Thursday, March 30, 2006

Guide to Choosing Retail NXT or Education NXT

Information on the Lego Mindstorms NXT sets is starting to drip in, but to most people, the distinction between the various sets isn't very clear. If you are a (private?) teacher, and you want to buy NXT, should you buy the retail version of Lego Mindstorms NXT, or the Education version? And what if you are not teaching?

For teachers, I would definitely say Lego Mindstorms Education NXT, because the rechargeable battery will save your budget, and the converter cables help you interface with parts you already have. If you don't need a class license on the software, you can buy an ordinary software license at $42 (which is cheaper than adding a battery pack to the retail set, at $48). So count on spending $292 on your first Education NXT set.

For other users, you need to consider how quickly you will run through your batteries. You could purchase the retail version and get the Education rechargeable battery pack, but this is more expensive than getting the Education NXT and adding the software.

If the rechargeable batteries are not your main concern, then go with the retail set. Although the education base set has an additional touch sensor, converter cables and lights, these items can be purchased separately as well. When you buy the retail set, you can start right away.

Given all possible choices and equal delivery for all, for myself I would choose the Educational Base Set, and add a software license. The battery plus sensor plus cables plus lights are just too good to resist. However, as a MDP applicant who didn't get selected, I can get the retail set delivered in the first week of July, so that is why I have purchased the retail set anyway. Any additional sets I buy will be Education sets, because with the retail set I already have the software that is missing in the Education set.

As a short summary, the retail version is better suited for casual use, and the education version is for people more serious about NXT robots. However, the education version comes without software, which will set you back another $42.

You don't want to compromise and are willing to spend even more on your first set? Then buy the Education NXT Base Set ($250), and in addition buy the Eduction Resource Set ($59), plus an extra software license ($42). The Lego Education Resource Set is a separate box from Lego Education, that contains 670 (mostly Lego Technic) parts, and an additional storage box. By the way, the FLL Robot Set appears to be a Lego Education NXT Base Set plus a Lego Education Resource Set, plus a software license, but a FLL Robot Set cannot be purchased directly.

Here's the list of features of the various offers as I currently understand them:

NXTEducation NXT Base SetFLL Robot Set
NXT Intelligent Brick111
NXT Servo Motors333
NXT Touch Sensor122
NXT Light Sensor111
NXT Sound Sensor111
NXT Ultrasonic Sensor111
NXT Cables777
NXT Converter Cables033
USB Cables111
NXT Rechargeable Battery011
NXT Charger011
Total Parts5774311102
Technic Parts519??
Storage Boxes012
LabView Software Licenses101
Robot Educator?11
Course Materials000
Pre-ordering FromApril 1, 2006April 1, 2006May 8, 2006
Who Can BuyAnyoneAnyoneRegistered FLL Teams

A word of warning: it is not clear how big the rechargeable battery is, and I suspect it is slightly bigger than the 6 AA's of the retail NXT set. I assume it sticks out a bit from the bottom of the NXT Intelligent Brick, and has its own connector to recharge. However, I don't think that for most inventions these size differences matter much.

All pricing is for the U.S., in U.S. dollars. In Euro, the price ranges from €250 to €300 for the retail and educational set, depending on your country. Pricing in Australia and Canada is higher than in the U.S.

All of the advice will be updated as new information becomes available. I'm trying to help you make your decisions, but remember that I may change the advice in the future based on new information from Lego, or the comments to this article.

UPDATE: Lego Education also posted an interesting comparison between the old Lego Mindstorms RCX and the new Lego Mindstorms NXT.

Saturday, March 25, 2006

Pre-ordering Today (Shipping July 1)

Friday, March 24, 2006, some people applied for the MDP but were not selected got an email with the opportunity to preorder. Ordering is not possible on the web. You have to call the Lego Shop At Home phone number, and provide a special confirmation code. Shipments begin July 1.

