 hex I had the battery and the hubs on the lower deck so I could easily get to them without bothering the other electronics.


5752

KurtEck
12-28-2014, 09:40 AM
Hi again Kåre,

Glad to see you on the Dark side :lol: I too have lost interest in doing anything these days with RC servos and SSC-32 and will probably sell off a few of my Lynxmotion based robots... Sorry I don't want to pollute your thread.

I can not wait to see what you come up with, both Hardware wise but software wise as well.

Personally on the electronics side, I am not sure what I would suggest. As you know I really like the Teensy 3.1 and you have one of my breakout boards. It is great to see that Trossen now sell them (http://www.trossenrobotics.com/teensy-3-1) This is a great little platform, which I will continue to use. You can probably do most everything you need with this platform. This will be especially be true when he comes out with the more powerful version.

However, ever since Trossen mentioned that the new HR-OS1 (http://www.trossenrobotics.com/HR-OS1) will have an Intel Edison, which they are now selling (http://www.trossenrobotics.com/intel-edison-arduino and http://www.trossenrobotics.com/intel-edison-breakout ), of which I now have one of each of these setups (actually two with breakout but one is semi toast). These processors are nice in that they are dual core 500mhz (plus a 100mhz Atom which they are not making use of yet), floating point, wifi, BT, USB,... And they do support some IO (UARTS, SPI, I2C, GPIO), but IO performance wise the Teensy rocks. Both in speed of the IO pins as well as more real time control.

My PhantomX currently has the Arduino version of the Edison with XBee and 2.8" TFT shield and USB2AX to control servos. Over the next couple of days, I will post more on this on a different thread. (Playing with control console app for TFT display). Then I am thinking of maybe switching out the USB2AX to a Teensy, which I believe will be closer hardware wise to what Trossen will be doing (Arbotix Pro?). Not sure when I do that if I will also move Display to this one as well? Who knows maybe need to design new version of Teensy board specific to this. Or may try Arbotix Pro once I can get my hands on one and know more about it.

If you (or someone else) wish to give Edison a try, not sure which way suggest to do this. That is, I have a version or two of the Phoenix code bases running on the Edison, through different means.

The first way is using the Intel Arduino IDE. I have a version that works with the Arduino breakout board. It may also work on Mini-breakout. Problem is that Intel has only setup an Arduino board variant for the Arduino board. I have a version that I am playing with, with pin definitions for the mini...

Then their is doing the stuff completely native using the linux setup on the board. Gives you complete control, but a pretty big learning curve. Again have version for this (My Raspberry Pi project). Some day this will probably be converted to ROS and cmake files...

Currently I am doing stuff using the Eclipse IDE with the Intel stuff including cross compilers. You are still programming for the native Linux system, but you are doing all of your editing and the like directly on your PC (Windows, Mac, Linux). I am still adapting for the Eclipse IDE as I am used to others, so I am still trying to get automatic on what is the key stroke for find next... But, the IDE does a lot of nice things for you, like it scans the sources and knows what the members are of a class, so it detects if you do something wrong, can paste in function prototypes... But the most fun and useful thing, is it is setup to download your programs to the Edison using wifi, and you also have the ability to do Source level debugging again over wifi. This has probably a little bigger learning curve than using the Arduino IDE and Intel's documentation on how to do things is not the greatest, but if you (or anyone) is interested I will be uploading my Eclipse workspace for the PhantomX to github at some point.

Again sorry if the above is a little off topic. If you are interested, can continue this in different thread, or PM or...

Again hope you are having a great holiday season!
Kurt

jwatte
12-28-2014, 12:14 PM
That looks like a great project!

From Onyx, I learned that the femur probably doesn't need the cross braces (which add weight) because it's tight in both ends against the servos. The tars, however, needs some cross brace to fix the foot -- the blue pictures you have make me fear that the feet will bend the tibias out of true.

Another option to consider for controllers is the OpenCM-9.04. It's $10 ($15 if you don't have the connectors to solder in in the parts bin alread, $20 if you don't want to solder yourself) and it comes with libraries from Robotis to read/write the Dynamixels. I find the software stack to be slightly more robust for the Teensy, but the integration specifically with the TTL bus to be better on the OpenCM.

