Powered Up - A tear down...

[Lok24]

Hi,

I have no problems.

How do you connect the Hub, handheld or SmartDevice? Is connection ok?
Do you have another motor? Or the chance to use a battery and some or a tester to check the motor?

 

 

 

[Electricsteam]

On 10/18/2020 at 8:17 AM, Lok24 said:

Hi,

I have no problems.

How do you connect the Hub, handheld or SmartDevice? Is connection ok?
Do you have another motor? Or the chance to use a battery and some or a tester to check the motor?

 

 

 

I use both the controller and the smart phone. I think the motor has just given up the ghost.  Does lego still do replacements and repairs?

[TheSpaceCadet]

On 9/29/2020 at 12:58 PM, biasedlogic said:

I bet it is protected in some way, but you can't just 'patent a connector'. If we talk patent, we talk some novel property or solution. The plug of LEGO PUP isn't novel in the technical solution - it's just a modular crimped connector in a bit unusual shape, but the shape itself also does not solve any old problem in a novel way. You don't get through the European patent office with something like that easily.

What is easier, is to protect a pattern, like you protect a logo or a trade name. But this has to be recognizable pattern for your company and does not necessarily refer to the technicalities.

Getting around a patent for a crimped connector is not so much of a problem - see my design, it's not crimped and the technical solution to obtain the necessary form is vastly different from LEGO's

Getting around a protected pattern is a bit harder thing, because it depends on what is exactly protected. It may be as easy as making the back of the plug round, leading cable sideways out and making it all lime-green so it looks nothing like LEGO, or something entirely impossible, if the actual shape of the contacts is protected pattern... I don't have time now to poke around and find the relevant IP, but if someone has it on hand I'd love to have a look.

greets!

M.

I had the same problem when I was trying to develop a controller/hub for Lego's new devices. I use a connector that you crimp right onto the cable, leaving the original cable fully functional but also breaking it out in a more universally usable way. If you'd like to see the full implementation and hub, check out the Kickstarter I just launched. Hopefully this plug isn't too shameless. Also, I ended up cutting all my cords - I just can't bring myself to fork out over $300USD for a spike prime when I've got plenty of technic pieces and I just want to program in Arduino.

Edited October 21, 2020 by TheSpaceCadet

[le Suedois]

Hello everyone,

I just joined eurobricks after seeing this topic. I did a little "game" last year and I am looking to do a update to the new powered up motors. Take a look here : 

 

There is no link with recent games released by nintendo or lego. This is only fan art and was only used to promote the maker movement in france.

First I want to share you this for all those peoples looking for pup, spike and control+ connectors. The electric component is for use with Eagle/Fusion 360, the stl is for your 3D printer! It's free for non commercial use ;)

https://github.com/LeSuedois/PoweredUpConnector

Second I want to ask if somebody managed to get the motor encoders working with Arduino? Or is there a page grouping information about the serial communication? I already know the BLE way but I don't like for this project, the hub is just to big...

I want to replace the RC servo with the new control+ sensored motors, more lego is better..

Tank you for this topic!

Edited November 18, 2020 by le Suedois

[Lok24]

34 minutes ago, le Suedois said:

Second I want to ask if somebody managed to get the motor encoders working with Arduino?

Yes, you could check the legoino-library

[ruppie]

Ask @GianCann about his basic example he made with color sensor and ESP32, i think sample code here will be good point to start with.

Plenty of other info can be found within "pybricks" documentation.

Edited November 21, 2020 by ruppie

[Jetro]

11 minutes ago, ruppie said:

Ask @GianCann .

I'm sorry to inform you that @GianCann passed away today, as a result of complications after contracting Covid-19

https://www.facebook.com/groups/PoweredUp/permalink/2498381407129982/

[Phil B]

6 hours ago, Jetro said:

I'm sorry to inform you that @GianCann passed away today, as a result of complications after contracting Covid-19

https://www.facebook.com/groups/PoweredUp/permalink/2498381407129982/

That is quite the shocker .... I didn't know @GianCann personally but I saw his frequent replies here on this forum. My thoughts go out all who knew him and of course his family and friends.....

[ruppie]

I did not know him , too . I allways enjoyed the controversal, but  inspiring discussions with him, very much. I am looking foreward , to continue this dicussion and exchange when we once meet on the "other side". Too the family and friends: He has left many footsteps behind, we can still follow him.