This is a really nice gesture from Lego towards the people not selected in the MDP. You're not getting the set in March, but you're still getting it about two months before it is in the shops.

Thanks to Louis, Robert, and Michael for bringing this information to my attention!

Unfortunately, I did not personally receive the email. This could indicate one of three things: perhaps the early preordering is only open to residents of the United States, perhaps Lego will notify Europeans early next week, or perhaps there was another email issue. Why would Lego wait to notify Europeans? In Belgium, the Shop at Home phone lines are only open during business hours, so Lego might be holding the email for Europeans until the shops are actually open and they can order immediately.

To find out if the first option was valid, I called Lego Shop at Home in the U.S. on Saturday, 2 AM U.S. local time (Saturday 11 A.M. Central European Time). I provided name of the set, the number of the box, and the special code. I did not get around to explaining I was calling from Europe, because the lady at the phone had absolutely no idea what box or preordering opportunity I was talking about.

Was the order valid only on Friday? Did they rotate people for the night shift and not brief the new people because the offer expired? It would be odd that Lego would only have a single day of ordering. Right now, the U.S. Shop at Home is my only source of information. In a few hours, their customer support is opening. Perhaps that will provide some answers.

I'll keep you posted as I learn more. Any additional information you might have is appreciated! I suggest NOT to make the ordering code public, I do not think Lego wants this information public. Let's not ruin the opportunity for those who can preorder.

Thursday, March 23, 2006

Jim's Blog

Keep an eye on Jim's blog. It has some great content, including a new video on Steve Hassenplug and a question on what you'd like to know on Lego Mindstorms NXT. Since Jim actually has the NXT bricks and his camera at home, you can expect answers to simple questions in about 40 days. So make sure you are subscribed to Jim's blog, or to bNXT's "breaking news" (which includes Jim's blog amongst others). That way, you won't miss a thing. On my own blog, I will not be copying other people's content unless I can add something to it.


On Lego NXT Motors

Phil from Kid Technic, the Australian site I reported on before, alerted me to an update on its page. The connector information is corrected, and new information regarding the Lego NXT motors was added. There is some good information here for those who want to start thinking about their designs today. Read Phil's page here.

Tuesday, March 21, 2006

Call to Teachers

I believe it is great that so many people teach kids Mindstorms. Many of you are just doing this after hours, mostly for the fun of it. I remember when I was 14 (20 years ago), I enjoyed similar "Commodore 64" hobby computer classes in my little village.

But I've been wondering: what would hold someone back to start their own after-hours robotics class in their own town? My first guess would be: knowing what to teach.

That is something many of you have already solved, so maybe teachers can help each other. If you have tought Mindstorms in the past (any version), do you have some sort of syllabus you used, and would you mind sharing it? I would like to collect a number of these, and try to publish some of this on bNXT (adapted for the NXT robots). That way, anyone can start a robot class in their own city. And you don't have to convert your classes to NXT yourself.

To send me attachments, just email me at Make sure the title stands out from spam, e.g. by mentioning the words "Lego Mindstorms". If you don't have a syllabus, but used a book written by others, please let me know about those as well.

Monday, March 20, 2006

HiTechnic Bridge's "Hello World"

Quite a while ago, I wrote about the HiTechnic Bridge, which is is essentially an onboard "IR Tower" that you can connect directly to the NXT Brick. The purpose of this device is to enable direct communications between the old RCX brick, and the new NXT brick. Over the weekend, Steve Barker from HiTechnic released some images of the Bridge.

HiTechnic Bridge

As you can see, HiTechnic placed Lego Technic compliant holes at the bottom of the sensor, in a way that resembles other sensors. The device is connected to the NXT Brick using a normal sensor cable. It communicates with the RCX via the IR light.

By the way, in case you were wondering, you cannot command a TV, radio or other device by simulating the remote control using the HiTechnic Bridge. The device is specially designed to communicate with the RCX.

