The "Platypus" my 3...
 
Notifications
Clear all

The "Platypus" my 3d printable chassis.

9 Posts
3 Users
4 Likes
827 Views
triform
(@triform)
Member
Joined: 5 years ago
Posts: 324
Topic starter  

This is a robot chassis I made to use for an indoor bot. This is the continuation of a failed project I had several (5) years past.

One of the smaller issues before was the chassis I had for the bot.  It was homemade from PVC sheets, a few 3d printed parts with bolts for platform standoffs/support. It had an RPI as the main controller and nine Pro minis on an I2c bus. One Pro mini was a master connected to the RPI via i2c and via one GPIO pin going to a homemade multiplexer. It figured out the IRQ requestor and then told the RPI via a GPIO that had a daemon watching it and would make a request to the master for the node information. All of that works okay for the most part, but the bot was plagued with noise and power issues. In the end, I was able to address the power issues but the I2C bus was still a problem. One of the first problems I started to address about three years ago was the chassis.  I sat down and made the Platypus in Tinkercad. I wanted something that I could add to and change easily, which was small in weight and had a fairly large usable area. It also had to be simple to assemble and print.   

Scott

 


   
Duce robot and Spyder reacted
Quote
Spyder
(@spyder)
Member
Joined: 5 years ago
Posts: 846
 

@triform

Nice !

Elegant and simple, and uncluttered. I should take lessons from you

I'm currently having a problem with my own openwrt running on a pi3, so I looked into seedstudio for the 7688, and there are 2 models... both out of stock of course. I would also need their breakout board (I think). I found the same item on amazon... for $40 !!! and their same breakout board is $30 whereas the same parts on seeed are only $15 each... wow

Would you know if the extra microUSB port could be used for a micro USB to ethernet connector ? Cuz this item might solve my problem (maybe)


   
Duce robot reacted
ReplyQuote
triform
(@triform)
Member
Joined: 5 years ago
Posts: 324
Topic starter  

@spyder

Thanks!

I have my Linkkit on a Naan board.  One USB is power only and the other is host.  both or OTG ports. You can find them on ebay now here for 10$+ 2.50$ shipping. https://www.ebay.com/itm/Seeed-Studio-Linkit-Smart-7688-Brand-New-USA-Seller/192956045706?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

If you want to make a mount PCB, send me a message.

 

Scott

IMG 20200802 210341

 

 


   
Spyder reacted
ReplyQuote
Spyder
(@spyder)
Member
Joined: 5 years ago
Posts: 846
 

I just ordered 2 from that ebay link. Thank You (just in case I break one of them... I do that from time to time)

Now, when you say the other is "host", what exactly did you mean ? Host, as in host/client ? I've got a few OTG-ethernet dongle thingies, would one of those work, or would I need  a breakout board of some sort ?


   
ReplyQuote
triform
(@triform)
Member
Joined: 5 years ago
Posts: 324
Topic starter  

I do the buy extra thing as well 🙂

I would assume it would work with at least some USB devices. I would give them a try. It does have Wifi built-in and I run an external antenna on mine. I have the board made in the pic above with a 5v LDO reg board and pins for the GPIO's. I also have more python modules installed on mine.  I'm planning to add more space via an SD card and will do it on a new one since I have the current one working. 


   
ReplyQuote
Spyder
(@spyder)
Member
Joined: 5 years ago
Posts: 846
 

I see the antenna. And I have a pair of extra antennas laying around. Did that antenna connector come already on the board, or did you attach it yourself ? It looks like it's part of the board, but I can't really tell

All I'd really need it for is the openwrt. I'm about to do something really stupid to try to untether my bot from the ethernet. No idea how this experiment is going to turn out

I figure it's a 50/50 shot... It'll either work, or it won't


   
ReplyQuote
triform
(@triform)
Member
Joined: 5 years ago
Posts: 324
Topic starter  

The connector cam on the board.  I just added the cable and a rubber-ducky antenna.  


   
ReplyQuote
byron
(@byron)
No Title
Joined: 5 years ago
Posts: 1112
 

@triform - I see you are on youtube, now I will be following your progress of your satisfyingly large indoor bot.  Very interesting indeed.   

I perused your site to see if you had documented any of your previous forays with your gps robots.  (nothing there 😥). Although its all just plans at the moment, though there has been some testing on a small bot, I have more or less convinced myself that for my eventual outdoor rugged terrain robot I would need centimeter accurate GPS with a couple of rtk gps boards  Before I go to that expense I will see how I get on with just a normal GPS board.  I think a normal GPS may be OK for roving about my field, but I think I would need some more precise navigation aids to negotiate the areas closer to the house. 

Thanks for sharing your platypus robot progress. 

This post was modified 4 years ago by byron

   
ReplyQuote
triform
(@triform)
Member
Joined: 5 years ago
Posts: 324
Topic starter  

@byron

Sorry, with my first GPS bot (2010/2011) I did not have a camera on a phone. I was using a laptop and a wifi router on the bot. I had a socket connection to the bots computer that was a hacked up notebook with the uC on USB. The GPS module was one I got out of an older GPS that had a serial GPS unit in it that I got to work.  I did the math to calculate the distance between two points and drove in the compass direction adjusting every so many seconds (30 I think it was). It had an older compass on it that was analog (I forget the part) I did the calc, took a compass reading, and adjusted the direction to match and drove forward.  It had to be in a clear area with no obstacles or it would get stuck.  There was no shortest path finding or the like. 

In 2014 I did the second. It had an RPI and Arduino. A U/Box GPS mod and an i2c compass. It also had sensors to detect obstacles and subsumption logic to get around them.  The task priority was GPS driving until an object was detected and then that was driven past or around then re-read GPS/COMPASS and go again.  I cheated and did not do any maths at that time. It drove until the bot got into a dead type zone cost it could get.  I had some figuring calcs for that area. It actually worked ok, but could get stuck in say a parking lot with a lot of cars. I know I didn't do this the best way at all and I was going for "make it work".  The subsumption logic worked out well for getting around most objects.   There was no XYZ adjustment on the compass reading either.  I was quite lazy that time out. 🙂

 


   
ReplyQuote