Buck-Boost Wrangling: Part III – Success!


IMG_0570     IMG_0573

Same setup, nothing different, except… the protoboard.  Had I not decided to switch boards (the one I was working with was too heavy to be convenient), I would have never known.  I am frustrated because I essentially burned a week unnecessarily traveling down a rabbit hole.  But, it works.

Before, I was incredibly concerned because I had seen a comment in one of the DIY articles that mentioned the USB’s data pins had to be engaged, which I found strange.  Good thing it was a non-issue.

MAKE excerpt

… also it won’t charge an iphone unless you have the correct voltage ref on the data lines…. other USB devices like Apple products also need correct voltage ref on data pins

Next steps for now till Friday: Integrating the mechanical and electrical systems.  Karen, Allieberry, and Shakeena will be presenting at ISTE 2015 (International Society for Technology in Education).  We’ve got quite a bit of work to do.  Just on the EE side, I have to install the buck-boost, solder and heat-shrink all the joints, and test.  However, all this is on pause until Lindsey prints all the parts for me to install the EE system onto, which will happen at some point tomorrow.

Prior: Buck-Boost Wrangling II

Reverse Engineering


Harnessing vehicular motion to charge small electronics is not a new concept by any stretch of the imagination, current EE wrangling challenges aside.  Examples courtesy of Instructables, MAKE Magazine.

There’s a difference, though, between something that is hacked together versus a product that is polished in all senses – designed for mass manufacture, affordable, reliable, aesthetically appealing, easy to use (and perhaps troubleshoot).  That’s what we’re striving to develop.

There are a number of stakeholders involved:

  • Karen
  • Shakeena and Allieberry
  • Autodesk
  • Simeon the toy designer
  • Lindsey and myself
  • EASE Lab

Axle Stability Brainstorm

We went to the National MakerFaire in DC and saw some really cool projects and products.

We didn’t actually have the gearbox working by the time we got to DC because the gears didn’t mesh. I was able to get that working yesterday when I figured out what gear modules actually were. There’s a formula that relates gear module to number of teeth and reference diameter.

Module = (Reference Diameter in mm) / (# of teeth)

I knew the number of teeth and the module of the gears I had so I was able to use the reference circles to make a gearbox that actually worked.

It looks like I screwed up the Faulhaber motor we were given by getting some adhesive into the internal shaft area. This is pretty unfortunate given that this motor was around $50. But it also forced me to find a new motor to work with. Hooking up a cheep motor we found in a box of stuff that Erica gave us to the gearbox actually showed promising results with the output current and voltage.

Spinning the input shaft of the gearbox by hand gave us around 0.7 volts and around 0.5 amps directly from the motor. This is pretty impressive for just hand speed. It looks like this could be a viable product after all.

The next issues I’m addressing is the stability of the entire device. Currently, there is no single shaft that extends through the entire device holding it all together. There are two separate stationary shafts that have a spinning component between them. This causes a discontinuous line of stability, as Myles* calls it, and could lead to failure at that discontinuous point.

(Above: The grey components are spinning while the red, blue, and yellow components are stationary. There’s no stationary component that runs continuously between the grey end caps.)

After some brainstorming with Myles, we decided that it was impossible to make a continuous shaft down the middle of the device while still having the input to the gearbox be in the center of the wheel. There would be interference with the spinning and stationary components that would be very complicated, if not impossible, to resolve.

So we came up with a solution to offset the input shaft of the motor and allow for a continuous support shaft through the entire device. It involves a bearing mounting inside a gear to allow for a stationary shaft to pass through a gear that is mounting rigidly to the spinning wheel of the EnGen. If this sounds pretty sketchy, it’s because it is, but I think there is some good reasoning behind it.

Having an extra set of gears would allow for another reduction and could produce even more power from the device. And pulling the gearbox input away from the center would allow for a continuous support to run through the device.

But this could also add an extra layer of complexity that we just don’t have the time for. The biggest gears we have in stock are too small to fit around the bearing we have and downsizing the bearings would reduce the shaft size, lessening any strength we’ve added to the device. We would have to find bigger gears and then bore out a hole for the bearing to be pressed into. This adds cost and labor to the manufacturing process.

So I’ve got a plan: I will continue along with my current prototype – the one with a discontinuous support system – until we get some sort of working model of it. Liani is still playing around with the electronics but pretty soon we’re going to need to test this baby out. If our first prototype doesn’t show signs of weakness at the discontinuous point in the support, we may not need this offset input idea anyway. Although, my guess is that there will be some obvious weakness and that this idea will have to be implemented eventually.

*Myles Cooper is a rising senior at Olin College who is working in the same lab space as Liani and I.

Starting a Linear Gearbox

Today I CADed a linear gearbox for the motor. I found nylon gears in the ME stockroom and was able to create a 15:1 gearbox in two stacked stages. These are so much better than the planetary gearbox I was trying to create from scratch.

Processes I tried for making my own gears:

  • 3D printing
    • Absolutely sucks for the small teeth profiles I needed. The profiles weren’t even there. They just sucked. Nothing meshed. Nothing fit. Ugly, awful gears that were only saved from being thrown in the trash by the wish to document the process.
  • CNC routing
    • Worked a whole lot better than 3D printing gears but still kind of sucked. The profiles were a lot nicer but it took forever to cut the gears out of 1/8″ plastic. They meshed and fit together very nicely but the time put into setting it all up was way too much.

And then I found nylon gears. So pretty.