Zenta
12-28-2014, 01:55 PM
Thanks for your input guys!

I think I'll stick to Teensy at first and at a later stage move over to a more sophisticated platform(s)..
FP would certainly help though. Kurt, are you planning to make a new board for that, when it is available?

So far I think I'll try to fit the battery and electronics inside the body. By rotating the battery 90 degree there is much more extra space for the rest.

I've been playing with different ideas for the upper body plate. Using CAD really help visualizing.

5753

Another approach using a top deck. I've not added any spacers, just for visualizing.

5754

I'll play more before settling down to the final design.

Having fun!

Zenta
12-28-2014, 02:16 PM
That looks like a great project!

From Onyx, I learned that the femur probably doesn't need the cross braces (which add weight) because it's tight in both ends against the servos. The tars, however, needs some cross brace to fix the foot -- the blue pictures you have make me fear that the feet will bend the tibias out of true.


Thanks for your input Jwatte! I didn't see your post until after I had submitted my last post.
I do plan to make some holes for cross spacers on the tibias. From what I've experienced the extra support on the femur wouldn't hurt since the sections are rather long.

The tars section on the tibia isn't finished. At first the foot/tars will look something like this (upside down):

5755

You can see the slots that will fit this at the end of the tibia.

Zenta
12-28-2014, 02:36 PM
Like I said before I don't know what technique to use for bending the tibias on the simple CAD program.

For better illustrating I've just placed the tars part at the end of the tibia. As you can see I'm planning to bend the tibias to fit the tars bracket:
5756
5757

I'm also using two cross spacers for holding the tars bracket in place.

jwatte
12-28-2014, 09:45 PM
That looks like a well designed foot knob!

If you bend the tibias, won't they rub against the servo body in certain angles? Or are they flush to the servo flanges, and bend below/after that?

Btw: If you're really into CAD, I highly recommend Autodesk Inventor. It costs a fair bit if you buy it for professional use, but there may be other ways of getting it legally for hobby use, including:
- If you have a Tech Shop membership, they can install a 6 month copy for you, and you can refresh this eery 6 months.
- There is a 30 day free trial. Depending on your appetite for installing on fresh virtual machines, and creating new email accounts on autodesk.com, this can last you a long time. As long as you don't sell anything made this way, it's probably not on the dark side, maybe. (This is not legal advice)
- If you are writing a book, you may qualify for the author discount for ADN (their developer support.) I think this is a few hundred a year, and it includes almost all versions of almost all Autodesk software (3ds Max, Maya, Inventor, Mudbox, ...) You will be expected to actually publish something at some point, I think :-) These are non-commercial installs -- not for use for commercial work.
- If you are a developer, you can buy a single-seat ADN license, which has the same benefits as author's license. This is a yearly cost, but may be cheaper than the full price. Again, these are non-commercial copies, made available so that you can build plugins and integrate with Autodesk software. (They do accept plug-ins for internal-only tools, I think, so there's probably no publishing requirement here.)

I'm a TechShop member, AND we have a 5-seat ADN subscription at work because we ship 3ds Max exporters for our art path, so I have two ways into that path. Others may have to look at the other options. Even if you mainly just use the solid construction and drawing tools, Inventor works better than the free/share/cheap-ware I've tried. All the FEM simulation/structural and materials stuff is wasted on me :-)

KevinO
12-28-2014, 11:40 PM
Zenta,

In regards to feet on my 64 hex I went with these.



5758

http://www.bhphotovideo.com/bnh/controller/home?O=&sku=489566&gclid=CPvUv7C36sICFewF7AodAFMA-w&Q=&is=REG&A=details

They are tripod feet. They are all rubber with great impact absorption.

Zenta
12-29-2014, 05:28 AM
Hi,

Thanks for your advice about CAD Jwatte. Another option would be the student edition, only problem is that I'm not a student.. Do they really check that?

I don't think bending the tibia would cause problem to the servo, like you said they are flush to the servo flanges, and will bend below/after that.