On Lego NXT Connectors

Dave Breaker just pointed out to me this site in Australia. It has some great photos and closeups.

It contains a good general introduction to the hardware, and shows the connectors in more detail than previously shown. Here's an interesting image.

The motor connectors in the NXT brick, and the connector inside a sensor. Original image here.

The article has some interesting information on the connectors used by Lego.
There is a new wiring and connector system for all the motors and sensors. Uses a 6 wire flat cable and plugs known as a RJ12 6X6 DEC MMJ Modular Plug. It is similar to an standard RJ12 plug but has an offset clip. [...]

On the 4 robots I was playing with there were already several of the plastic release clips broken off from the plugs so I think a crimp tool and a bag of Dec MMJ plugs will always be close by.
Despite the potential brittleness of the connectors, it would be great to be able to make more cables. However, I'm not sure about that RJ12 DEC connector claim. Doesn't that connector have its clip offset to the wrong side?

RJ12 6x6 DEC MMJ connector

Will Lego Execute?

When reviewing Lego bricks and third party hardware, I intend to tell you the good and the bad about the item. Otherwise, I don't feel like it is a real review. Pointing at a few minor issues with a product doesn't mean I cannot like it as a whole.

Likewise, when I look at a company, I can see some things they did really poorly, and still like the company or their products. Lego is in this case.

With Mindstorms NXT, Lego set out to engage their audience. They got 4 users involved in the design, and they went out of their way to engage more people via the Mindstorms Developer Program. So far, the culmination of that effort was the story in Wired, that got a lot of people excited about Mindstorms.

Unfortunately, Lego seems to have taken too much on its plate. When Lego started the MDP, it boldly announced that it would notify everyone who applied for the MDP program about the result, even those not selected. It turned out that over 9600 people applied for the MDP. Still, in the marketing business, sending out 9600 mails is just business as usual. So sending out a polite "sorry" note to all participants seemed like the right thing to do.

Unfortunately, to the surprise of many, Lego did not accomplish this seemingly simple task. Unofficially, Lego has offered a story about emails that got caught in people's anti-spam filters, but this story is not in check with how the Internet really works. It appears that what really happened is that Lego's IT department has killed the "sorry" messages that went out from Lego's Marketing department.

Regardless of what happened or what event was to blame, where Lego really failed to execute is when, despite being well aware that the majority of these "sorry" messages did not arrive with their recipients, they made no (succesful) effort to resend the message. They have also not made any attempt to get a more credible story out on what really happened. It appears that the executives running the NXT MDP did not investigate what really happened, or did not use unbiased advise to investigate this. For a marketing program that tries to show off Lego's user-centric approach, such an attitude seems counterproductive.

We can only hope Lego improves its act as it goes forward. And it will need to do so, because more of its choices are already irritating future customers. As the product gains market visibility, these issues will only become more widely discussed.

For instance, in a tightly knitted community that is spread over the globe, there is an expectation that everyone is treated equally by Lego. Including when it comes to pricing. As it stands, the "official" pricing is only valid in the US and selected other parts of the world. Canadians have to pay 20% for no apparent reason. Australian prices are 20% higher. Northern Europe also pays 20% more than the Southern Europe, yet they use the same currency, and taxes across Europe do not follow the same north-south divide. To show the absurdness of the situation: if I drive 30 kilometers (19 miles), I cross the magical North-South border and pay 20% less. There is nothing to declare when I cross the border due to free traffic of goods within Europe. With eBay replacing the middleman, I suspect online shops in the South of Europe will mail lots of NXT bricks to the Northern parts. No import/export taxes apply...

Perhaps Lego is right in taking the actions it does. But it will need to communicate better about its choices to its customers. Executing in a connected world is very different from the old business. Let's hope they deliver better on their core business, because pre-ordering the Lego Mindstorms NXT starts in just 10 days, on April 1, 2006 (no joke!).