I also found some dowel pins that pressed into the gears to get a good, solid axle.

These gears were bought in bulk by someone (no idea who the actual manufacturer is) and can be used in a mass production setting. I like these a lot better than anything that I was able to create myself.

The only issue I had with these gears was getting one of them to fit on to the motor that was provided for us. The hole in the last 10 tooth gear was too big to be pressed onto the motor shaft. It was down-right sloppy. I was hesitant about gluing the gear onto the motor because this is a $50 super fancy motor. But adhesive is what I ended up using.

I used Permatex medium-strength threadlocker (blue) to attach the gear to the motor shaft. The threadlock created a film around the motor shaft that helped to center the gear on the motor shaft. It didn’t take very long to cure and seems to be holding fairly well. I’m sure that if this ends up being the wrong gear to use, I can force it off the motor shaft without damaging the motor.

Currently, I’m designing a housing to hold the gears in place. I’m planning on 3D printing the housing but it could easily be injection molded if this were put into mass production. The entire gearbox will be around 2 inches long which is considerably smaller than the planetary gearbox monster I previously made.

The issue I’m having right now is trying to figure out what the pitch diameter is of the gears I found. I couldn’t find spec sheets on the gears so I have absolutely no idea what any of the designed dimensions are. I’ve contacted Nagy Hakim, who is the ME stockroom NINJA, about getting specs for the gears and I’m awaiting for a response. If he doesn’t get me that information in the next day or so, I’ll just do some trial and error designing to find the correct center-to-center distant for these gears.

Buck-Boost Wrangling II: Picking Erica’s Brains

This past weekend, Lindsey and I took a day trip to Washington DC for National Maker Faire.  We connected with Erica and met Karen, Shakeena, and her twin.  (We did not meet Allieberry, the other girl, because she had last minute commitments.)

I spent some time before our 1:30 auditorium presentation slot with Erica, who is an electrical engineer by education, attempting to debug further, but no dice.

This week, I’m going to try straight up testing from the power source unit to phone, and see if the phone can draw the appropriate current.

Prior: Buck-Boost Wrangling I.  Continued: Buck-Boost Wrangling III

Buck-Boost Wrangling I: Part Selection and Testing

When I joined the project on June 1st, Lindsey had already been working on the mechanical aspect for a couple of weeks.  The electrical system had only been discussed in very high-level terms.  By this point, Erica had suggested using a buck-boost converter circuit.  A buck-boost circuit takes some voltage input and spits out some voltage output.  The natural acceleration involved in riding a scooter would result in variable voltage when backdriving the motor; the buck-boost would act as a voltage normalizer.  For context, a cell phone charges on 5V and anywhere from 0.5-0.9A.

This naturally leads to the QUESTION: Which buck-boost converter should I use?


Choosing a buck-boost:

Quick Google searching resulted in a handful of options, but I decided to test use the LM2596 from Amazon because:

  1. Two-day Amazon Prime shipping = reduced lead time.
  2. Not only would it come in two days, I would also be getting a 2-pack.  Spares are always a good idea.
  3. The input/output voltage range and output current seemed appropriate.  Spec sheet here.
  4. There was a potentiometer onboard so I could customize Vout. Hooray for debugging tools!

Testing the buck-boost:

I didn’t have high hopes for the buck-boost converter, but testing was particularly frustrating.

This was the setup:

WP_20150618_001     WP_20150612_003

 (Right: The computer is not the power source; the power source unit is outside the picture frame.)

I set the power source to 5V and… nothing happened.

Debugging Step 1:  Tweak parameters and probe

I trust the power source to be outputting the correct voltage, so I connected a multimeter to measure the current drawn by phone.  I slowly tweaked the voltage supplied to 12V.  The buck-boost did its job and outputted 5V all the way, but still no current was drawn.

Debugging Step 2: Return to known system

I know that if I plug my phone directly into a computer, my phone charges.  I disconnected my male / female USB ports from the protoboard and reconnected to the computer’s USB, adding the multimeter to get a reading.


Plug in phone, and… no dice.

Disconnect multimeter, plug in phone – still nothing.  Conclusion – borked USB cable.  Find another male/female, strip, splice, heat shrink.  Repeated this process twice until I found a working USB.

The multimeter read 5V and 0.4A.  This is my baseline.  The question now: was it simply the USB cord that was messing up?

Debugging Step 3: Add a layer of complexity

Move system back to protoboard.  This time, the multimeter registered a current draw of 0.1A, which stayed consistent no matter how I tweaked the voltage.  The USB cable is functional; therefore it must be the buck-boost converter that is limiting the current draw.

I was baffled, though, because the LM2596 chip was spec’d to allow a draw of maximum 3A.  I was at an impasse, and decided that I would pick Erica’s brain about the matter the next day at the National Maker Faire.

Continued: Buck-Boost Wrangling II

Issue with Parametric modeling in Fusion 360

Issue encountered with Fusion:

I had already created a wheel housing for the 3rd version of the EnGen scooter when I created a linear gearbox to go inside. I created the gearbox in a separate Fusion file so that the part was clean and organized. When I inserted the gearbox into the wheel housing, I found it was too large for what I had previously designed.

This wasn’t supposed to be the end of the world – just make the housing larger and slip the gearbox in. But the size of the housing was determined by the very first sketch I made in this Fusion file.

Changing the height dimension from 50mm to 55mm suddenly broke a series of features in the timeline.

I had to roll back to the first error and start editing features and sketches until I was back where I started.