Thanks for you tip about the foot Kevin. I would really like to see a perfect round foot though. Looking at other options on your link I found this one looking very interesting:

5759
I wonder what the dimensions are for this one?

Zenta
12-29-2014, 08:28 AM
Anyway, I think I'll try the bumper I got from DigiKey (http://www.digikey.no/product-search/en?KeyWords=SJ5017-0-ND&WT.z_header=search_go) some months ago.
5760
My plan is to stick them onto the top of the tars section that contains of these parts. I'm also using a long 3mm screw to hold the parts together.
5761

KurtEck
12-30-2014, 05:55 PM
I think I'll stick to Teensy at first and at a later stage move over to a more sophisticated platform(s)..
FP would certainly help though. Kurt, are you planning to make a new board for that, when it is available?


I am not sure if I fully answered this yet. I am planning to, my guess is it will be a killer little processor... Extracts from the teensy thread on what will probbly be in this version:



My current plan for Teensy++ 3.1 has 48 "outside" pins, where the 28 on the left side (the USB side) very closely mimic Teensy 3.1's pinout. The idea is you could put the new board into a project originally using Teensy 3.1, assuming there's room for the extra 1 inch of PCB to hang over. Or if you design something with Teensy++ 3.1 and later want to step down to a less expensive board, the others will be mostly pin compatible.


First of all, the low cost board is coming first, based on a Kinetis L-series chip. I know many of you want higher performance hardware, and that will happen, but I feel it's important to get a solid product out that really allows good 32 bit Arduino compatibility at approx $12 retail.

Teensy++ 3.1 will definitely not be released this year. The simple truth is I've spent most of this year on software. Hopefully a lot of good has come from all that effort, but it has pushed back plans for new hardware.

Until recently, I had planned on using the K22 chip. The K22 very well may still be used. In fact, I have a (not yet working) K22-based prototype on my desk right now.

However, I'm seriously considering delaying Teensy++ 3.x, for a new chip Freescale will be releasing next year. Obviously I can't talk about anything that's under NDA. But I can tell you I have only minimal technical info (not a datasheet or reference manual), but what info I do have says it will be very well worth the wait!

Freescale and ARM have already made press releases about Cortex-M7 coming soon. So far, I don't have pretty much any technical info about it, other than what's publicly available. But it certainly looks like it's going to be awesome... and perhaps worth delaying a Teensy++ 3.x board?

Many of the M7's features are optional, much like M4. But I'm really hoping for a FPU that supports double.

I'm *really* excited by some of the other optional features, though it may be a few years until we see semiconductor manufacturers using 20-some nm silicon processer for microcontrollers (eg, clock speeds approaching 1 GHz). In the near term, we're probably going to get only 200 to 400 MHz speeds.

So as I mentioned, this for sure sounds like something I will probably want to play with ;)

jwatte
12-31-2014, 03:32 PM
A Teensy with hardware FPU (even for 32-bit floats) would be the awesome!

For the $10 price point, it's hard to beat the OpenCM 9.04...

Zenta
01-05-2015, 11:33 AM
Thanks for posting the information Kurt, it sure sounds worth waiting for the next Teensy++ 3.x.

I think I'm more or less finished designing the first MX hexapod. I've played with different styles for the body top plate.

5774
I might go for B or C. Any suggestions would be appreciated.:)

KevinO
01-05-2015, 12:06 PM
I'd pick C. One thing I regret on the 64 Hex is that I didn't plan on any additions so any new cables have to be run around the edge to the sensors on the top. Having those gaps to run wire might come in handy.

Also if I can make a suggestion.. I wouldn't wrap all the way around the servo. I've had to remove a leg or two and I really liked I could slide the leg out without removing the top or bottom. Here is a pic from way back when of some of the body parts so you can see what I mean.

5775

KevinO
01-05-2015, 12:09 PM
In case you are curious Length is 30cm and width is 10cm.

jwatte
01-05-2015, 12:13 PM
I wouldn't wrap all the way around the servo.

+ infinity!

Having to remove the plate from ALL the servos just to get at ONE of the servos is a right pain. I made that mistake on an early iteration of Onyx, and I quickly learned from that experience.

KurtEck
01-05-2015, 12:28 PM
Your designs look nice.