I, for one, will not pre-order. I'm waiting till August, and I'm going to drive those 30 km. Well, at least for my second and third box.

Mindstorms Development Program Updates

Judging from Jim's blog, the Mindstorms Developer Program (MDP) members are allowed to break their silence on May 1, 2006. If that interpretation is correct, there should be a flurry of information on the web in about one and a half months.

Brick-labs has started to collect the names of the MDP people on this page. If you are member of the MDP, and you haven't claimed your place in the hall of fame yet, just go over to Brick-labs and leave them a little comment.

Monday, March 13, 2006

NXT Firmware Likely as Open Source

Andre Wiesner from the Humboldt University of Berlin brings the following news on LugNet, which I think is important enough to republish here:
Loren Sund announced in an interview with German paper "Die Welt" that the NXT software might be published as open source!

> Die Firmware wird von Lego voraussichtlich als Open Source ins Netz
> gestellt, alle Treiber und Schnittstellenprotokolle sollen
> veröffentlicht werden. Die Firmware läßt sich wie beim RCX durch
> fremde Betriebssoftware ersetzen.

For those who don't know German, a loose translation:

"The firmware will most likely be available on the net as open source, all drivers and protocol interfaces will be published. The firmware can be replaced by other operating systems like in the RCX."


Congratulations Jim!

Congratualations to everyone selected, but especially to Jim from The NXT STEP blog on being selected for the MDP!!

Jim is back from holidays, and is going to be live on his blog soon. He wrote this about it:
Interestingly, my blog got 'spam blocked' while I was on vacation and I cannot post anything on it until someone at reviews and confirms that, Yes, I am a Human Being and my blog is Real. (It says 24 hours to review...).
There is also a good article on CNET about the MDP. Be warned, it is a kind of sad article to read for those NOT selected.

Saturday, March 11, 2006

Lego MDP: What I Think Really Happened

In a previous posting, I raised the question if Gmail and Yahoo account holders could have missed their ticket to the MDP due to some problem at Lego.

Lots of people helped me out by providing extra information, I would like to thank all of you. Basing myself on that information, here's what I think really happened.

The unofficial position of Lego is that the bummer mails were caught in people's personal spam filter. I can only speak for myself, but I think it is pretty unlikely that my bummer message was caught in my spam filter. I have checked my spam in Outlook for Lego messages since the beginning of February, and still do, and I have checked the spam filtered messages on Gmail itself. Neither contains the message, unlike many real spam messages.

So that means that my message never arrived at GMail. To the best of my knowledge, routers don't do content filtering, which leaves two options: either it didn't leave Lego as it should have, or it was lost in transit due to unreliable hardware/software. Because so many people have not received the email, the odds on the lost-in-transit scenario are impossibly small.

That means the bummer emails never correctly left Lego, which is not the same as saying that they were not send. I assume that the PR guys did send the message. But even the virus filters on my own PC ask permission to send emails that are addressed to too many recipients. I imagine Lego corporate has protection against undetected virus infections, and if one account sends out too many emails at once or in a short period, the emails are simply killed by their email server before sending them out. Such a anti-virus protection prevents the spread of a Lego virus to its business partners and customers. Such protection would also nicely explain why the first bummer emails arrived (before the treshold was reached, on February 27 and 28), but after that the mass of bummer emails never went out.

One other question I previously asked, is whether you could have missed your invitation as a result of not receiving their email to congratulate you. Lego's answer, as Eric Burdo found out (thanks Eric!), is that almost everyone who was invited for the MDP responded in a positive way. Only two positions had to be filled in a second round, and they were filled easily.

That statement could be consistent with the scenario I just discussed. Lego invited only 100 people for the MDP, and it is likely it did not invite them all at once. The Lego anti-virus software would have considered the email flow as normal, and let those emails pass. It is only as the number of emails went up drastically, say beyond 1000, that the antivirus heuristics started to kill the emails.

