Skip navigation
1 2 Previous Next

Make Life Accessible

19 Posts authored by: dougw Top Member

This update to the Clear Walk project is to demonstrate its performance in the snow, but first I should show the android app I made to run this test. Note it is in the style of Star Trek LCARS because the Clear Walk system is also the "Main Deflector" in my Star Trek Alcove project.

ClearWalkApp

I have been itching to try out the Clear Walk system, but it never seems to be sunny when I have time to work on it. We have lots of snow, but today was the first day that had sunshine while I was at home. So I went down to the basement to bring out the apparatus only to discover it was so large it couldn't  quite fit up the stairs. After removing the solar panel and rotating the mirrors in both axes several times during the trip, I was able to negotiate the wrap-around staircase and get it outside.

ClearWalkOutsideYou can see from the shadows the sun is at a low angle - lower than I had expected when designing the system in the summer. This caused some issues getting the mirrors to tilt down enough, but I was able to tilt the apparatus enough to try the following experiment, melting snow and ice from the steps. The air temperature is -4C and this is also the temperature of the snow in the yard.

This sequence of pictures was taken over the course of about 25 minutes...

snow0 snow1

Canadians don't generally get the "Mind The Gap" welcome mat, but Londoners will certainly know where the phrase comes from.

 

The brick in shade is at -3C

The brick in direct sunlight is at +1C

The brick in sunlight plus mirror light is at +16C

The snow in direct sunlight is at -3C

The snow in direct sunlight plus mirror light is at +1C on the surface

The 2 bright stripes are from the 2 mirrors.

snow2 snow3

snow4 snow5

snow6 snow7

snow8 snow9

Note - I went back for a second pass to melt the snow at the end of the black mat.

snow10 snow11

snow12 snow13

The brick temperature in combined sun and mirror light had risen to +26C by the end of this 25 minute sequence. The snow in shade was still at -4C and the snow in the yard (direct sunlight was at -3C.

It takes a while to melt snow even when the temperature is well above zero, but the Clear Walk mirrors definitely made a difference in heating the bricks up. I discovered an amazing phenomenon - when the brick was warm but just still slightly wet and the air was still, remaining moisture would suddenly evaporate in a little cloud of water vapour. I tried to get a picture of this happening, but it is not visible in a still image. I guess the moving vapour (as it rises) is much easier to see.

 

Observations and Notes

  • Increasing the amount of sunlight reaching a location using mirrors does help raise the temperature enough to melt snow, even when the outside temperature is below zero.
  • The mirrors need to be pretty large to make a big difference - it is obvious from the images above that the fairly large mirrors only resulted in fairly narrow strips of extra light.
  • The mirrors must be able to tilt to significantly negative angles (pointing down) to focus light on the ground.
  • The quantization of movement needs to be small - the system was set up to have much coarser vertical rotation than horizontal. Horizontal quantization was fine, but vertical was a little too coarse.
  • My walkway is not ideally positioned to make optimal use of reflected light - a second set of mirrors might improve efficiency.
  • If you build one in a basement, make sure it will fit up the stairwell.

 

Conclusions

The Clear Walk system works, but it is more of a proof-of-concept system than an optimized solution. I think a simple manually positioned mirror resting on the ground would work just as well without as much danger of blowing over.

I did try beaming sunlight into a window, and that works well.

Will I end up using it? Probably not much - it currently needs some supervision and the number of days where conditions are ideal is quite limited. There are a few improvements I could make, such as better granularity of vertical motion and a more acute downward tilt angle, but I don't think the effort is worth it - I would rather design a better system at ground level.

Standing around waiting for snow to melt is like watching grass grow - not terribly exciting. But while standing there, I came up with a new use for the system that may prove very useful. Our house has a tendency to grow icicles off the roof directly over our front stoop. They are difficult to reach to remove, but the mirrors may do a great job of melting the attachment point to the roof.

 

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

I have to squeeze in one more update before I have to drive to Boston. The video below shows the fully functional Clear Walk system, with mirrors installed,  capable of redirecting sunlight anywhere it is needed under remote control. It could be an icy walkway or driveway that needs to be cleared, or it could be an indoor plant (redirecting sunlight through a window), or it could be some laundry hanging on a line.

Here is a close-up of the electronics module showing the Bluetooth module, the arduino, the Dc-DC converter, and the H-bridge motor drivers all crammed into a neat little box:

electronics

It may not be too obvious from the videos but there are over twenty 3-D printed parts in the system, not including the parts that did not work out.

