Search This Blog

Friday, May 22, 2015


Haven't had time to work much on my rocket. The bigger problem is the space. My small apartment doesn't have much room for hobbies :(.

Thesis research is eating all of my time. Hundreds of hardware and software problems have put me behind schedule about 3 months, pushing graduation off another semester.

Exciting stuff in the news, though. Apparently I'm a trend setter. There are some commercial electric hubmotor longboards hitting the market now!  M1 (formerly Monolith)MellowNext Board, Eon, Stary, Acton Qu4tro and Blink. The Acton Qu4tro is really the embodiment of what I set out to achieve with ELB (which ended up being a rather poor prototype). A few of the others are what I hoped to achieve with LiteBoard. There's even something like EHB now. Hopefully what I tried and documented helped them avoid the same mistakes. Send me free boards please? Haha. There have also been a bunch of other personal hubmotor longboard builds in the past few years. There's also a large DIY movement that's happened in the past couple years, particularly on Electricskateboard builder forum and electric spheres. Very cool stuff. There's an open source FOC ESC now, and tons of ready made hub motors for sale.

I'll get around to finishing LiteBoard and EHB one day. Some of these new designs coming out will give me some ideas for improving their designs. One of them has a neat design for screwing the hub onto a standard axle...will have to look into that more. Considering how well the technology is maturing, I may just buy one instead of finish LiteBoard. I'm sure I can use the parts for something else.

Sunday, January 11, 2015

L3 rocket update

I managed to get some work done on my L3 rocket today.

The centering rings' ODs were about 1/32" too large. I used a dremel and a sanding block to sand them down. I also drilled and mounted the U-bolts in the forward centering ring.

Marked the hole locations first. The X markings are to remind me to not line up the
U-bolts with fins because the front fin tabs will butt up against the back of this ring.

CA used on all nuts for anti-vibration.

I completed most of the nosecone modifications today. As mentioned in the last post, the goal is to be able to adjust the mass in the nosecone easily. I cut the bottom off of the nose cone and sanded it. I'll be using a ~5" diameter, 1/8" thick plywood bulkhead to brace the 1/4-20 threaded rod that will be glued into the tip. I drilled a 1/4" hole in it for the threaded rod and sanded a chamfer in the outer edge so that it would sit in the cone better. The threaded rod was cut to length so that an inch or so would stick above the bulkhead for attaching an eye nut for shock cord attachment. I also drilled some small holes in the bulkhead and superglued in some thread so that I could pull it out. This bulkhead will not be glued in place; instead, it will be held in place using nuts and washers. This allows me to add and subtract mass (in the form of 1.25lb lifting weights) from the nose tip be sliding weights on or off the threaded rod. The weights will be held in place using large washers and more nuts.

I then marked and drilled holes for 3/16" wood dowels to go through the nose tip. These will help hold the threaded rod and the epoxy plug in the nose.

I had to dremel these because I didn't get them straight the first time.
Markings for the second set of holes. I realized the lower
(original) markings were too close to the other holes.
Both sets of holes drilled. I also sanded the inside of the nose
 tip to help with epoxy bonding.
Supergluing washers to nuts. Probably should of bought flange nuts.
After many test fits, the final locations for the nuts were set with CA.
5 minute epoxy + silica was used to glue the dowels in.
These will be cut and sanded flush later.
Inside view before potting with epoxy
Washers help hold the threaded rod in.
Camera didn't resolve threads for some reason.
For potting the nose tip, I used 3x pumps of West Systems 105 resin and 206 (slow) hardener mixed with about a tablespoon of colloidal silica. I carefully poured it all into the tip, avoiding the threaded rod. I then shook/tap/hit the nose to get the epoxy into all of the crevices. After that, I poured a few ounces of steel BBs into the tip and repeated the settling exercises. After that, I put the bulkhead in and placed the nosecone upside down in a piece of body tube. It takes a few hours to cure, so I'll check it tomorrow. After this, I need to cut the dowels, sand and fill the tip, prime, and paint it with high temperature BBQ paint.

Left to do:
  • Finish nose cone
  • Purchase a bunch of stuff
  • Make motor mount
  • Install fins
  • Filleting 
  • Build altimeter bay
  • Minor stuff
  • Final primer coat, sanding, then painting
  • L3 paperwork

Joule Thief Night Light Update

I plugged it in and measured voltages across some of the components using a super cheap volt meter. With black tape on the photoresistor (LEDs on): (all approximate)

  • across projector LED: 2.8V
  • across other LED: 2.7V
  • across proposed joule thief location (see last post): 6.25V
  • across 1Ohm resistor: 0.017V
  • across 33Ohm resistor: 0.57V
  • across photoresistor: 1MOhm
With the tape off (LEDs off):
  • across proposed joule thief location: 1.1V
  • across 33Ohm resistor: 0.6V
  • across transistor (collector-emitter): 0.5V
  • across photoresistor: 4kOhm, 0.43V
What this is telling me is that some sort of voltage regulation is happening. I think it has something to do with the zener diodes. The AC-DC converter looks a lot like typical capacitor (transformerless) DC power supply, but with zener diodes as part of the rectifier. Looking at the circuit diagram in last post, can anyone tell me how this power supply works?