Remember, all this is just theory and interpretation. But what it adds up to, is that not receiving your email, means that you were not selected. Nevertheless, you did not miss your opportunity to be on the MDP due to email problems. For me, although it makes no difference whatsoever, that is a relief.

Friday, March 10, 2006

Lego MDP Was Not Open to Gmail and Yahoo Users?

UPDATE on March 11: Also check out the follow up posting, on what I think really happened.

It appears that so far, no Gmail or Yahoo users have received either a "congratulations" or "bummer" email from the Lego Mindstorms Developer Program. One theory is that the Lego email server has a problem with Gmail or Yahoo email addresses, which prevents them from sending email to these addresses. The problem would allegedly be related to the spam filters Lego uses.

Everyone that signed up did receive an initial "Thank you for signing up" email, because that email was sent by a different company.

If the Lego sent out an email to invite you to the MDP, and it was caught by its spam filters, then you never got it. As a result, you could not reply to it. Assuming Lego did not notice the problem in time, that means that everyone who was invited to join the MDP, but was on a Gmail or Yahoo email, did not reply "in time", from Lego's perspective. The position was declared vacant and someone else was invited.

The theory would also explain why Lego appears to have needed a second and even third round to fill all the positions, while I have not heard of any applicant who would not immediately accept were a MDP position offered.

Right now, it is just a theory. But there is a way of testing it: if you applied for the MDP, did you use Gmail or Yahoo, and did you receive a "congratulations" or "bummer" email? So you applied to the MDP, please comment below and help prove or refute this theory...

Let me go first: I signed up using my gmail address, and I did not receive any notification from Lego after the initial "thank you" note.


Sign Up for NXT Pre-ordering Notification

Lego has a webpage available where you can enter your contact information. The webpage promises to contact you when pre-ordering starts. Lego posted this information to the people who applied for the MDP, but did make it in.

Simply click on the link below to fill in your contact information. LEGO Shop At Home will then notify you by email when MINDSTORMS NXT is available to pre-order, so that you can be one of the first to reserve yours!
Alternatively, you can just monitor this blog and I'll tell you when pre-ordering starts...

Dutch Feeds Added !

I would like to thank the guys over at BrickBash Robotics for the special effort they made to add an RSS feed to their Dutch discussion forum. As a result, bNXT can now add Dutch as an official bNXT language! Dutch is actually my mother tongue...

As you'll notice, the menus are in place and the news feed will start filling later today, if all goes as planned.

bNXT on Robot Dreams

Thanks to Robot Dreams for the nice review.

Jim told his readers to monitor my site for news, but I'm afraid little news to be published. I guess Jim picked a perfect time to go on holidays.

Tuesday, March 07, 2006

MDP Team Public Soon?

Today, Steve Hassenplug wrote the following on LugNet:

Do you really want to know who's in the MDP?

It's a very impressive mix of experience, and fresh blood. I suspect you'll hear more soon...


I'm looking forward to hear more...

Friday, March 03, 2006

Ordering NXT Begins In One Month

According to the LEGO MINDSTORMS NXT Education blog, the online store will take orders for the LEGO MINDSTORMS Education NXT set early in April, just one month away from now.

The ordered sets will be shipping in August, in time for the new school year.

Given that for the (paying) Mindstorms Developer Program, 9610 people applied in one month time, it is not unthinkable that Lego may be facing supply problems for the first units shipped. Would they use a first come, first served principle? In that case, better order as soon as the online shops opens.

Lego Announces 100 MDP Chosen out of 9610 Applications

Lego announced today that for the Mindstorms Developer Program, the 100 participants have been selected. It also said that 9100 people have selected, with ages ranging from 35 to 75 years old, coming from 79 countries. About 50% of the applicants was younger than 35.

The 100 selected people will focus on one of three aspects:
  • software and hardware development;
  • inventions and creations;
  • community.
Personally, I would create an invention like My First NXT Project, to develop software in libraries for all to use in different ways, and write about that on my blog. This seems to go right across the categories.