I haven't had time to reflect on this project yet but perseverance and adaptability are necessary when you run into a project like this. Although I have to leave town right now, for the critical last few days of the project, and consequently cannot do a grand finale, I will continue improving the system when I get back - until I can finally try it out on some snow. After all this work, I really want to see how well it performs.

 

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

The Clear Walk system is finally fully operational under battery power. I have packaged the electronics in a small case which can be mounted on the frame. You can see a demo of the mechanics in operation in the following video.

It has been a rocky project, with all sorts of setbacks and ornery issues. This project is characterized by large efforts to make small progress and overall a major effort trying to close on a workable solution. However, it looks like it is all coming together just in time to meet the deadline. It has been fun at least most of the time and I learned a lot. Overcoming the shear number of problems and the significant magnitude of some of them has made it a very satisfying result. It was costly, but could have been much more costly.

Unfortunately, my presence is required in Boston this weekend, so I will not be able to polish the project up as planned. This is a major disappointment after months of passionate work, but other priorities intrude.

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

It has taken months to get the Clear Walk mechanics and motors built up and functional, but only one evening to get the motor driver electronics working under Bluetooth control. Finding the H-bridge motor controller card amongst my old projects was quite fortuitous as the module I ordered has not arrived yet. It is proof that hoarding sometimes pays off. The video below is a first power up of a bread-boarded system with one motor under Bluetooth control.

The Bluetooth connection can be run from any Bluetooth terminal or maybe even a Bluetooth keyboard, but I expect to make a proper android app so sun tracking can be automated.

Next I will clean up the wiring and power supplies so it can be connected to the real Clear Walk apparatus for a proper demonstration of the full system rotating mirrors.

I am still planning to connect the Kinetis motor controller and Linix motor to demonstrate that they can easily handle this application. However it turns out they are overkill for this particular application, so I am implementing something more rudimentary first. The permanent magnet synchronous motor algorithms and programming tools from Kinetis are very sophisticated and I have learned a lot from experimenting with them. The main issue right now is building a suitable bracket to rigidly position the big motor and associated gear train while allowing belt tensioning.

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

This update demonstrates motorized azimuth rotation. Finally there is a fully mechanically working system. Only the mirrors need to be mounted, plus some weatherproof covers.

Here is a closeup of the azimuth motor. It has a pivoting motor mount that allows the belt to be tensioned easily.

I am still working on mounting the big Linix motor that took so much effort to get geared down, but I wanted to show the full mechanism as soon as possible.

Next I will work on some electronics to hook everything up.

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

 

Another intense week - including 2 all-nighters trying to get the drive mechanisms designed, build and assembled. The image below has eighteen 3D printed parts, and there are more not shown.

Progress in this update:

  • I have mounted the solar panel which will keep the battery topped up.
  • Designed and built the mirror frame from carbon fibre (hockey sticks) and 3D printed elbows
  • Designed, built and assembled the solar panel brackets
  • Designed build and assembled brackets to hold the elevation drive arm to the elevation pulley
  • Designed built and assembled the cross bar brackets to mount the mirror frame
  • Designed built and assembled the gearmotor chassis
  • Designed built and assembled the motor mount bracket and clamp
  • Installed the elevation drive belt and tested motorized operation
  • Bought a bunch more parts including the big mirrors

A lot of effort went into design. It may not be apparent, but the entire system is very well balanced and the mirror plane is coaxial with the elevation pulley axle - requiring very little motor torque to hold or move the mirrors.

ElevationMechanics

So far I have used almost a kilogram of plastic in my 3D printer - you don't realize the cost of printing a few parts until suddenly you have used a whole roll of filament.

The mechanism seems robust and stable although not quite as much travel as I had planned - making the brackets robust causes early interference.

Here is the elevation motor in action driving the mirror frame:

Here is another angle of the mechanism: You can just see the motor between the forks:

ElevationMechanism

Next I need to mount the azimuth drive motor and then hook up all the electronics and then get on with the software.

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

It has been a very busy week at work, plus working on 3 major projects at home, plus playing in a softball tournament - so 5 softball games and 2 hockey games, each taking several hours. I did manage to get in about 12 hours on the Clear Walk project, but sleep is hard to come by.

Designing and building the drive pulleys for the Clear Walk solar mirror has been quite an adventure, and a lot more 3D printing than I expected, and it isn't over yet.

