Monday, December 27, 2010

Blue Shield of California rate hikes in 2011 (clueless and/or greedy)

UPDATE Feb 10, 2011. I have started a new health insurance policy with Health Net to replace my overpriced Blue Shield plan. I researched health plans on and found one that offered nearly the same benefits as Blue Shield's but at half the price. Here's the basic comparison:

Blue Shield: $4000 deductible, no coinsurance, no dental/vision --- $115 (soon) / month
Health Net: $4500 deductible, no coinsurance, no dental/vision --- $65 / month.

Yeah. Blue Shield claims they are losing money in the individual health care market. I don't know what they are doing wrong, but it seems pretty bad. Force health insurance companies to compete by choosing the company with the best rate. It's the only way to keep them honest.

If you are concerned about this isssue, please sign the petition that I have started:

A few weeks ago, I received a 32-page booklet from my health insurer, Blue Shield of California, titled "updates to your health plan." Of course, the updates included higher premiums and less coverage. Today, I received another missive; this time just a two-page letter. Apparently, Blue Shield is so excited about raising rates, they have already issued another rate increase even before the original rate increase has gone into effect. The total premium increase will be a 55.5% increase within a three-month period and a 98% increase since I bought the individual policy in March 2009.

When I first bought health insurance with Blue Shield in March 2009, the policy cost $58. This is almost the least expensive health plan available from any company. It has a $4000 deductible, and I am in the second-cheapest age bracket: 19-29. I also bought a dental plan for an additional $18. At some point in late 2009 or early 2010, Blue Shield raised the health insurance premium to about $75.

The letter says that hospital costs went up 15%, and prescription drugs and physician costs went up 12% and 9%, respectively, in 2010. In addition, the letter states that only 2% of premiums contribute to Blue Shield's "net income". Therefore, they must raise my premium by over %55 in the next three months, not including the first rate increase they issued in early 2010 to cover their costs. Either their math is wrong, or I do not understand their definition of "net income."

In early December, Blue Shield sent me the booklet with the chart on the left. It says my premium should be $97, but by late December, Blue Shield had sent another letter and decided the plan should really cost $115. What changed between the first and last week of December? Payroll tax cut? I mean, hey, the government didn't give any sort of explicit tax cut for health insurance companies in this latest round of wild tax cuts, and so insurance companies just needed to step in and take what's rightfully theirs. You know, people will have more money in their pocket due to the payroll tax cut, so they can just send that extra money directly over to their favorite health insurance company. That seems fair.

If you are concerned about this issue, please sign the petition that I have started:

Saturday, December 25, 2010

Holiday woodworking projects

I made this vase by laminating walnut, maple, and bloodwood, into a 5" square by 14" long block, then turning the composite block on a wood lathe. I drilled a 1-5/8" hole with a forstner bit and extension, then inserted a closed acrylic tube so that the vase can hold water (and fresh flowers).

This picture frame is made from walnut, yellow heart, and bloodwood. The pattern is NOT inlay, it is solidly constructed by gluing small blocks of bloodwood and yellow heart between two long pieces of walnut. The entire glued assembly was planed, then cut into segments to make the frame. The needlework of the rose was done by my aunt.

Sunday, December 19, 2010

Fiberoptic joystick with quadrature encoders and arduino

I have designed and built a joystick that contains no electronic parts -- only fiberoptics. The motion is sensed via quadrature encoding, and the signal processing is handled by an arduino microcontroller.

Here is a similar project for linear position tracking:

This joystick is designed for use in MEG and MRI scanners where the electronics may interfere with the scanner's operation.

Fiber transmitter and receivers:
Avago HFBR-2412TZ (mouser)
Avago HFBR-1414TZ

Quadrature decoder:
Avago HCTL-2032-SC (mouser)

Fiberoptic cables:
62.5 um glass fiber custom cables from

Code wheel from

Saturday, December 11, 2010

DIY 10-finger flex sensor gloves for possible VR or video game control

I built a pair of flex sensor gloves for capturing the motion of all ten fingers. This system uses individual flex sensors made by Spectra Symbol and a National Instruments analog capture device to record the flex sensors' values.