I think he will be always missed, but never forgotten ;-)

 

[andrei66]

Hi, I have been a LEGO fan as a child, and I recently received a Lego Technic set (42099, the 4x4 offroader). As an engineer myself (now, at almost 22), I began modifying the truck to make it faster and more powerful. I have already added a 4S LiPo pack using a regulator to 10V output, inside a box I found on @Philo's website (for the Control+ hub). The performance is now a lot better, and it does not decrease with battery discharging (since the output is a fixed 10V from the regulator). However, I am now thinking about changing the motor drivers, by bypassing them. I am only taking about the 2 driving motors, the steering motor will be left as it is.

I studied the datasheet of the LB1836 driver and found that the IN1..4 signals that drive the 2 motors must be PWM (since there are no other pwm control inputs), and they should be 3.3V (since the STM32 is providing them).

I bought an L298N dual driver module after studying its datasheet, and found that it has an almost exact input structure: 2 IN signals (which can be PWM) per motor, and an extra EN signal (enable, usually used with arduino projects for pwm speed control, while the IN signals are either 01 or 10 for direction), but from the arrangement of the logic gates and the statement that the IN signals are "TTL compatible", it must mean that you can directly use PWM on the IN signals, while setting the EN signal to 1 (using a provided jumper), so in theory, I could break the M1 and M2 lines from the C+ hub, cand power the motors from this driver instead, using the same input signal as the existing driver, by soldering wires to the PCB and routing them to the inputs of the motor driver. This would mean powering the motors from about 15V (full charge) to about 11V (considering the ~2V voltage drop of the driver)

My question is, since I have seen this is a very active and engaged community, whether anyone has ever done this kind of thing before or is even thinkinh about it.

My quest for power comes from the fact that I previously had a modified RC car that I fitted with a 3000W electric motor that hit 120kmh and had immense power. I know I am never going to reach that with this offroader, but I will at least try to get the most out of those motors.

And yes, I know about the resettable fuse in the motor case and its current limit, and I do not think it will be such an issue.

Thank you kindly for your help! Any ideas are appreciated.

You can see for yourself the datasheets that I referenced:

https://www.sparkfun.com/datasheets/Robotics/L298_H_Bridge.pdf (new driver)

https://www.onsemi.com/pub/Collateral/ANDLB1836M-D.PDF (original driver)

https://www.flickr.com/photos/147573503@N04/albums/72157711852953692/with/49092205906/ (PCB with traces, where I would take the signal from the driver contacts, according to the datasheet) 

Edited December 26, 2020 by andrei66

[dr_spock]

2 hours ago, andrei66 said:

Hi, I have been a LEGO fan as a child, and I recently received a Lego Technic set (42099, the 4x4 offroader). As an engineer myself (now, at almost 22), I began modifying the truck to make it faster and more powerful. I have already added a 4S LiPo pack using a regulator to 10V output, inside a box I found on @Philo's website (for the Control+ hub). The performance is now a lot better, and it does not decrease with battery discharging (since the output is a fixed 10V from the regulator). However, I am now thinking about changing the motor drivers, by bypassing them. I am only taking about the 2 driving motors, the steering motor will be left as it is.

I studied the datasheet of the LB1836 driver and found that the IN1..4 signals that drive the 2 motors must be PWM (since there are no other pwm control inputs), and they should be 3.3V (since the STM32 is providing them).

I bought an L298N dual driver module after studying its datasheet, and found that it has an almost exact input structure: 2 IN signals (which can be PWM) per motor, and an extra EN signal (enable, usually used with arduino projects for pwm speed control, while the IN signals are either 01 or 10 for direction), but from the arrangement of the logic gates and the statement that the IN signals are "TTL compatible", it must mean that you can directly use PWM on the IN signals, while setting the EN signal to 1 (using a provided jumper), so in theory, I could break the M1 and M2 lines from the C+ hub, cand power the motors from this driver instead, using the same input signal as the existing driver, by soldering wires to the PCB and routing them to the inputs of the motor driver. This would mean powering the motors from about 15V (full charge) to about 11V (considering the ~2V voltage drop of the driver)

My question is, since I have seen this is a very active and engaged community, whether anyone has ever done this kind of thing before or is even thinkinh about it.