I still need to mount the mirrors and the motors, but most of the drive train is figured out, and I think I have most of the hard-to-find components. I need. Still need a bunch of screws etc.

DrivePulleys

 

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

The drive train from motor to sprocket is finally approaching a viable solution. This installment adds a flex coupler to allow some eccenctricity between shafts and minimize any resulting vibration.

FlexCoupler

It is getting closer to a full drive train for this motor. The next phases will involve building the pulley, mounting it to the steering column, mounting the motor/greabox to the Clear Walk frame and connecting  the two with a drive belt.

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

Still plugging away at getting the gear box aligned with the motor. I can sometimes get them lined up pretty well, but every time I make a tweak to try to improve it, it most often gets worse.

The video below shows significant misalignment, but it is useful to show what I am fighting.

The brass coupler I ordered a while ago finally came in so the connection to the gearbox is pretty clean. Also the aluminum sprocket for the gearbox arrived. It did not have a 6mm hole to fit the gearbox shaft, but I knew that when I ordered it (I bought this one because it was cheap) Fortunately it was very simple to drill the hole to the correct size, and the set screws still work perfectly.

The gearbox mount is bolted to the motor mount with large clearance holles, so it can be positioned vertically and horizontally to align the shafts. However this is a very touchy adjustment, being off by the thickness of a sheet of paper creates a large wobble. I expect I can get it aligned well enough to work - it only has to run for about 1 second every 10 minutes.

DriveTrain

The video is uneditied - no titles, not great lighting, just letting the camera do its thing. I am curious to see how it shows up online. I believe the camera is set to record 4K video. Normally I convert it down to lower resolution, but this is a short video.

My brother finally convinced me to use a lapel mic with the camera (by giving me one), so the audio in the video should be a little better as well.

 

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

I have been spending entirely too much time trying to get the drive train to be functional, but I still need to devote some time to describing what I have been up to. The Linix motor is great - lots of power, but it runs between 300 and 4000 rpm. I need this speed to be reduced by a factor of about 4000 to get reasonable control and accuracy in my solar tracking application.

There are nice gearboxes (like the one below) made for this particular motor, but I cannot afford them:

LinixGearbox I explored a lot of options as this can be an expensive aspect of the system. I procured several adpters, flex-couplers and gearboxes, to try to cobble a workable combination together, but every component in the system has its own set of problems, each of which requires time and usually money to resolve. The picture below shows a gearbox I obtained by butchering a perfectly good gearmotor, but of course the gearbox driveshaft is a different diameter from the motor driveshaft, so an adapter is needed. The commercial coupler I ordered has not arrived and might not arrive in time for this project, so I have been attempting to make a suitable coupler. The unconnected cyclindrical coupler in the picture caused huge vibration because it could not maintain accurate axial alignment. The clamping coupler shown connected, was a little better, but still not good enough. I hind sight, butchering the gearmotor the way I did was not the best solution. It left only a single bushing in the gearbox to control the driveshaft. And of course I bent the driveshaft a bit in the process - very difficult to straighten accurately.

It appears that I cannot hold tolerances accurately enough to avoid significant vibration, so I will have to try some kind of very flexible coupling. I expect that will take a few more hours and probably several iterations.

One other nasty issue is that all the nice dimensioned drawings of the gearbox do not show the exact location of the input driveshaft (because it is internal to the gearmotor) and it is not in the middle either vertically or horizontally. Which is going to make building a chassis a tricky problem.

Motor Gearing

The gearbox has a ratio of about 266:1 which is a good reduction but it is not quite enough - it still leaves the output shaft rotating at over 6 degrees per second. My plan is to put a pulley on the output shaft that will reduce the speed by 14:1. This will get the speed down to about 1 degree every 2 seconds. The sun is progressing much slower at about 1 degree every 4 minutes, so the motor only needs to run 0.2% of the time to keep up.

DrivePulleyThis image shows part of the pulley - I expect to use a separate take-up reel. The 2 clamps will be used to mount the reels to what is left of the handlebars described in earlier blogs. You can see the convoluted threading path for the belt internal to the reel. This is a very tight fit and does a good job preventing the belt from slipping. The aluminum drive sprocket is also shown in the picture just to get an idea of the gear ratio.

 

The project is proceeding too slowly despite putting a lot of hours into it, however I am learning a lot about what not to do. Some things like the pulley look like they will work on the first try while others such as the shaft coupler look like they will take many iterations. The drivetrain chassis is definitely going to be difficult.

