Servo motor with screw drive
The electronics that control the stepper motors on my CNC mill were purchased from Ocean Controls (An Australian company) in kit form. Specifically, I purchased Continue reading
Y-axis Servo and pusher assembly
Since my last post I have constructed the Y table servo mount and pusher assembly. The Y-axis servo is mounted using a pair of L-Brackets, M4 machine screws, washers and nuts that were purchased specifically for this project and actually fit the screw holes on my servo! (Unlike the X-axis servo, which is connected with a 3/32 machine screw that doesn’t actually fit the threaded hole in the servo, and relies on friction to keep it from moving.) Since I was at the hardware store, I also bought an 8 foot piece of wood (2 inches of which I used for the pusher assembly with a lot left over for things like the X-axis limit switches.)
The video below shows the router drawing a diamond (twice as difficult as a square, because you have to make two motors move at the same time). I am controlling it using a small bit of python code to issue motor commands over a serial port to the stepper controller board. The diamond is drawn from 0 to 18,000 in stepper coordinates on each axis. (The 5.5 inch range of each axis ranges from 0 to 46,000 in stepper coordinates, so the working area is about 2.5 times larger than the drawn diamond.) As you can see from the (lengthy) video, the screw drive assembly on my (1.8 degree per step) steppers makes the actual motion speed of the X and Y axis quite slow. It is unlikely that I will have to artificially slow down my cutting speed when cutting actual material.
The device holding the sharpie marker is a (cheap, and inexpensive) “Dremel Drill Press” from Harbor Freight Tools that cost me $15. I can no longer find them on the Harbor Freight website, but they do have a “Drill Press Milling Vise” for $60 that is basically an X/Y table with 8 and 7 inch travel and a 5 inch jaw capacity. Adding a few servos to that would effectively replicate my X/Y table made from IKEA parts for only slightly more money and a bit less labor, although you’d still have to figure out a way to anchor the servos and connect them to the screw drives. I plan on raising and lowering my dremel drill press by hand to cut out the parts I need for my next generation CNC mill. So my computer control program will run the X/Y axis automatically, and then give on-screen prompts to the human Z-axis operator.
X-Table with Servo and linear bearings made from IKEA drawer rails
I am building a poor man’s Computer Controlled Router (Milling Machine). My first step will be to bootstrap the ability to mill needed components. Before I can make my own linear bearings, I am using cheap linear rails from the AS-IS (Surplus, Scratch & Dent) room of IKEA. Specifically, drawer rails make relatively good linear bearings. (Well, relatively good compared to what I could make myself without a mill.)
Continue reading
My iRobot Roomba 4150 started to exhibit a behavior that I hadn't seen before. After starting up, the robot would "wiggle" as it went forward for a few feet, then turn, then move forward a few more feet while wiggling. After a few turns, it would stop and issue a double "uh-oh" sound. After cleaning and removing the vacuum brushes, the behavior was still hapening, so I decided it must be a problem with the sensors or motors on the robot. The robot's wheels would still turn manually, and they would pop up and down into the body as normal. I tried cleaning the clif sensors and that didn't fix the problem. Finally, I blew compressed air into the drive wheel compartment and that fixed the problem. Aparently, the rotary encoders on one of the wheels had gotten jamed up with dirt.
The iRobot website did mention an OSMO update that is supposed to fix this issue, but according to the serial number on my roomba, it had already received the update. I guess removing the debris from the wheel is the better solution.
I refilled the Li-ion cells in my laptop battery about 18 months ago.The battery has worked as expected since then. The only "special" treatment this battery gets is that it travels in my laptop (as opposed to being the extra battery in my backpack) when traveling through TSA security checkpoints. (I figure it's better the have the battery with the fully applied IBM label sticker visible in the backpack, and the battery where I pulled off the label sticker to dissasemble it safely integrated with the laptop.)
Over the last 18 months, the battery capacity has reduced itself to 44% of the original value (This happens with LiIon cells as they age). My last full capacity ACPI reading is now 21060 mWh, and provides 1-2 hours of working time on my IBM X31 Thinkpad. In a few more months it will be time to replace the Li-Ion cells again. Luckly, a friend gave me an IBM T60 extended battery (containing 8 cells) that had failed. When I took it appart, it appeared that 6 of the cells worked fine, and two of the cells were dead (killing the entire battery). TaDa! 6 free replacement cells!
A system downtime announcement I received today (slightly anatomized to protect the guilty):
Summary:
On Tuesday, July 1, 2008, at approximately 6:30 PM, the XXXX XXXXXXX server instance, which monitors the availability of XXXX servers and network devices, went offline. This outage was not noticed until Monday, July 7, 2008. After discovering the outage, the server was brought back online and was again monitoring services at 6:00 PM on July 7.
If you have any questions, please contact the XXXX Help Desk (XXXXXXXXXXXX). Thank you.
I purchased a north American version of the Motofone F3 (Manufactured in Brazil), which operates on the 850/1900 Mhz GSM bands (used almost exclusively in North America) for $34 including S/H from dakmart.com. After plugging my SIM card in the phone it reported that it was on the AT-T network. My original Motofone F3 from India worked only on the 900/1800 bands, and wouldn't work on the AT&T network (or any other North American network).
My encrypted home directory worked well with 8.04, I only had to make one change caused by the new xml format of the pam_mount.conf.xml file as follows:
<volume user="summetj" fstype="crypt" path="/dev/sda6" mountpoint="/home/summetj" options="cipher=aes" fskeycipher="aes-256-ecb" fskeypath="/home/summetj.key" />
After I installed 8.04 from scratch I had to mount the directory and chown the files to the new user.
I have used Myth to rip a large number of my DVD's to AVI files (using either the Excellent or Good setting). Now that I am starting to play back the movies, I have found something that is very troubling.
On many of the movies, the first 95% of the movie (e.g. 1:13 min of a 1:26 movie) plays flawlessly. However, near the end of the movie, the audio and video starts to speed up and get jerky. (Audio/video sync is also thrown off.) Continue reading
I have been having problems connecting to a wireless network that does not broadcast its SSID. (A so-called "hidden" network.) I am using an Intel IPW 2200 802.11b/g mini-pci wireless card that uses the ipw2200 wireless module.
I found that reloading the ipw2200 module before attempting to connect to the network fixed my problems (but only if I put a pause between the unload/load/connect sequence). I use wifi-radar to manage my wireless connections, and was able to call my unload/reload script using its "before" connection command.
The script is simply:
sudo rmmod ipw2200
sleep 1
sudo modprobe ipw2200
sleep 1
This does add two seconds to the connection time, but this is not a terrible penalty to pay as without the above pause the connection would usually only work about 25% of the time.