My quest for power comes from the fact that I previously had a modified RC car that I fitted with a 3000W electric motor that hit 120kmh and had immense power. I know I am never going to reach that with this offroader, but I will at least try to get the most out of those motors.

And yes, I know about the resettable fuse in the motor case and its current limit, and I do not think it will be such an issue.

Thank you kindly for your help! Any ideas are appreciated.

You can see for yourself the datasheets that I referenced:

https://www.sparkfun.com/datasheets/Robotics/L298_H_Bridge.pdf (new driver)

https://www.onsemi.com/pub/Collateral/ANDLB1836M-D.PDF (original driver)

https://www.flickr.com/photos/147573503@N04/albums/72157711852953692/with/49092205906/ (PCB with traces, where I would take the signal from the driver contacts, according to the datasheet) 

Welcome aboard, Andrei.  The spec sheet say min high 2.3V.  Test and see if you don't need to use a 3.3V / 5V logic converter..  :-) . 

[biasedlogic]

8 hours ago, andrei66 said:

I studied the datasheet of the LB1836 driver and found that the IN1..4 signals that drive the 2 motors must be PWM (since there are no other pwm control inputs), and they should be 3.3V (since the STM32 is providing them).

I bought an L298N dual driver module after studying its datasheet, and found that it has an almost exact input structure: 2 IN signals (which can be PWM) per motor, and an extra EN signal (enable, usually used with arduino projects for pwm speed control, while the IN signals are either 01 or 10 for direction), but from the arrangement of the logic gates and the statement that the IN signals are "TTL compatible", it must mean that you can directly use PWM on the IN signals, while setting the EN signal to 1 (using a provided jumper), so in theory, I could break the M1 and M2 lines from the C+ hub, cand power the motors from this driver instead, using the same input signal as the existing driver, by soldering wires to the PCB and routing them to the inputs of the motor driver. This would mean powering the motors from about 15V (full charge) to about 11V (considering the ~2V voltage drop of the driver)

My question is, since I have seen this is a very active and engaged community, whether anyone has ever done this kind of thing before or is even thinkinh about it.

Well, I'd say this isn't the way to go.

1. You are replacing halfway modern motor driver with something ancient, dropping almost 4V at full chooch, leeching 70mA for itself when idle, and to boot slow to switch. If you want to upgrade, get some decent modern mosfet based motor driver and not a slow dinosaur.

2. The PoweredUp connectors are not good for currents noticeably higher than what the original driver delivers. I'd eyeball them at 500mA constant, 1A momentary, going by these being basically goldpin sockets with single-sided contact, single point of contact, contact force provided by the contact alone (no separate spring element). You ssure need to bypass these when increasing motor power.

3. While the internal gearing of the motors sure has some leeway in terms of torque overload, what it does not have is lubrication. You may do a short proof of concept run at high power, but you will be replacing expensive motors fast. If you are going for modding LEGO, give a thought to replacing the motors first. Can be done, needs some basic tooling and/or a 3D printer. If you are chopping up the bricks already, you may give this a thought as well.

4. If you want to boost the hub, better leave it as is, pick the motor PWM signals at the PoweredUp plug. Yes, they are at power levels there, but a simple voltage divider is all you will need. With a bit smarter circuit you could try detecting the free-wheeling state of the driver, but that will unlikely be necessary for performance applications. You can make own Powered-Up plugs or splice the wire to the motor. The power to the driver anyway has to be routed directly from the power battery, so you may leave the brick power at its original batteries, no motor means low battery drain from the brick.

Good luck!

M.

[andrei66]

55 minutes ago, biasedlogic said:

Well, I'd say this isn't the way to go.

1. You are replacing halfway modern motor driver with something ancient, dropping almost 4V at full chooch, leeching 70mA for itself when idle, and to boot slow to switch. If you want to upgrade, get some decent modern mosfet based motor driver and not a slow dinosaur.

2. The PoweredUp connectors are not good for currents noticeably higher than what the original driver delivers. I'd eyeball them at 500mA constant, 1A momentary, going by these being basically goldpin sockets with single-sided contact, single point of contact, contact force provided by the contact alone (no separate spring element). You ssure need to bypass these when increasing motor power.