Note: with the current PhantomX hexapod plates, I have found the new optional top plate to be handy as it allows me to easily add the battery below it (and remove it after I kill it ;) )

As for the new Teensy and carrier board for it, will have to wait and see what the holes and like are on it and also what features I/we want on it. Things like, build in more advanced sound stuff like his add on board or stay with simple built in speaker. AX Buss- yes, How many 3 pin connections... If I want to use for IO board connected up to main processor. Use USB to communicate? Or something else (SPI, I2C, ...). XBee?

Will be fun.

Zenta
01-05-2015, 12:49 PM
I'd pick C. One thing I regret on the 64 Hex is that I didn't plan on any additions so any new cables have to be run around the edge to the sensors on the top. Having those gaps to run wire might come in handy.

Also if I can make a suggestion.. I wouldn't wrap all the way around the servo. I've had to remove a leg or two and I really liked I could slide the leg out without removing the top or bottom.


Awesome, thanks for your input guys!

I suppose you mean something like this Kevin:
5776

As for replacing the battery, I'm planning a similar solution as I did for A-Pod.
5777

KevinO
01-05-2015, 01:01 PM
Yep! That way if you have a issue with a leg you just unscrew and slide it right out.

Zenta
01-05-2015, 03:43 PM
Hi,

I believe the body sections are (almost?) ready now. I'll probably make an optional top plate for more electronics.
5778

I'm planning to place the Teensy breakoutboard inside the body mounted upside down on the top plate.
5779
5780

Kurt, did you make a 3D file of the board? I'll check your thread. It's not necessary, just for fun.

5781

Trying to visualize the LiPo.

KurtEck
01-05-2015, 04:16 PM
Kurt, did you make a 3D file of the board? I'll check your thread. It's not necessary, just for fun.


Not sure what you need. I have done pictures showing the 3d output. But for the fun of it I output a couple different formats in case one will work for you:

KevinO
01-05-2015, 04:21 PM
Big fan of top plates... well heck I'm a big fan of bottom plates as well. :) I was astonished how quickly room was taken up once wires and sensors were added on mine.

Zenta
01-06-2015, 02:54 AM
Thanks Kurt, I was able to use the step file.

blegas78
01-10-2015, 03:28 PM
Wow! Can't wait to see this move around! While you're moving in the direction to make one bigger, I've been making a smaller hexapod.

Zenta
01-11-2015, 04:30 AM
Hi,

Is that you Matt Bunting? Welcome to the forum! I really admire your work! Just saw a video on your channel https://www.youtube.com/watch?v=JyxFMALAf9A "Can they jump?" sigh.. Well, I do get questions if my MorpHex can fly, lol.
Nice hexapods! Using the XL-320? I've also thought of using them in a project. What payload capacity and overall weight does your hexapods have?

Yeah, I'm going for a bigger version this time, but compared to Matt Denton's Mantis its pretty small..

KevinO
01-11-2015, 02:17 PM
I'm pretty sure that's Matt. :) Welcome to the forums Matt!

blegas78
01-11-2015, 03:20 PM
Hi,

Is that you Matt Bunting? Welcome to the forum! I really admire your work! Just saw a video on your channel https://www.youtube.com/watch?v=JyxFMALAf9A "Can they jump?" sigh.. Well, I do get questions if my MorpHex can fly, lol.
Nice hexapods! Using the XL-320? I've also thought of using them in a project. What payload capacity and overall weight does your hexapods have?

Yeah, I'm going for a bigger version this time, but compared to Matt Denton's Mantis its pretty small..
It's me :)
Fortunately I suppressed the "Can you use a battery?" question I always used to get. For some reason 5% of the people I show it to thinks that it can fly.

Yeah I'm using the XL-320. The tibias are much too long in my design, but that's what the fashion designer requested. I'll be making a smaller version and make sure that they walk well thought he life of the battery. I'm not sure about payload capacity or weight at this time, still a work in progress.

Have you calculated the expected payload of BMX?


I'm pretty sure that's Matt. :) Welcome to the forums Matt!
Thanks!

Zenta
01-12-2015, 02:05 PM
Did you post some info about your XL-320 hexapods?

I'm trying to keep the total weight as low as possible, so I don't think there will be much payload capacity, barely enough for electronics and battery. Currently I'm aiming for a total weight of 5,5 kg. If its going higher than that I'll probably need to adjust some dimensions. For this project my goal was to have as long/large leg segments as possible.