Monday, June 26, 2006

NXT Shipping in U.S.

As Jim said last Saturday: shipping notices have been sent by Lego Retail Sets for people who ordered their set with a delivery point in the United States.

Those in Europe will have to wait a few months, but even there Lego will be well on time for the holiday season! Since I can't wait till October (offcial Belgian launch), I ordered mine to a friend in the U.S. So I hope to receive my sets in the next weeks.

People in the U.S. who ordered for Lego Education will have to wait one more month, I believe Lego Education starts shipping in August.

9 Comments:

At June 26, 2006 6:27 PM , Anonymous Cloned Milkmen said...

Orders to Canada are also shipping right now.

 
At June 26, 2006 9:29 PM , Anonymous Filip said...

Thanks for that!
Filip.

 
At June 27, 2006 4:43 PM , Blogger rb95403 said...

July 3rd delivery date -- due to East Coast to West Coast travel time?!

First delivery please post!

 
At June 27, 2006 6:06 PM , Anonymous Anonymous said...

Mine arrived this morning! New York, 10:00 am June 27th 2006.

 
At June 29, 2006 5:34 AM , Anonymous Anonymous said...

Mine arrived Tuesday afternoon, but due to office surgery, I was only able to just finish using the Wonderful Dearth of Information to count parts (at least they gave us a beam measurement chart, something absent in a big Technics kit I ordered for parts and because it looked like a Great Base for Robotics - but they could have named and numbered the parts inventory instead of showing pix of rows of (well I think they're black connecting pins?) etc.

The dearth of available info is surprising, as is the release by Lego Store of pre-orders for a product schedule still aimed at 'Official Rollout in August, according to a source associated with Stamford.

What bodes even worse is a close examination of information admittedly given before, which many of us, in our hopes, chose to ignore:(manual p 11 of 76(almost total doc)) Technical Specifications...32 Bit ARM-7 [the Good News - an Arm-7 powered many a Palm pilot, can handle any complex math you throw at it faster than a robot can, etc.
Bad News: 256K flash mem [disk equivalent] 64K RAM - now maybe I've been spoiled by modern machines and it's been a Long Time since I toggled programs into 1K core pages of my 1965 PDP-8 with a 4K memory block. But the ARM design usually accepts SD flash cards with up to 2 GB of memory, and I'd say the box has room for a 1 gig SIM, not that we need that much space.
BUT it's a 32-bit RISC CPU and RISC means Lots of Fast Instructions needing a big memory area.

and now for the
WORST NEWS: Rather than go with a modern 32-bit hot I/O chip enabling oh, things like analog granularity into the terabits, and the ability to run a fast chain of 128 USB devices from each of several ports, I/O is apparently (I say apparently, because I haven't had the time to warm up the 'scope and break the manufacturer's warranty before even trying my new toy) run through an AVR 8-bit microcontroller with a seperate 4K Flash and 512 byte RAM.
512 bytes, ONE HALF a KByte, when a KBYTE fits on a pinhead!!!
That means, in addition to making multiplexing I/O that much harder, limiting us to the limited range of the Particular 4 input and 3 motor output (with the idea that movement should be controlled by a right and left motor ONLY [the kit confirms this by dumping the old transmission and differential gears]) and the maximum 8-bit granularity of 0-255base 10, 8-bit sound quality with a maximum frequency of 8 KHz!
We do have the BLING factor of the ability to build a "fun" (yes, the mostly-wordless manual manages to use "fun" as an adjective about 40 times, and also always promises na "cool" reaction from pressing a button) "graphical"[folks, no matter what Redmond tells you, the word is g-r-a-p-h-i-c] 100x64 monochrome display - probably the largest job assigned to the 32-bit CPU, that and running the USB, and, admittedly, useful, though short-ranged Blutooth port, I can see many looking to fill the USB port with an 802.11g card and the firmware to declare the "slave" port (which, technically means it lacks a +5V line) to run as a master, and use a second channel to handle a USB ping-pong ball camera to watch a robot stroll the house. Of course, that means carrying more power, both for the USB devices and the 6xAA-powered NXT, which could be accomplished by using less-hungry motors (servos eat amps)and, for starters, separate banks of D-cells for both the NXT controller and USB/devices.
Hey, Lego's got non-servo motors that can pull Lego Brick TRAINS, and I doubt Lego will be the last word in sensors or sensor multiplexing or firmware and programming!
I'm just a hardware bug, will get my soft-side friends to optomize the firmware while I do the interface work - I mean I'd been holding on to the great optical sensors and microswitches I scored recently when I ripped down a pair of ancient laserprinters, and I don't intend to waste 'em.

Hey, Software Types, get working! Another set of Alternate Firmware to make the ARM a USB (including USB flash memory card) aware, and I'll build the controller for the live communications with range and find the motors that'll get us there without carrying a 10-lb gel-cell and 6-foot solar panel to keep it charged.

Good Luck to all!
dmr

 
At June 29, 2006 9:53 AM , Anonymous Filip said...

Hi dmr,

Interesting comment!