3. While the internal gearing of the motors sure has some leeway in terms of torque overload, what it does not have is lubrication. You may do a short proof of concept run at high power, but you will be replacing expensive motors fast. If you are going for modding LEGO, give a thought to replacing the motors first. Can be done, needs some basic tooling and/or a 3D printer. If you are chopping up the bricks already, you may give this a thought as well.

4. If you want to boost the hub, better leave it as is, pick the motor PWM signals at the PoweredUp plug. Yes, they are at power levels there, but a simple voltage divider is all you will need. With a bit smarter circuit you could try detecting the free-wheeling state of the driver, but that will unlikely be necessary for performance applications. You can make own Powered-Up plugs or splice the wire to the motor. The power to the driver anyway has to be routed directly from the power battery, so you may leave the brick power at its original batteries, no motor means low battery drain from the brick.

Good luck!

M.

1. I did not know this was an "ancient" driver, I only worked with BLDC drivers before, and a Sabertooth 2X60 for DC motors, which is way too overpowered for these motors. Could you recommed a modern DC motor driver that would be good for this task?

2. I saw they were quite thin, I was also worried about this, maybe changing the wire too would be a good option

3. This was actually my main concern, at first I wanted to completely replace the motors (I also have a 3D printer), but I figured it was to much fiddling around, and so I thought about boosting the current motors. By the way, do you know what kind of DC motor size would fit in place of an Control+ XL motor? I know the interior motor is quite small, but I'm not versed in DC motor sizes.

4. Ah yes, I didn't realize the output of the driver is also PWM, but instead of 0 - 3.3V it is driven at 0 - Battery voltage. Good point there, I missed this observation. Yes, practically a voltage divider that divided this voltage by about 3 would be the same deal when it comes to the input signal for the new motor driver.

Thanks for the advice. I'm having doubts now whether this would be beneficial in the long run. Certanly a good thing I wrote here, maybe I would have done it and toasted the motors soon after. I think I'm going to think better about this mod, maybe explore some other motor options and check out other drivers.

Thank you!

7 hours ago, dr_spock said:

Welcome aboard, Andrei.  The spec sheet say min high 2.3V.  Test and see if you don't need to use a 3.3V / 5V logic converter..  :-) . 

Yes, it does. And the datasheet for the original driver says min high is 1.8V, so I would bet that the actualvalue is 3.3V from the MCU, since that is what it delivers on its output pins, although I could be wrong. I will measure it too to be sure. But @biasedlogic suggested a better idea of using the motor output directly, with a voltage divider (that would map the 0 - batt voltage values to 0 - 3V for example), so I think this is a better idea

Edited December 27, 2020 by andrei66

[biasedlogic]

10 hours ago, andrei66 said:

1. I did not know this was an "ancient" driver, I only worked with BLDC drivers before, and a Sabertooth 2X60 for DC motors, which is way too overpowered for these motors. Could you recommed a modern DC motor driver that would be good for this task?

2. I saw they were quite thin, I was also worried about this, maybe changing the wire too would be a good option

3. This was actually my main concern, at first I wanted to completely replace the motors (I also have a 3D printer), but I figured it was to much fiddling around, and so I thought about boosting the current motors. By the way, do you know what kind of DC motor size would fit in place of an Control+ XL motor? I know the interior motor is quite small, but I'm not versed in DC motor sizes.

4. Ah yes, I didn't realize the output of the driver is also PWM, but instead of 0 - 3.3V it is driven at 0 - Battery voltage. Good point there, I missed this observation. Yes, practically a voltage divider that divided this voltage by about 3 would be the same deal when it comes to the input signal for the new motor driver.

Thanks for the advice. I'm having doubts now whether this would be beneficial in the long run. Certanly a good thing I wrote here, maybe I would have done it and toasted the motors soon after. I think I'm going to think better about this mod, maybe explore some other motor options and check out other drivers.

Thank you!

Ad.1: The datasheet current revision has "Jenuary 2000" (yes, there's typo there) on front page, which would date the design at 20 years old, but don't be fooled! The hand-drawn reference PCB layout is a strong indication the truth is deeper yet! The chip has been around since at least 1988, it was actually designed and made by SGS Thomson (1998 renamed to ST), the first public preliminary datasheet was dated September 1988. The design of the chip is older still, the sister chip L297 (stepper driver with 4 channels) was designed around or before 1987... As about what to use instead - it's hard to recommend something, best look at what polulu.com has on their website. The thing is, chips are plenty, but the packages are not breadboard friendly, so best select something that's available in single numbers on a breakout board.