The announcement also raises some questions. So far, not everyone has who that applied with Lego, has received either a “congratulations” or “sorry” note. Is that notification delay purely of a technical nature, related to the quantity of notes to send out?

In the press release, Lego indicates that the units are shipping now to the MDP participants. Lego will accept selected changes to the product until just one month before launch.

As a side note, Lego intends to provide ordinary batteries with the kit to MDP participants. So it appears that the rechargeable battery kit that comes with the educational NXT may not to be ready for shipping yet. If Lego wanted the MDP kit that they sent out to be cost-efficient, they would ask participants to buy their own batteries. Within the 100 selected MDP participants, there are teachers. Lego would want to hear from them how they feel about the new rechargeable battery pack. The pack also has a wall socket, which would have allowed testers to run the robot software for longer test periods.

Overall, if the pre-launch interest is any measure, Lego should have a huge success on their hands with Lego Mindstorms NXT.

Thursday, March 02, 2006

Multi-Language Support on Website

The website has had a "breaking news" section for some time. The Breaking News section is bringing together all the blog and website feeds on Lego Mindstorms NXT that I know, and that are in English.

Twelve year olds who do not speak English at home, rarely can read English. That's why I believe it is a very good thing that new blogs are starting all over that are written in the local language.

Therefore, I have added a second "Breaking News" section. This section brings together all the news on Lego Mindstorms NXT that is written in Portuguese as well as new that is written in Spanish.

That way, you can subscribe to the feeds for which you know the language, but do not get bothered with news entries that you don't understand because of a language barrier.

If you have a Lego Mindstorms NXT website with an RSS feed in any language, please let me know. I'll make an effort to support as many languages as I can.


Wednesday, March 01, 2006

HiTechnic's Digital Compass

Previously, we reported that HiTechnic will have a digital compass ready for Mindstorms NXT in time for the launch. Today, Steve from HiTechnic sent us more details about the product.

© HiTechnic.
A picture of the new NXT Compass sensor

Why Do You Need a Compass?

A common way for a robot to know its position, it to count the number of rotations its motors, a process called dead reckoning. Dead reckoning is a great technique, but there are two physical problems with it. Slipping wheels cause errors, and collisions may throw it off course. Another problem with dead reckoning is that if you want the robot to retrace a previous path, it must be exactly in the same position. Often, it is very hard to position the robot so accurately at its starting position. Turning the robot by one degree is hardly visible.

A high accuracy compass makes it a lot easier to determine your position. You still count the number of rotations of your motors to understand how far the robot has traveled, but you use the compass to steer the robot in the right direction. When the robot starts, it does not matter as much whether the robot is a few degrees off course. Using its compass, it will immediately correct to drive in the intended direction.

More advanced software can combine the directional calculations of dead reckoning with digital compass measurements to provide the best performance, but this is beyond the level of the average Lego Mindstorms NXT builder.

If you are building a walking robot, dead reckoning is even more complicated. In these cases, the compass may be all you have to go on.

HiTechnic’s NXT Compass

HiTechnic's new compass for Lego Mindstorms NXT is well suited for its job. The compass is accurate to 1 degree, which is just what you need to rely on it for navigation. Reading the compass is very easy. The compass returns a value from 0 to 359 degrees that indicates the direction the sensor is currently pointing in.

Off course, there is only 1 degree difference between 359 and 0 degrees. Since a robot easily slips one degree, you may be fluctuating from 0 to 359 and back. Make sure your robot software shouldn't overreact to such seemingly large differences!

A compass works by detecting the magnetic field of the Earth. Magnetic sources nearby can seriously disrupt any compass, and electric motors create a magnetic field when they work. Putting a compass on a robot can therefore create very strange results. Fortunately, the effect can be compensated by measuring the distortion of the local magnetic fields. Measuring and storing the amount of compensation is called calibration. To calibrate the compass, you need put it in calibration mode, and drive the robot in a circle.