Monday, December 6, 2010

Installing beer taps in my house

Since moving into my new house, I have been planning to install beer taps into a dining room wall. Tonight, I have completed my goal and drank the first beer from the new taps.

The faucet on the right is frosty since I just poured a victory beer. I haven't found/made any handles yet.
In this photo at the bottom left, you can see the white keg refrigerator under the counter in the kitchen, which is adjacent to the dining room. I chose this location for the fridge, since the beer lines could be kept short, and would lessen the need for beer line cooling.

Here's the view from the kitchen.

There is a foam-insulated bundle of hoses running from inside the fridge into the wall between the kitchen and dining room.

The fridge holds two of my beer fire extinguishers ("thirst extinguisher"), a CO2 tank, regulator, all kinds of different tubing and a cooling fan that blows cold air from the refrigerator up to the space behind the beer taps.

This fan is coupled to a 5/16" ID hose that travels up the foam insulated tube to the chamber behind the beer taps. The air returns to the fridge through the space between the beer lines and foam tube.

I later sealed the back of this styrofoam chamber with another panel of styrofoam so that any air blown into the chamber would build up a small amount of pressure and be expelled through the foam tube. I wrapped the foam tube in duct tape to provide an additional moisture barrier.


I used 1/4" ID hose between the kegs and the taps. This was not a good choice. For short runs, 3/16" is better since you can keep the keg pressure at 10-12 psi at around 40*F, and open the tap wide open without the beer exiting too quickly. The 3/16" hose provides substantially more backpressure than 1/4" hose. I will experiment with in-line restrictors since I am NOT taking all of this apart to change the hose size. There is a lot of vague/wrong/incomplete information on the internet about dispensing beer, so that doesn't help either.

Sunday, November 21, 2010

Adding a pressure gauge to a Delonghi espresso machine

I bought a new espresso machine with the main intention to hack it and modify various parts while enjoying really good coffee along the way. This machine is a DeLonghi EC155 and cost $90 at Fry's Electronics with $20 mail-in-rebate. I normally don't buy anything that includes a rebate, but I made an exception in this case since $90 is already a good price for this machine.

The EC155 has a stainless steel boiler, "15-bar" pump, large water reservoir, separate baskets for single and double shots, and a pressurized portafilter. Espresso purists will tell you that a pressurized portafilter is just a kludge that is meant to help newbies get the right extraction time despite improper grind size or tamping, but I might disagree -- more on that later. In any case, this machine is a really great value, and a HUGE upgrade from steam-powered machines.

I previously hacked a steam-powered machine and added an air pump and temperature regulator to have more control over the brew process: The machine worked pretty well, but it was time to retire it because of a few nagging problems including a leaky portafilter, overall junky basket design, and other very difficult-to-fix items.

The DeLonghi EC155 with added pressure gauge.

I could not believe the unit would ever reach its stated "15-bar" pressure, and I was also curious what the brew pressure really was for an average shot. I started by removing all of the machine's guts from the case. It's built surprisingly well. The grouphead is made of very thick cast metal (probably an aluminum alloy), and the boiler is all-stainless with brass fittings and silicone seals.The pump is a ULKA brand pump. There is no obvious model number, though.

The boiler contains a large heating element and two separate thermal switches -- one for espresso, the other for steam generation. The boiler empties into the grouphead via the brass tube sticking up through the boiler's bottom. This way, the water level must rise above the brass tube's top for water to enter the grouphead. When the machine is off, there is no chance of water leaking out.

The brass tube also contains a spring-loaded valve that keeps the grouphead closed until the pressure in the boiler reaches 75 psi (5 bar). I suspect this is to prevent the grouphead from leaking while using the steam wand.

I removed the original 4mm plastic tube that ran from the pressure regulator to the boiler and replaced it with 1/8" PFA tubing. I added a tee in the line and attached the pressure gauge with more pipe fittings and 1/8" compression adapter. The machine will reach 13 bar when the portafilter is completely blocked (no flow), but is usually 7-8 bar during a double shot. Thus, the pressure regulator doesn't do anything in normal operation since the pump cannot maintain more than 8 bar during an average brew flow rate. I may be hacking this part in the future to have more control over brew pressure. However, I think the espresso experts agree that the brew process should have constant flow, not constant pressure.