I was close with the 5V guess. It's actually around 6.2V when the LEDs are on. But the problem is that this drops to 1.1V when the LEDs are off (light room). While that's good for power dissipation while off (estimating about 20mW using 33Ohm resistor current and 1.1V), it means that if I have a 6.25V joule thief hooked up to it in a light room, a lot of current will have to be dissipated by the 33Ohm resistor and the transistor, which is bad.

One (very non-elegant) work around would be to cut a trace, have a way to switch between JT or AC operation, have a transistor inline with the JT, and a tiny microcontroller with leads spliced across the photoresistor. In JT mode, the microcontroller would read the photoresistor and tell the JT transistor to turn on or off. In AC mode, the JT would be taken out of the circuit by a switch and it would operate as it was designed to. 

Another thing I might be able to do is splice off of the photoresistor-base transistor connection and use that to turn a transistor in the JT circuit on or off directly. 

I'll have to think about this some more. Figuring out how the power supply works would probably help. 

To do:
-learn how the power supply works
-design JT circuit for 6.25V and ~20mA
-figure out efficient way to integrate JT circuit 

Saturday, January 3, 2015

HP Split 13 X2 (product 13t-m000 or E1N79AV) SSD Replacement

This process has been such a pain in the ass, so hopefully this will save you some stress.

The HP Split has two hard drives. A small SSD (64gb) with the OS in the tablet portion and a 500gb hard drive in the dock. I wanted to upgrade the SSD in the tablet with a larger one. I also didn't want to reinstall the OS or the many programs I had on this computer. I bought a Samsung 840Evo 250gb mSATA SSD for a very reasonable price off Amazon. I also bought a USB 3.0 mSATA bay so I could clone the drive (even though this computer doesn't support USB 3.0, I'll use it later). I backed up all of the files first, then started the process.

First problem: The OS is Windows 8. I never got around to upgrading it before. I tried 4 or 5 times to get Windows 8.1 to complete the upgrade, but it never would. I tried just about everything I could find online: did all updates (OS and drivers), uninstalled antivirus software, etc. I eventually gave up on this. I'll try again one more time after I get the new SSD installed.

Second problem: The new drive (in the USB bay) wasn't recognized in windows explorer. You have to go into disk management and format the drive, then it will show up.

Third problem: Samsung has free migration software available on their website. Turns out it is total shit. It failed 1/4 the way through cloning. I found the EaseUS Todo Backup Free software online that seems to be fairly adware free. I installed it and cloned the C: partition. I then switched the SSD's.

Fourth problem: It wouldn't boot after the switch. Shit. I switched them again, windows ran a system recovery, and I got back to where I started. I reformatted the new SSD. I then read a bit more. Turns out you might need to clone the whole drive including the tiny partitions, so I did that.

Fifth problem: That didn't work and now the original SSD won't boot at all. It says I need recovery media, which of course I don't have. So I've actually gone backwards. Damn it.

I did a factory reset of the original SSD. This required hours of installing updates. I was then able to update to Windows 8.1, which means some piece of software was preventing me from updating the drive to Windows 8.1 earlier. I then installed missing drivers from HP, set a system restore point, installed EaseUS Todo Backup Free again, and attempted another clone to the new SSD. I first tried booting off of the new drive while it was still in the USB bay, but that didn't work. The computer restarted and booted off the original drive. At least the original drive didn't fail this time. Next, I just replaced the original drive with the new drive and unplugged the USB bay to see if the computer was getting confused by having two identical copies of Windows. It booted up! YAY! I had to resize the volume of the partition under disk management to make full use of the volume. Now I have a hybrid laptop with ~230gb SSD and a 500gb HD!

So, steps you should follow if you want to do something similar:
1. Backup your files and programs.
2. Set a system restore point.
3. Reinstall the OS (I did a factory reset). I couldn't update to Windows 8.1 nor make a successful clone due to some (unknown) software preventing it from working.
4. Perform all software and driver updates
5. Set a system restore point
6. Clone the whole drive with windows on it. I suggest staying away from manufacturer provided migration software. There are a lot of good, free cloning software out there.
7. Swap the cloned drive with the old drive.
8. Boot.
9. Install programs and put your backed up files back on it.