On HiTechnic's legacy RCX compasses, there is a button on the sensor that puts the compass into calibration mode. For the NXT, HiTechnic has raised the bar. You can put the compass into calibration mode from within your program. So when the robot wakes up, you can automatically put the compass sensor into calibration mode, have the robot turn (at least) 360 degrees, and finalize the calibration. How cool is that?

HiTechnic's NXT Compass remembers the calibration values, even when powered off. So after you build the robot, you only need to perform the calibration once. However, if you use the robot later at a different location, you may still need to recalibrate the sensor to compensate for the local magnetic environment. Perhaps you could provide a calibration option in your software? For instance, "if the touch sensor is pushed when starting up, then recalibrate". As a result, you can hide the compass sensor somewhere in the body, and trigger the calibration by pressing an external sensor (or button) when the robot starts. You could even use my previous post on Bluetooth, to recalibrate the robot using your mobile phone as remote control.

Because the calibration is under software control, you can also prevent calibration when it is not appropriate. For instance, recalibrating in the middle of a maze or halfway on its route is bound to disorient the robot! It would be better to drive the robot back to its base station and recalibrate there. With HiTechnic's NXT Compass, it is up to you to decide when you allow to the sensor to recalibrate, and what other actions you take when it does!

Accurate When Horizontal

The compass sensor is designed to be accurate when horizontal, although you may get away with not mounting it perfectly horizontal. Steve from HiTechnic explains the errors related to the horizontal position:
Compass dip, or the angle of the earth's magnetic field (up or down) at the earth's surface, together with pitch and roll which produce compass errors will not typically be a factor in using this sensor because of the relatively small distances and slow speeds involved unlike a compass in an aircraft for example, which will be more prone to dip error, especially when close to the poles, and errors when rolling in and out of turns.

Any compass error will depend on a number of factors including distance from the equator (the closer to each pole, the greater the dip angle) and pitch and roll angles, and the heading relative to these factors. Compasses on ships are gimbaled so they can stay horizontal to compensate but in the typical NXT application I don't expect these to be significant factors. Keeping the Compass Sensor horizontal and calibrating it for each installation will minimize errors.

Illustration of dip angle

If you didn't get that, don't worry. Just keep the sensor horizontal and recalibrate the sensor from time to time!

Technical Specifications

For the technically inclined, here a short note by Steve from HiTechnic on the NXT compass and its calibration:

© HiTechnic.
Interior of the NXT Compass sensor

The compass uses a Honeywell dual magnetometer. The results obtained from the two magnetometer channels are processed by an onboard PIC micro controller which computes the ArcTan of the two channels' readings and outputs a magnetic reading value between 0 - 359.

To ensure accuracy in every installation, the compass should be calibrated to negate the effects of any local magnetic interference from sources such as the NXT. The onboard firmware can be set to calibration mode under program control from the NXT.

The compensation consists of measuring the response of the two channels during a complete revolution of the device and computing the locus of the readings. This locus should be circular, centered on (0,0). Calibration coefficients are calculated and are those values necessary to adjust the raw readings to make them circular and centered on (0,0). Once calibrated, the NXT Compass will retain the compensation values, even when powered off.

MDP Slightly Delayed?

It appears Lego is running slightly behind schedule for the Mindstorms Developer Program. Although Lego has contacted the first 100 selected people for some time, it has not contacted those that did not make it into program so far. They might be waiting for the last confirmations to arrive from people selected in the second or third round. Lego previously intended to complete the selection of the MDP lucky 100 before the end of February.

UPDATE: Emails have been arriving since February 27 in some people's inboxes stating that they were not selected. However, this does not necessarily imply that the MDP is now finally closed. Perhaps it no longer made sense keeping thousands of people waiting when only a few positions still needed filling.