I'll rephrase my statement. The Penguin mastered two servo walking "in my opinion". Running with two feet off the ground? I wish. I attempted a race against a Boe-Bot across my desk and the results were impressive even though the Boe-Bot clearly won by over 30%. Penguin was tetering on the brink of tipping over the whole time.
By the way, do you walk with two feet off the ground (or only run that way)?
I wouldn't rule out dynamic walking with hobby servos. Standard servos might be slow, but high-speed metal gears with a spring-loaded leg might be able to bounce along quite nicely. Especially if the power supply were lightweight.
I'm really impressed with the power-to-weight ratio achieved by some of these new super-cap toys (mainly the R/C cars and planes). My little boys and I went tonight to fly two of those Harbor Freight free-flight airplanes at the local college field. After charging the plane for 45 seconds with two AAs, it flew at least to 600' high and maybe a quarter to a half mile away before it went out of sight. It literally disappeared from view·it went so high. I couldn't imagine such a $6.99 airplane existing 15 years ago.
I don't run with two feet off the ground, but the running gait for humans do include a period time in which both feet are off the ground.
Granted, a dynamic gait with a two-legged robot is pretty hard, much harder than with four legs. A really simple lego bot does trotting with 4 legs.
I'd like to share with you the latest hardware revision on the Penguin robots. The latest design incorporates several design improvements to the mechanics. Improved precision, a new foot design and better integration with the BASIC Stamp system are some of the changes.
In this picture you'll see the prior prototype (the clear aluminum one in back with the electronics) along with three new bodies. The parts are either nickel-plated or anodized red, black or blue. These new versions are not wearing any electronics yet.
Meanwhile, we're working on a new PCB design which includes a USB mini-B (with a brand new FTDI chip, not even available to our customers yet!) and an RF module for PC-based control.
Of interest is that everybody who sees these small robots assume they're less expensive to make than their larger counterpart, the Toddler. Actually, as a design gets smaller and more precise the cost of parts increases significantly. Machining isn't the only big expense - cost goes up with small servos, stainless steel hardware, mixed SMT/through-hole·boards, and components on both sides.·The cost of building these is almost twice that of a Toddler. Ouch.
Ken Gracey
Parallax, Inc.
Post Edited (Ken Gracey (Parallax)) : 9/3/2005 4:11:27 AM GMT
Any time frame on these going to market? I assume that the development has progressed far enough that this will be a viable product line for Parallax. I know I will be the first in line to buy one. *grin* Can I pre-order now? My credit card # is..... (ok I may be a little too exited).
These will be availabe in November for purchase. We just launched production for the robot bodies and hardware and the electronics are in their last prototype.
It's a limited run, too. Only 50 nickel-plated robots, 50 anodized blue and 50 anodized red. No target price yet, but under $250.
Ken, those are awsome! I can't say for sure I'll get one to keep my Gold Toddler and Boe company but they are still awsome.
Ken Gracey said...
Actually, as a design gets smaller and more precise the cost of parts increases significantly. Machining isn't the only big expense - cost goes up with small servos, stainless steel hardware, mixed SMT/through-hole boards, and components on both sides. The cost of building these is almost twice that of a Toddler. Ouch.
I think most of the folks here understand that all too well
I need to answer a couple of questions. I'm using aluminum because I can machine it to exact tolerances, it looks good and can be anodized a variety of colors. This is a high-quality [noparse][[/noparse]low-volume] desktop robot, so plastic was not considered as a material. Once you get a chance to hold a Penguin you'll see how the aluminum provides the perfect mass of well-balanced material.
The update is that the Penguin hardware is in production at a local machine shop. The electronics are in a final prototype and have not gone to production. The new PCB includes an RF receiver for PC control, a USB port for programming, HM55B compass, speaker, infrared, Ping))) socket, 7-segment blue LED, photoresistors and a power switch. We're still aiming for mid-November, but no promises.
Has the price for the Penguin been refined any further than "under $250", I understand it would still be an estimated price, but maybe if I start saving now, I can afford one when it becomes availible .
We'll be selling it somewhere around the cost of materials and assembly. This should be about $250. Okay, we'll probably make fifty bucks. It's a special robot that we are making to provide as a gift to key distributors, but it will also be for sale in limited quantities (don't believe me? You will when we're out of stock [noparse]:)[/noparse]. It will be 100% manufactured in the USA, with the best QA/QC we have available. I consider it an experiment in quality, precision hobby robotics.
There is one business benefit of low-volume manufacturing in the USA. Machined parts can be made in low quantities, without tooling. Minimums can be reduced with high-end products. Inventory levels can be much lower. And considering we're currently swimming in a pool of continuous rotation servos this is an important point at Parallax.
Now you will want to become a distributor to get a free one.
Thanks Forrest - just made the fix. For some reason I always get the years wrong. That's because they go by really fast, and each subsequent year of one's life is a smaller proportion of their total life.
The large bipeds were made using all manual tools. This includes a Bridgeport-style mill in our office, a shear, a brake and sander.