Total time spent upgrading to a new SSD: 12 hours

Good resource:

Friday, January 2, 2015

L3 rocket update

I haven't done much fabrication. I installed the AeroPack 98mm retainer on the rear centering ring. This was kind of a pain in the ass. I tried the method in the instructions on friend's centering ring last year. It suggests holding the retainer aligned with a motor casing, marking the holes, then drilling. The problem is that getting the marks for directly in the middle of the holes is pretty much impossible. This time, I CAD'd the AeroPack retainer and the centering ring, assembled them in CAD, made a drawing of the assembly with hole center marks, printed that out, cut out the center, and taped it to the centering ring. I then punched the center of the holes and drilled them. It came out much better this time, so I did my aft centering ring next.

The holes match up very well.

Screwing in the thread inserts.

Tapped over retainer to prevent epoxy from gluing the retainer to the centering ring.
My friend has a fiberglass aft centering ring (the one I messed up the first time around). The instructions suggest tapping the holes, but I don't think the fiberglass will hold threads well. We'll be using steel nut plates glued to the back of the centering ring to provide threads.

I also cut the motor tube to length and sanded it. Because phenolic tends to fray after awhile, I treated the ends with CA. I did the same to the two tube couplers (I did this to the body tubes earlier).

I did some modelling Modelling and Design updates. I took out the 12" section of body tube and one of the couplers from the stretched version of the rocket (which is the one I'll be using for my L3). The 12" section will be used instead of the 36" section for the scale (non-stretched) version.

Updated CAD
Because my Dad wanted to buy a CTI 5G 75mm case, I've chosen either a CTI M1400 or M1401 as the cert motor, and updated the OpenRocket simulations accordingly.

1/2 Scale
It predicts about 7500ft on the M1400.

Avionics bay design: The avionics bay end caps will have an airframe bulkhead and a coupler bulkhead glued together. Each end cap will have an eyebolt for recovery system attachment (see below). Two 1/4" threaded rods will run through the avionics bay and end caps to link the two end caps together and provide rails to mount the electronics sled. Two altimeters will be mounted on the electronics sled. The end caps will have high temperature screw terminals and charge holders glued to them. The vent hole(s) will be drilled through the upper body tube section and into the avionics bay. The av bay will be secured to the upper body tube section with 4-40 stainless steel screws. Overall, a pretty classic design.

Nose cone design: While running OpenRocket, I realized that, if I wanted to be able to use the same nose cone for both the stretched and non-stretched versions, I would need to make the nose weight adjustable. I'll show pictures of this later when I actually make it, but the plan right now is to epoxy a 1/4-20 threaded rod into the tip of the nose cone with a eye-nut on the end of it to attach the shock cord. I'll mix steel BB's and chopped carbon fibers in with the epoxy to increase weight and strength respectively. I'm also planning on drilling through the tip of the nose in a few places and running wooden dowels through in order to help anchor the tip/threaded rod in place. It'll make more sense once I post pictures of the process. Weight can be added to the 1/4" threaded rod as required.

The recovery system will be as follows: All connections are done with 1/4" quicklinks unless otherwise stated. A Kevlar Y-harness connects to two 1/4" U-bolts in the forward centering ring. The Y-harness connects to a ~20ft long Kevlar or tubular nylon shock cord. The drogue parachute (~48") is attached to a loop in the shock cord about 1/4 of the way from the nose/payload section. The other end of the shock cord attaches to a 1/4" forged, galvanized steel eyebolt in the avionics bay. There is an identical eyebolt on the other side of the avionics bay. This connects to a shock cord similar to the previously mentioned one. The main parachute (Sky Angle Cert-3 XL) attaches to the shock cord about 1/4 of the way from the nose cone. The nose cone will have 1/4" forged, galvanized steel eyebolt in it that attaches to the other end of the shock cord. Nomex cloth protectors will be used to shield the parachutes and shock cords from the ejection charge gases.

Hardware acquisition: I got two PerfectFlite altimeters for Christmas, so that covers the electronics. I managed to find some of the hardware I needed from Home Depot (U-bolts, threaded rod, etc.). The rest I'll get from McMaster.

Time outlook looks abysmal...I kinda have a thesis I need to do. This rocket may not get finished for another year :( .

Left to do:
  • Purchase a bunch of stuff
  • Sand centering rings
  • Make motor mount
  • Install fins
  • Filleting 
  • Build altimeter bay
  • Minor stuff
  • Final primer coat, sanding, then painting
  • L3 paperwork