The machine is designed to use a pressurized portafilter, and produces LOTS of crema with the pressure valve in place. I removed the valve and tried a few shots, but they did not taste as good or have enough crema, and it was very difficult to make the shot last more than 15 seconds. I used the finest grind that I could muster and tamped very hard, but the shot was still underextracted. I suspect that the machine's designers didn't worry about the portafilter geometry, or depth of the coffee puck since they knew the valve would regulate the brew pressure and extraction time. It may also be true that the machine's pump flow rate is not high enough to produce enough pressure without the portafilter valve. I did some tests:

10.4 seconds per ounce at 7.5 bar (170 ml/min)

9.1 seconds per ounce at 5 bar (192 ml/min)

3.3 seconds per ounce at 0 bar (530 ml/min)

I am guessing that the pump is specifically sized so that the brew pressure will be about 7.5 bar and a double shot will take exactly 25 seconds.

Sunday, November 14, 2010

Laminated wooden bowl

I made this bowl by a method that I have not previously tried. The idea is to make a flat laminated board of contrasting woods, then cut rings out of the board and laminate the rings into a stacked shape. This link describes the method for the bowl in the place setting

I rotated each ring 45* with respect to the ring below it. This is good since the ring seams do not line up, and the bowl has a more wild, interesting pattern.

I am not good at turning feet on bowls, since I do not have a very good chucking system on my lathe, and don't get much practice with turning feet. Instead, once I cut the bowl off the lathe (or remove it from the drive plate), I just sand the bottom flat, and add some felt to cover the surface.

Testing Tenergy 9V NiMH batteries (discharge curve)

I recently bought some 250mAh 9V NiMH batteries and a charger made by Tenergy. I plan to use these batteries in a Shure PG58 wireless mic. I first tested the current draw by the mic from a bench power supply, and found that is was very close to 50mA from 6.5V to 10V input voltage, and did not fluctuate based on sound input to the microphone.

I connected a DMM in parallel with the battery that is plugged directly into the microphone.

Test 1 = battery charged until the charger LED changed to green, then the battery was removed and left on the shelf for 2 days before testing in the microphone

Test 2 = battery charged for two days (fast charge, then automatic trickle), removed from charger and immediately tested in microphone.

I am slightly disappointed since the battery is supposed to be 250mAh, and I am only discharging it at 50mA ( .2C), so I should get nearly full capacity, but instead only get 200mAh effective capacity when the battery is fresh out of the charger -- 150mAh capacity after a couple days . It will probably be good enough for my application, but this is more evidence that battery manufacturers are inflating their ratings.

I popped open the charger and found a circuit board with a single IC (part number is obfuscated -- possibly not intentionally), lots of SMD transistors and assorted passives.

Here's an oscilloscope screenshot that shows the charging current and voltage behavior. This battery is almost completely discharged. The charger maintains a constant current of 105mA, with a 167ms gap every 1.7 s where no current is put into the battery. Presumably, this is so the charger can monitor the unloaded battery voltage. As you can see, the battery voltage drops from 9.6V to 9.5V in the 167ms gap. As the battery charges, the charge voltage (10 V in this screenshot) rises to accommodate the battery's increasing state of charge.

Wednesday, October 13, 2010

DIY searchlight housing for 1000W xenon arc lamp

Original test run of 1000W arc lamp:

I finally finished the 1000W xenon searchlight project that I started earlier this year. The power supply is a slightly modified arc welder coupled with an automotive ignition coil for the starting pulse.

This is the searchlight's beam shooting skyward behind a large pine tree in my back yard. The beam is very difficult to capture on video.

Thursday, October 7, 2010

Low-cost DIY thermal imaging -- liquid crystal paint testing

I am trying to develop a low-cost DIY thermal imaging device. The commercially available thermal imaging cameras still cost well over $1000 because of high production costs and low demand. Many hobbyists would like to have a cheap thermal imaging camera even if performance is not as good as commercial or military units. My goal is to build such a camera.