The small Penguin was made using a Wabeco F1210 CNC desktop mill. I use FlashCut signal generator (Mach2 with Gecko drives is more popular, though) and CAM work was done with OneCNC. I also used a Wabeco D6000 CNC lathe to turn a small part you can't see, which is the stride linkage between the legs. The specific part is 1" long and has a ball on each end which fits within the leg slot so the bottom servo can be used for stride control. Additionally, this part (which is Delrin) was drilled and broached with the shape of the servo spline. To broach it, I made a tool with a sideways-mounted rotary table. Using a small V-cutter, I machined a drill blank rod with 21 cuts to match the servo spline. Then, this tool was used to punch the shape into the Delrin.
The other tools I used include end mills, drills, and lots of coolant. I also used a benchtop sander to clean up any tabs which I left on the aluminum parts from CNC (to avoid making fixtures for the parts).
The CNC tools·I used are among the more expensive for desktop machines, but·a Sherline machine could certainly do the same. A Taig mill could also make the parts. In fact, you could make them with a bandsaw and drill if you were careful (especially if you used acrylic).
The drawings are posted, so·you're welcome to try to make one.
I'm just curious...I know that you said that the penguins go on sale in Q1 2006, but is there any chance that we could get a more detailed release date? Maybe like what month it looks like you'll start selling them? And if you can't release that, maybe some more photos of the penguins...so we could see more detail?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Truth can be dangerous... Trust can be deadly.
I received my Servo magazine and just had a chance to download the drawings for Penguin. I didn’t realize it functioned with so few parts. In this age of bloated hardware and software its good to see such simple elegance. Any “engineer” can design something to be complex. But it takes real inspiration and commitment to make something simple.
It got me to thinking about the historic designs that I consider elegant in terms of “form following function.” Among my favorites are
The fiberglass sailplane
The British regulator pendulum clock
Browning’s design for the M1911 pistol
The software/hardware designs used in the Apple II computer
The only part not shown is the 5/8" machined standoff which holds the servos in place. It's nothing more than a 4/40 5/8" long round aluminum standoff with the middle milled flat. Then, tap a 4/40 hole in the middle of the flat spot and this will be used to mount your servo.
The servos mount on two 5/8" round standoffs to the 1/8" holes on the Penguin body. What kind of dimensions are you looking for? The DXFs should provide the exact detail you need, from which you are free to extract dimensions.
Jimmio, pretty soon. We're just finalizing the PCB. As this is not a revenue-generating product we've not been in much of a rush. We're only making 1,000 units and you'll never see it again after that. The reasoning is that the product is entirely USA manufactured and machined, so the cost is truly through the roof.
Comments
I'll rephrase my statement. The Penguin mastered two servo walking "in my opinion". Running with two feet off the ground? I wish. I attempted a race against a Boe-Bot across my desk and the results were impressive even though the Boe-Bot clearly won by over 30%. Penguin was tetering on the brink of tipping over the whole time.
By the way, do you walk with two feet off the ground (or only run that way)?
I wouldn't rule out dynamic walking with hobby servos. Standard servos might be slow, but high-speed metal gears with a spring-loaded leg might be able to bounce along quite nicely. Especially if the power supply were lightweight.
I'm really impressed with the power-to-weight ratio achieved by some of these new super-cap toys (mainly the R/C cars and planes). My little boys and I went tonight to fly two of those Harbor Freight free-flight airplanes at the local college field. After charging the plane for 45 seconds with two AAs, it flew at least to 600' high and maybe a quarter to a half mile away before it went out of sight. It literally disappeared from view·it went so high. I couldn't imagine such a $6.99 airplane existing 15 years ago.
We lost the airplane, by the way.
Ken Gracey
Parallax, Inc.
Granted, a dynamic gait with a two-legged robot is pretty hard, much harder than with four legs. A really simple lego bot does trotting with 4 legs.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
It's me.
I'd like to share with you the latest hardware revision on the Penguin robots. The latest design incorporates several design improvements to the mechanics. Improved precision, a new foot design and better integration with the BASIC Stamp system are some of the changes.
In this picture you'll see the prior prototype (the clear aluminum one in back with the electronics) along with three new bodies. The parts are either nickel-plated or anodized red, black or blue. These new versions are not wearing any electronics yet.
Meanwhile, we're working on a new PCB design which includes a USB mini-B (with a brand new FTDI chip, not even available to our customers yet!) and an RF module for PC-based control.
Of interest is that everybody who sees these small robots assume they're less expensive to make than their larger counterpart, the Toddler. Actually, as a design gets smaller and more precise the cost of parts increases significantly. Machining isn't the only big expense - cost goes up with small servos, stainless steel hardware, mixed SMT/through-hole·boards, and components on both sides.·The cost of building these is almost twice that of a Toddler. Ouch.
Ken Gracey
Parallax, Inc.
Post Edited (Ken Gracey (Parallax)) : 9/3/2005 4:11:27 AM GMT
Any time frame on these going to market? I assume that the development has progressed far enough that this will be a viable product line for Parallax. I know I will be the first in line to buy one. *grin* Can I pre-order now? My credit card # is..... (ok I may be a little too exited).