Ad. 2: Wires are an issue too, but these are easily replaced when using a custom PoweredUp plug (see https://www.biasedlogic.com/index.php/lego-powered-up-connector/). The problem is the socket in the hub itself.

Ad. 3: Again, it's hard to give recommendations. LEGO is quite good at being consistent, i.e. the axles, the gears, the connectors, these all have a breaking point somewhere, and the power of LEGO motors chosen so as not to hit that breaking point. You have to be careful about interfacing LEGO geartrain to anything more powerful. I had some success with using a DFRobot motor https://www.dfrobot.com/product-1617.html but not with PoweredUp, but with Mindstorms EV3. With PoweredUp you are losing position/speed feedback with 3rd party motors, if you want that you have to go down the EV3 route (see https://www.biasedlogic.com/index.php/running-third-party-motor-with-your-mindstorms-ev3-nxt/). For a race car you, however, likely won't care about feedback (remember to code the motor as a simple motor in that case)

I'd say the LEGO system is the kind of system you can't really improve on power-wise without changing a lot of elements and then it stops being fun. There's place for mods, but function-wise, not performance-wise. I use, for example, 3d printed O-ring wheels for robots to make movement more precise. Or there are 3d printed railway elements, like crossings, that are not available from LEGO directly. Or there are 3rd party sensors for EV3 that can expand the capabilities of robots. Or, as in the case of aforementioned motor, reducing backlash. 

best regards!

M.

[andrei66]

Hi again, I managed to make everything work with the L298N driver.

The PWM signals of the MCU are given straight to the driver via jumper pins, the motor wires were spliced and screwed into the OUT terminals, while power comes from a 4S LiPo 1300mah wired directly to the driver.

The speed and power difference from previous 10V is significant. I am very pleased with the result.

All in all, I can say the 2 to 3V of voltage drop is benefcicial in this case, since there is less of a chance of burning the motors.

Here are some pictures:

https://imgur.com/a/yihaVEC

After running it for a while, the motors barely get warm, only the driver's radiator heats up. Haven't had any issues except for the occasional clicking gears, but I think that is to be expected since the power output is increased.

Thank you all for your suggestions and ideas!

Edited December 29, 2020 by andrei66

[biasedlogic]

9 hours ago, andrei66 said:

Hi again, I managed to make everything work with the L298N driver.

The PWM signals of the MCU are given straight to the driver via jumper pins, the motor wires were spliced and screwed into the OUT terminals, while power comes from a 4S LiPo 1300mah wired directly to the driver.

[...]

After running it for a while, the motors barely get warm, only the driver's radiator heats up. Haven't had any issues except for the occasional clicking gears, but I think that is to be expected since the power output is increased.

Glad to hear that you managed to get it running!

The heating up of the driver is not surprising - you are losing likely 20-60% of power there, depending on operating point (remember: it's a dinosaur...).

What I can't see from the photos is where did you lift the ground reference from? In the photo there are only the two PWM signals visible. Or is the Hub powered from a common battery (+step down) with the motors?

[andrei66]

On 12/30/2020 at 12:19 AM, biasedlogic said:

Glad to hear that you managed to get it running!

The heating up of the driver is not surprising - you are losing likely 20-60% of power there, depending on operating point (remember: it's a dinosaur...).

What I can't see from the photos is where did you lift the ground reference from? In the photo there are only the two PWM signals visible. Or is the Hub powered from a common battery (+step down) with the motors?

The lipo is connected directly to the driver, and in the same terminals I routed 2 wires that connect to a regulator that provides 10V for the hub. The ground is common there.

These were previously connected to the lipo balance and gave power to all motors when there was no external driver.

And Happy New Year to all of you!

[biasedlogic]

On 1/1/2021 at 1:23 PM, andrei66 said:

The lipo is connected directly to the driver, and in the same terminals I routed 2 wires that connect to a regulator that provides 10V for the hub. The ground is common there.

These were previously connected to the lipo balance and gave power to all motors when there was no external driver.

And Happy New Year to all of you!

That sounds good!

Happy New Year to you too!

[AntonV]

Friends, I have Lego City train with power function 2.0, today, with the train platform connected, I turned on the motor mode, when I pressed the plus - full throttle (not the first speed as usual) when released - the brake. I could not repeat it, maybe someone knows how to switch to this mode?