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

This blog shows the result of a lot of pondering what materials to use to make a sturdy low-cost stand to support the rotating mirror apparatus, a lot of design and pondering how to build a suitable structure given my tools and resources, and some metal bashing to put it together. It is not painted and it is not finished yet, but the hard work is done and the unknowns about the stand are history. There is still a similar amount of pondering required to design and build the dual motor drive train and mirror mounts, and after the mechanical is complete the eletrical and software will dominate the activity. There are still some worrisome unknowns on that side of the project.

ClearWalkStand

The tripod stand has turned out to be nice and sturdy and the rotating bearings are positioned well, with no interference. So far I have managed to keep stand costs down to about $15. Although it does not include costing for apparatus that has not been assembled yet, this part is an order of magnitude less than some of the options I was exploring.

There is still a long way to go, but it is nice to start retiring some unknowns and see tangible progress.

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

This entry is a first exploration of the Kinetis Motor Suite capabilities, demonstrating motor reversing and sequencing through several states.

Next time I will try to hook up the Hall effect sensors to see if I can work at lower speeds.

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

It took me about 33 hours to get the motor running, mostly because I had installed the latest version of Kinetis Design Studio (3.2) but it appears that Kinetis Motor Suite doesn't work with the latest KDS and it wasn't until I tried the older KDS3.0 that things started to work. Here is a short video to show that the Kinetis Motor Suite does work if the right KDS is installed:

 

I will have to experiment with the Kinetis Motor Suite to see how it can be used to program the motor for my application. There is still a lot I don't understand about the Kinetis system.

I have also procured most of the chassis materials for my system stand, but it is going to be built from raw materials, so there will be quite a lot of metal bashing and woodworking to do.

Then I need to start getting a Bluetooth interface setup to control the system remotely.

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

There is an extensive step-by-step tutorial series on getting started with the  FRDM-KV31FFRDM-KV31F Development Platform here

FRDM-KV31F|Freedom Development Platform|Kinetis MCU|NXP

Following through the steps and installing all the software took a couple of hours on a very fast PC - lots of steps and lots of software. Here are some of the software packages that are part of the process:

Kinetis SDK 1.3.0 Mainline - Windows.exe (Kinetis Software Development Kit - 312 MB)

kinetis-design-studio_3.2.0.exe (Kinnetis Design Studio - 701MB)

Eclipse add-on to add Kinetis SDK V2.x Project Wizard.zip (Elcipse upgrade for the Software Development Kit - 243 KB)

KinetisMotorSuite.exe (Kinetis Motor Control Suite - 120 MB)

KSDK-Project-Generator.zip (Kinetis Software Development Kit Project Generator - 32 MB)

Additionally, some of these packages need to be upgraded immediately through an online process.

I was trying for a vanilla default installation using all the default options, but it did not go smoothly for me.

Before installation I had spent a lot of time to collect all the info I could find on all the hardware and software involved in the kit. One issue I ran into was documentation of the motor. After lots of searching I found several different datasheets on the motor, but none of them had any detail on what the 8 wires were coming out of the motor. Freescale had lots of documentation on how a 3 phase brushless DC motor works, but again, almost no information on wiring this particular motor up. They seem to use this same motor for many of their demo applications for other products, so eventually I came across a note that indicated it didn't matter which phase was connected to which terminal. This still did not help identify which 3 wires to use. There are 2 white wires, 2 blue wires, 2 green wires, a red wire and a black wire. I noticed that one each of white, blue and green wires were slightly heavier gauge so I am assuming these are the main motor phase wires. I gather from some of the docs I found that the other wires are probably 3 Hall effect sensor outputs (white, blue and green) and the red and black wires are 5 volt power and ground. I came across some images that have the red wire going to the ground pin and the black wire going to the power pin as well as some images where they are going to the phase sensor pins, so I still have no idea which wires do what. I am not at all comfortable connecting this motor up without better documentation - I doubt Hall effect sensors (assuming that is what some wires go to) will be happy with 24VDC on their outputs. Likewise if they had their power pins reversed, they probably wouldn't be in great shape.  Anyway, after several hours of scouring the web, I am stuck with guessing which 3 wires to use and leaving the rest disconnected. If anyone has a wiring diagram for this motor, I would dearly like to see it.

I have purchased a 150W DC-DC converter to supply 24 VDC from a 12 VDC battery:

DCconverterIt seems to work well, but I will know more once I get the motor working. It will require a 12 V battery capable of putting out at least 10A.