In this video, I am testing one possible approach: Using a very thin projection screen that is painted with thermochromic liquid crystals. These liquid crystals change color in the temperature range 77*F to 86*F. Ideally, the projection screen housing would be heated (or cooled) to 77*F, so that all incident thermal radiation would raise the screen temperature higher than this, and immediately cause a color change.

Wednesday, September 29, 2010

AES tops DMV as most-aggravating government entity


If you ship a package internationally, and the contents are valued at $2500 or more, you are legally obligated to record your shipment with AES (automated export system). This is true for all shipments -- personal, small business, corporate, etc. OK, so no problem if recording the export were easy. Unfortunately, AES is the most backward, bureaucratic, aggravating, and time-wasting entity that I have ever dealt with as a business owner.

I first must register online with AES and create an account (understandable, but still a turn-off. Since the data is ostensibly for census, it could be reported for each transaction without knowing the sender's intimate details). Oops, my old AES account from years ago is no longer active and I cannot create another account with the same name. How can I reactive the old one? The only way is to fill out a paper form and fax it to AES. No online form, no phone call. It must be a fax, and then I have to wait for someone from AES to call me. OK, fine, that's done. Now I can enter my shipment details, right? Nope, my original AES account was registered with my social security number. AES doesn't used SS numbers anymore. I must apply for an EIN (employer identification number) from the IRS. OK, luckily the IRS has an online form, and it "only" takes 10 minutes. OK, now am I ready? Nope, the AES account must be "transitioned" from the old SS number to the new EIN number. I first must create another AES account to administrate the change from the original SS to the new EIN accounts. Meanwhile every time I create a new account with AES (new password for old account, admin account and new account), I must create a 12-character password with six non-repeating characters, numbers, and special characters that does not contain any dictionary words. The form conveniently offers to print out your password since their is no hope of remembering such a thing. Great security. Tell users to print their password -- on a networked printer?!

After registering my original account, I had to study for a quiz. That's right, a quiz. AES will not let users enter shipment details until they pass a quiz. Can you imagine FedEx requiring customers to pass a quiz so they can ship packages online? Good grief! OK, finally, I have my EIN account setup, quiz passed, etc, etc. It's a good thing I did this in preparation for a large transaction that I will have coming up in the next few months. My account will work in November, right? No, actually the 12-character password must be changed once every sixty days or else the account is suspended. How do I re-enable it? A fax. Yes, the only way to re-activate the account after 60 days of disuse is to send another fax and wait for a call back. Holy Hell!

OK, I've got my account reactivated, everything is in place, and I am ready to enter shipment details. Here we go. The form is divided into seven parts, and each part has about 15 fields. Some of the fields are easy, like my name. Some of the fields require me to determine the "schedule B or HTS number" for the items in my shipment. These numbering systems attempt to codify all commodities in the world. And since this is a US government system, you can be sure it is updated to reflect the latest technology and trends. Searching the HTS for meaningful product designations can easily take 30 minutes. It's also very unlikely that a first-time user would choose an HTS number that is even remotely correct. The AES included schedule B search engine is truly an insult to all search engines: "spaces are not allowed in search queries."

So basically, it's going to take between 2 hours and half a work day to simply record the package details with AES. I hope the package will not be going to Canada. There is an exception for packages shipped to Canada, but then a NAFTA form must be filled out.

Anyway, I've pretty much vented enough. Here are some various approaches to the situation:

1. Charge international customers more, or raise prices generally to account for the additional time wasting that AES causes.

2. Refuse international orders.

3. Split shipments into packages that are each valued at $2500 or less. This might be illegal, but I am not sure.

4. Ship international packages with the US Postal System and enter the correct (over $2500) value, but then choose "NO EEI 30.37a", which indicates the value is less than $2500. For some strange reason the system will accept it, and I have never heard of people having problems shipping stuff this way. The system might depend on USPS clerks checking the value manually, and they never do this because they are almost always overwhelmed by handling an international shipment.