Doc
These will be availabe in November for purchase. We just launched production for the robot bodies and hardware and the electronics are in their last prototype.
It's a limited run, too. Only 50 nickel-plated robots, 50 anodized blue and 50 anodized red. No target price yet, but under $250.
Ken Gracey
Parallax, Inc.
I think most of the folks here understand that all too well
Cool bots!
Thanks,
Sean
I need to answer a couple of questions. I'm using aluminum because I can machine it to exact tolerances, it looks good and can be anodized a variety of colors. This is a high-quality [noparse][[/noparse]low-volume] desktop robot, so plastic was not considered as a material. Once you get a chance to hold a Penguin you'll see how the aluminum provides the perfect mass of well-balanced material.
The update is that the Penguin hardware is in production at a local machine shop. The electronics are in a final prototype and have not gone to production. The new PCB includes an RF receiver for PC control, a USB port for programming, HM55B compass, speaker, infrared, Ping))) socket, 7-segment blue LED, photoresistors and a power switch. We're still aiming for mid-November, but no promises.
Ken Gracey
Parallax, Inc.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·1+1=10
We'll be selling it somewhere around the cost of materials and assembly. This should be about $250. Okay, we'll probably make fifty bucks. It's a special robot that we are making to provide as a gift to key distributors, but it will also be for sale in limited quantities (don't believe me? You will when we're out of stock [noparse]:)[/noparse]. It will be 100% manufactured in the USA, with the best QA/QC we have available. I consider it an experiment in quality, precision hobby robotics.
There is one business benefit of low-volume manufacturing in the USA. Machined parts can be made in low quantities, without tooling. Minimums can be reduced with high-end products. Inventory levels can be much lower. And considering we're currently swimming in a pool of continuous rotation servos this is an important point at Parallax.
Now you will want to become a distributor to get a free one.
- Ken
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
There's nothing a good wack with a hammer won't fix
Darn I let the white smoke out again
Post Edited (Washer Medic) : 10/6/2005 3:56:17 AM GMT
Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Thanks
Bill
·· Up about 5 replies from yours Ken posted an anticipated time frame that we're not making promises on.·
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
Thanks Chris,
Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Thanks
Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
The Penguin is on the cover of Servo Magazine this month:
http://www.parallax.com/html_pages/robotics/penguin.asp
Ken Gracey
Parallax, Inc.
Shouldn't that be Q1 2006?
- Ken
Can you post a listing of the development tools you used, ie lathes, mills, CAD, etc. Just curious. Thanks.
The large bipeds were made using all manual tools. This includes a Bridgeport-style mill in our office, a shear, a brake and sander.
The small Penguin was made using a Wabeco F1210 CNC desktop mill. I use FlashCut signal generator (Mach2 with Gecko drives is more popular, though) and CAM work was done with OneCNC. I also used a Wabeco D6000 CNC lathe to turn a small part you can't see, which is the stride linkage between the legs. The specific part is 1" long and has a ball on each end which fits within the leg slot so the bottom servo can be used for stride control. Additionally, this part (which is Delrin) was drilled and broached with the shape of the servo spline. To broach it, I made a tool with a sideways-mounted rotary table. Using a small V-cutter, I machined a drill blank rod with 21 cuts to match the servo spline. Then, this tool was used to punch the shape into the Delrin.
The other tools I used include end mills, drills, and lots of coolant. I also used a benchtop sander to clean up any tabs which I left on the aluminum parts from CNC (to avoid making fixtures for the parts).
The CNC tools·I used are among the more expensive for desktop machines, but·a Sherline machine could certainly do the same. A Taig mill could also make the parts. In fact, you could make them with a bandsaw and drill if you were careful (especially if you used acrylic).
The drawings are posted, so·you're welcome to try to make one.
Ken Gracey
Parallax, Inc.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Truth can be dangerous... Trust can be deadly.
It got me to thinking about the historic designs that I consider elegant in terms of “form following function.” Among my favorites are
The fiberglass sailplane
The British regulator pendulum clock
Browning’s design for the M1911 pistol
The software/hardware designs used in the Apple II computer
I'll add Penguin to the list.
·
The only part not shown is the 5/8" machined standoff which holds the servos in place. It's nothing more than a 4/40 5/8" long round aluminum standoff with the middle milled flat. Then, tap a 4/40 hole in the middle of the flat spot and this will be used to mount your servo.
I'll make a drawing of that part next.
- Ken
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Info on all kinds of robosports:
www.freewebs.com/robotbattles
The servos mount on two 5/8" round standoffs to the 1/8" holes on the Penguin body. What kind of dimensions are you looking for? The DXFs should provide the exact detail you need, from which you are free to extract dimensions.
Jimmio, pretty soon. We're just finalizing the PCB. As this is not a revenue-generating product we've not been in much of a rush. We're only making 1,000 units and you'll never see it again after that. The reasoning is that the product is entirely USA manufactured and machined, so the cost is truly through the roof.
Ken