[Tcm0]

1 hour ago, AntonV said:

Friends, I have Lego City train with power function 2.0, today, with the train platform connected, I turned on the motor mode, when I pressed the plus - full throttle (not the first speed as usual) when released - the brake. I could not repeat it, maybe someone knows how to switch to this mode?

Did you use the remote or the app?
Normally the other motors have full speed - break as their mode. Maybe the motor was detected incorrectly.

[biasedlogic]

2 hours ago, AntonV said:

Friends, I have Lego City train with power function 2.0, today, with the train platform connected, I turned on the motor mode, when I pressed the plus - full throttle (not the first speed as usual) when released - the brake. I could not repeat it, maybe someone knows how to switch to this mode?

Not possible.

This must have happened due to intermittent bad contact on id pin.

If you want this mode you have to modify the motor by soldering in a small resistor. It will then stop being identified as train motor and instead will be recognized as simple (WeDo-style) motor. You will lose the gradual speed control and gain the bang-bang type of operation. From the app level there's no difference in operation

[AntonV]

10 hours ago, Tcm0 said:

Did you use the remote or the app?
Normally the other motors have full speed - break as their mode. Maybe the motor was detected incorrectly.

Use remote.

9 hours ago, biasedlogic said:

Not possible.

This must have happened due to intermittent bad contact on id pin.

If you want this mode you have to modify the motor by soldering in a small resistor. It will then stop being identified as train motor and instead will be recognized as simple (WeDo-style) motor. You will lose the gradual speed control and gain the bang-bang type of operation. From the app level there's no difference in operation

maybe I accidentally pressed the combination when turning on remote(+ stop + stop etc) or  maybe you are right, this intermittent bad contact (but the wire was visually without damage and I did not touch the connector after that, but when i off/on mode changed, it`s very strange).

one another version I have two remote and train, son tried to turn on his remote and train with me.

Thx for answers.

[Thyraz]

As a big fan of @Mr Hobbles node-poweredup, I started to mix it with other useful nodejs / Javascript libraries lately.
Really love the possibilities this opens up. 

I thought I try to share some of my tests and start with using the browser gamepad API in combination with node-poweredup.

Most of the stuff can already be done with Powered UP app + Powered UP remote or with BrickController2.
But it gives you the best of both worlds:

- You're free to code your own sequences that can be started with the gamepad buttons.
- You can use analog sticks for fine-grained control.

I already tested this with some Robots based on the Lego Boost Hub and 2 additional motors,
and the mix of good manual controls in combination with automated movement sequences is a lot of fun. 

There's a first version of the sample project available here:
https://github.com/Thyraz/PoweredUpGamepad

I haven't added a readme, but installation should be quite easy if you're familiar with node-poweredup.
(Otherwise look at the node-poweredup repository for necessary installations step on Windows/Linux. OS X should be fine out of the box.)

- Install node.js
- Download the content of the repository to a local folder
- Run npm install in the directory to install the dependencies
- Run npm start in the directory to start the software

A node.js express server will be started that loads node-poweredup and it will automatically start chrome and connect to this node.js server.
The browser frontend will ask to connect a gamepad and send gamepad events like pressed buttons or positions of the axes to the node.js backend.

There's an example program.js in the package that initializes the Powered UP hub and uses the gamepad events to control it.

Feel free to ask questions if you're interested to use it.
This might help me to write a proper Readme on Github. ;)

[Thyraz]

Another question as some of you posted questions about connectors / extension cables:

As we didn't see extension cables released, even after the new mindstorm set arrived, I don't think Lego will release something like that in the near future.
Another idea I had was to cut the cables of the motors in the middle and attach some small connectors (male + female).
Normally you can just leave these connected to stay with the default length.
But you could use self made extensions using flat ribbon cables and these connectors types instead of the ones Lego uses.

Does anyone know good and small 6 pin connectors?

JST-SM connectors can be crimped for both, male and female and have some protecting plastic around the pins, so they won't break too easy.
And they are cheap and easy to handle.
But not sure how durable they are.

Edited February 25, 2021 by Thyraz

[Tcm0]

There are currently some adapter cables available on aliexpress: https://de.aliexpress.com/item/32871195908.html

Also, I hope that we'll see something in that direction when lego makes a new renewable energies set for education.