5. Lie about the value, so that it is less than $2500.

Tuesday, September 28, 2010

Seriously hacked central air conditioning

In my old rental house, I had a large window air conditioner mounted in a window in my living room. It did a great job blowing cold air, but was ugly and noisy (as are all window air conditioners). When I moved to my new house, the need for air conditioning was a little lower since the new house has double-pane windows and the house is also shaded by trees. However, I play music with my band every week in one of the small bedrooms, and with six people in one room in the summer, the need for air conditioning is easy to grasp. I didn't want to mount the large window air conditioner in a window, since it is so big and ugly, so I mounted it in the house's crawl space and installed flexible ducting to connect it to the house.

I know it is very difficult to see anything in this picture; the crawlspace is very dark and cramped. The air conditioner is an LG 15,000 BTU/hr model. It's the largest LG model that runs on 120V. I modified it by adding standard register boots to its cold-side inlet and outlet. I also used a plastic storage bin to couple the inlet register boot to the evaporator coil. Everything is sealed with rubber foam gaskets. Currently, the hot-side inlet and outlet are not connected to anything and just vent into the crawlspace under the house. I know this is pretty dumb, so my plan is to connect the hot-side outlet (the back of the air conditioner) to some nearby crawlspace vents. I can use uninsulated flex duct for this.

I removed the unit's control panel and built some custom circuitry that allows me to control the device via the standard wall-mount thermostat in the house.

The inlet and outlet flex duct are 8" and travel up from the crawlspace in a utility closet that is open to the house's attic. In the attic, the return duct connects to a 14"x6" ceiling vent:

14"x6" return vent

The supply duct connects to an 8" inline duct fan that is mounted in the attic:
$130 shipped, new on eBay. This particular fan is very powerful and also very quiet. I was able to mount it to the ceiling joists, and it cannot be heard over the general air noise when the system is running. I originally tested the system without the duct fan, but the air conditioner's stock fan did not move enough air through the ducts to be useful.

10"x6" supply register

Another 10"x6" supply register

Yesterday, it was over 92*F according to this thermometer. It was 81*F inside my garage.

Ahhh, 70*F in the house!

I had just installed the new duct fan and was curious to see how well it worked, so I turned the thermostat way down and was pleased to see the system had plenty of cooling power. The house is only 1100 ft^2, so the 15,000 BTU/hr air conditioner is sized well. The supply air is just under 20*F cooler than the return air, which indicates a properly-sized system.

Today, I set the thermostat for 75*, and the system was able to hold that temperature while running between %25 and %50 (the thermostat uses "four cycles per hour"). It was about 92* today as well.

Retrofitting Ikea cabinet door dampers to old Ikea cabinets and non-Ikea cabinets

I recently went to Ikea to buy a new cabinet, and noticed that all of their display models now have these soft-close dampers that prevent the cabinet door from slamming. The damper catches the door as it closes, and allows it only to close the last couple inches very slowly. Ikea sells the door dampers ("Integral Door Damper" 2/$4.99) separately from the cabinets. I wondered if the dampers could be adapted to my existing Ikea cabinets. I bought some and found that old-style Ikea hinges can easily be modified to accomodate the dampers.

Left: Old-style Ikea hinge Center: New-style Ikea hinge Right: Door damper

Note that the hole between the adjustment screws in the new hinge is very rectangular, while the hole in the old-style hinge is rounded. The door damper needs to mount in a rectangular hole.
Less than one minute with a Dremel and 1/8" carbide die grinder bit made the hole in the old hinge much more rectangular.

The door damper snaps into the rectangular hole and works perfectly.

Next, I thought that I might be able to use the door dampers for non-Ikea cabinets and also Ikea cabinets that do not use the standard 125* hinge.

I started by making some mounting brackets out of shelving support rod. This stuff shows up in every hardware store. I used a milling machine to cut the pattern out, but it would certainly be possible to use a dremel. The large rectangular hole matches the size of the hole in Ikea's hinges.Next, I mounted the bracket to the interior of the cabinet with two screws. The top-side of this particular cabinet would not allow such mounting, so I mounted it on the bottom.

Snap the door damper into place, and voila! The door now closes silently. This method would work for nearly any kind of cabinet. The dampers cost only $2.50/each from Ikea, and the homemade mounting brackets are even less expensive.