The 4800 mAH Li-Ion battery I have may not be quite up to it, but I want to test the actual requirements in my application before writing it off.

12VLiIonBatteryHere is my motor test setup showing a 700 Watt MFJ power supply substituting for the battery:

MotorTest

Okay, back to software installation woes. After 7 hours of struggling with it, I still have nothing working. The installation tutorial seemed to go okay - a bit fitfully, since the video was not clear enough to read any of the text on the screen in the video. Some straight forward steps were left as an exercise for the viewer, but resulted in discrepancies with the video, so had to be redone or corrected.

Part of the process involved upgrading the Kinetis Design Studio, which could not complete because of missing repository files. I could not figure out how to correct this so I proceeded, hoping the upgrade was not crucial. Unfortunately, it probably was needed. Here is the error:

UpdateError

When I got to the end of the process there were some exercises to try loading demo programs, but none of the examples were available beyond the driver libs. I didn't see anything that looked like an example with this motor control system.

Update: I did finally get the examples sorted out and running, but no luck with the Motor Suite.

So I skipped on to try the Kinetis Motor Suite. I gather the MCU module comes loaded with motor demo firmware, but after sorting out and configuring the serial port, the Motor Suite could not connect to the MCU module because the .out or .elf file in the Motor Suite did not match  what was in the MCU module. Here is the error:

ConnectIssue

I could not figure out how to get the .elf files to match, indeed I could not get the Motor Suite to find the right directory, possibly because there wasn't a .elf file in the right directory. This may all be due to the failure in upgrading, but whatever the reason, I have not been able to get past this failure to connect yet.

 

I have not been able to find any description of the pre-loaded demo firmware in the MCU module beyond that it will blink red and blue when powered up. (which it does) Nor can I find the source code for this demo.

 

One thing I noticed is that the MCU needs to be connected to a PC to run its demo program - simply connecting it to a USB power supply does nothing, not even an LED lights up. I will have to change this as my application is stand-alone - no PC in the front yard.

 

I really want to get on with building my system, but I have to get this basic stuff working. It is clear from the gigabytes of software I have installed and the many steps involved in software project development that I have a massive learning curve ahead of me. So far it is pretty frustrating, but hopefully I will start making better progress shortly.

 

The Kinetis Motor Suite (in the Project menu) has a menu option to "Show Path Selections".

This reveals that  "Kinetis Design Studio: Not Selected"

The Project menu also has an option to "Select Paths" with a drop down to select path for "Kinetis Design Studio" which brings up a standard "Browse For Folder" window.

Selecting the directory where KDS installed pops up this error:

KDSrooterr

I have tried every level of directory and this same error always pops up. The directory in the first line of the error message is the one I selected on this attempt and it is the directory where the install program installed KDS by default. The Kinetis Motor Suite just does not recognize the install directory of the Kinetis Design Suite or the location of the .elf file.

I have not been able to get any answers from Kinetis, but one note on one of their forums indicated that KMS does not work with KDS3.2.

I will try earlier versions this weekend to see if that works. 30 hours on this issue so far is putting a serious crimp in my development schedule.

.........

After installing KDS3.0 things started working a little better. I did have to set up the directory where KDS3.0 was installed, but then I could get the COM port set up and connected. The motor wouldn't do anything at this point, but eventually I figured out the little fault indicator was trying to tell me the DC input was over voltage - even though it was set to exactly 24 VDC.

I had to turn the voltage down all the way to about 12 VDC to get the fault to disappear. It gives an under voltage fault below 11 VDC and an over  voltage fault above 14 VDC. I'm sure this can be altered as the motor and drive module are rated to run at 24 VDC.

With the input voltage at 12 volts I was able to get the motor running. There is a lot of flexibility in the motor suite that is going to take some experimentation to learn, but I will start with a short video in my next blog showing the motor running.

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible

The project kit of parts to be used in the Make Life Accessible design challenge has just arrived, so I made an unboxing video to show what came in the box:

BLDC

KinetisM4

MotorControl

controllers

 

 

 

Next Steps

I want to power up the BLDC motor and controller to see if my battery system can supply adequate power.

I also want to get my bike frame mounted on a chassis and progress with linkages to drive motors.

 

Relevant Links

MLA Design Challenge

 

The full set of Clear Walk project blogs can be found here:

Make Life Accessible