Running a Lego train is far less demanding on the motors than functioning as a robot leg. Actually, the Lego Racer motors are even more powerful than the NXT motors, but they consume too much energy. Lego probably did a good job when looking at the motor tradeoffs, and the "servo" side makes a lot of difference. But it shouldn't be to hard to take other motors, and build your own Lego servo motor the way you want it.

I certainly think the number of ports is a problem. I hope HiTechnic's multiplexer will be the solution.

About the rest of the hardware, wouldn't it make more sense to connect a PDA or even PC (UMPC?) to increase the computing and I/O capabilities?

I'm a Software Type, but I prefer working on robot intelligence rather than on operating system and other low level stuff. For me, it not just another CPU with devices. It's a robot. Let's avoid walls, navigate a maze, find our base station, auto recharge, stuff like that.

But I hope there are lots of types like you, that do what I don't like to do :-). After all, the current NXT Ultrasonic sensor was first developed for the RCX outside of Lego, and spotted by Lego as a good thing to add as a standard part. Perhaps things you design will make it into the next NXT set.

 
At June 29, 2006 2:34 PM , Anonymous Anonymous said...

I got mine on Wed June 28th. My son and I built the quick start bot and did some simple programming.

 
At June 29, 2006 6:24 PM , Blogger Brian Davis said...

Hmm, a couple of comments... first, keep in mind that whatever a lot of us would *like*, LEGO is making and selling toys... for kids. This means they need to make it easy and "safe" (as in preventing a kid from getting lost in the details) for 10+ year olds to operate, *without* adult supervision. That's a tall order. Second, it needs to be affordable for the target audience... and again, That's Not Adults. We are a portion of the market, and a vocal portion, but not even close to the bulk of the market (by some estimates, around 5%).

The documentation is rich... compared to any other childrens toy (and the in-editor documentation of the NXT-G language even better, including walk-throughs). The memory is a problem, IMO, but LEGO decided to go with the memory that was on the ARM itself, and *not* add the cost and complexity of external memory... and for this purpose, they used the most memory-rich ARM in the line. As to what the CPU is doing, well... polling four input ports (any of which can support a digital format... think I2C, bit-banged in the CPU to save the cost of still more specialized chipsets), handling the USB, dealing with the Bluecore chip and BT messaging, sending messages to/from the AVR (which handles PWM of the output ports BTW), decoding and tracking three motor encoders from the outputs, running the sound system (limited, yes... but It's A Toy, remember), handling the LCD display, monitoring the battery level and flagging battery level issues to the LCD before browning out the CPU... and in its spare time, executing the users program.

Also, I *think* the difference between a USB master and slave is in more than just providing power... IMS, there is also a different (and, again, slightly more expenssive) chipset needed for a master USB.

As to the motors, yep, they're power-hungry... but not *nearly* as power-hungry as the train motors mentioned (these need to be run off a wall current supply, and actually have low efficiency, relative to the other LEGO motors). And LEGO motors tend to be rather efficient, given their torque and shaft speed. (Also, sometime I'd love somebody to tell me what the true difference between a continuous-rotation servo and a "normal" high-efficiency DC electric motor... I'm not sure, to be honest, which LEGO has used in the NXT).

So is there any good news? Yep: there are already folks working on HW expanders for the input and output side, at least one alternative FW already exists (and I can think of a couple more on the way), and with BT built-in the potential exists to "network" a fair number of devices without a single physical wire, or even back to a remote computer to "teleoperate" the robot (not my cup of tea, but a lot of people like that option).

I can go on about the number of things I would have done differently (and I have rambled on about it, in some cases to LEGO). But in the end I'm not sure I could have done anything different or better... again, reference the very first paragraph as to what LEGO is selling here.

--
Brian Davis

 
At June 30, 2006 11:33 AM , Anonymous Filip said...

Thanks for the long comment, Brian.

I think you should also mention "size" and "battery life".

The lowest end PC's from DELL here in Belgium can be bought for € 199 (excl. VAT & delivery), and even a 1 GB RAM (80 GB HD) machine was available for € 260 (well, that promotion went away quickly). However, these PC's are power hungry [battery], need a fan (always a problem with robots), and fit nowhere into a nice-sized robot brick [size]. So at this time, these PC technologies were not a good option. I expect that to change in the future, but Lego can change when that happens in some future Mindstorms version, since onboard application backward compatibility is not an issue for Lego toys.

Regarding that 5% of adults, I think the stats are skewed. I think only 5% of adults SAY they bought it for themselves.

Plus, in my former robotics labs, we tried to use the RCX for robot modelling when testing out new ideas, but the RCX unfortunately wasn't up for the task. With Bluetooth connecting back to a PC, servo motors, and lots of new parts entering the market, there is not a lot that (combined) NXT robots cannot model. I would expect a pretty big take up from anybody doing research into robots.

Finally, Bluetooth encourages you to use multiple robots together. I already own more NXT robots than I had RCX robots. But I expect this feature to be used mostly by "power users", which will mostly be older teenagers and adults.

Regardless of how much better it is for adults, I believe Lego did a great job in making this technology usable for kids.

 

Post a Comment

Subscribe to Post Comments [Atom]

Links to this post:

Create a Link

<< Home