Friday, September 17, 2010

Time delay for relay (turn-on delay)

When power is applied to this circuit, the relay will not activate until a few seconds have elapsed. The parts count is very low and the circuit does not need external or regulated voltage. I used the circuit by keeping the 12V connected at all times, and switching the ground connection. Still works fine. Circuit will turn off very quickly after power is removed, but a rapid off/on cycle will have a shorter delay than expected because of residual charge in the cap. Not a big deal in most cases, though.

Wednesday, September 15, 2010

Visa gift card (credit card)

I recently participated in a usability study and was given a Visa gift card to thank me for my time. Unfortunately, the gift card is not as easy to use as cash or a check, but I found a good way to handle it.

The gift card cannot be used at an ATM or bank to get cash. If you try to charge more than the card's balance at a merchant, the entire transaction will be denied. A merchant cannot check the card's balance. So, if someone receives a Visa gift card, uses it a few times, then wants to use all of the remaining balance, he or she has to check the balance by calling an 800 number or going to a website, then go to a merchant and say "Please charge eg $23.14 to my gift card and I'll pay the remainder with another credit card." If the amount charged to the gift card is even 1 cent over the balance, the transaction will be denied. Also, the card's balance will deduct by $2 every month ("maintenance" fee), so you'd better check your balance on the exact day you plan to use it. I guess this fee only applies 13 months after the card was purchased, so not quite as brutal, but the card's value simply disappears after 2 years.

As you can see, Visa must be making a lot of money on unused balance, maintenance fees, and the $5.95 that the gift card's purchaser spent just to burden you with this pseudo money. Here's how to beat the system: Pay your utility bill with the gift card. As soon as you get the gift card, use it to pay the card's full value into your phone, cable, etc account. The utility company will just credit your account, and lower the amount due for the next month or two. This way, you can use the full amount of the card without ever checking balances or worrying about fees. Now spend your utility bill money on something fun.

Monday, September 13, 2010

Comcast internet

Comcast cable internet is cheaper and much faster than AT&T DSL in the SF Bay Area. Here's the comparison:

AT&T DSL. $33/month plus basic phone line charges of about $12/month. 2.5 Mbit down / 0.42 Mbit up.

Comcast Internet 15/3 plan: $19.99/month (even for new customers who are not buying TV or phone). 17 Mbit (average 15 ) down, 5 Mbit up.

Here's what I learned:

Do not call Comcast's phone number. Instead, use their online chat feature. It seems the people in the call center are unable or less willing to make a deal than the chat room people.

Each CSR (rep) acts as his or her own negotiator. They all have different quotas and desire/ability to make deals. Start a chat session and just ask for current promos. If nothing sounds good, refuse politely and try again in a few minutes/hours/days. Different reps will offer different deals.

In one chat session, I mentioned that I currently had AT&T DSL. A few minutes later, the Comcast rep offered me cable internet for $33/month. What a coincidence! That's exactly what I was already paying for DSL. The reps have more information than I thought. The rep would not go lower than $33/month and I eventually refused.

Buy your own cable modem. eg
For $30 or less, you can beat the $5/month rental pretty fast.

The "internet-only" service apparently comes with some basic cable TV channels:

I do not want cable TV service (even for free), but I brought in my shop monitor from the garage to test if cable TV was indeed available. It appears so. The Comcast tech did not install any sort of filter between my house and the poll. I flipped through it, and it appeared to be all working. The only thing that caught my eye was an old Star Trek The Next Generation episode. I will return the TV to the garage and not even worry about "stealing" cable TV.

Here's a photo of a device that the Comcast tech called a "trap." It's not a trap in the electrical sense. It's more of a mechanical security device.

The tech left it on the ground after completing my installation, so I decided to cut it open to see its internals.

If this were a 1/4" left-handed drill bit, it might grab the inner piece of metal enough to turn it counter-clockwise.

Remote focus control for telescope (remote follow focus)

The chips in the control circuit are:

Avago HCTL-2032-SC quadrature decoder
E-lab EDE1200 stepper control
Generic 558 timer for simple oscillator

Hi-res version

Moon at f/6.3 1/320 at ISO100. The lens is approx 1260mm (telescope with focal reducer).

Composite shot of Jupiter with its moons.