« March 2010 | Main | May 2010 »

April 28, 2010

Diverging Diamond

I read a neat article recently at work about a graduate student who came up with a new way of laying out a highway interchange. The problem with interchanges is the traffic that has to turn left to get onto ramps or the crossroad. Left turns make for conflict points for possible accidents and make traffic signals less efficient because only 1/4 of the traffic gets to move at one time. Also, since left-turning traffic backs up into traffic going straight, you put them in dedicated storage lanes. So a 4-lane road might have 4 through lanes and 2 left-turn lanes on the bridge adding expensive bridge square footage (about $100 per square foot). So what this guy did was said why not put an intersection at each end of the cross-road that shifts traffic to opposite sides of the road so that people actually drive on the wrong side of the road over the bridge? If you are on the left side of the road already, it is easy to turn left onto the ramp. And people coming off of the ramp who would be making a left could also make an easy left (if they are turning right they merge in on the other side of the intersection so it still works like an easy right). Anyway, this is called "diverging diamond" interchange as opposed to a regular old diamond interchange. The nice thing about it is that existing interchanges can be retrofitted to a diverging diamond. Another advantage is that by eliminating the need for left turn lanes, the bridge only has to carry two-way traffic, so it saves on the width of the bridge or lets you increase capacity of the interchange without widening the bridge and buying up the gas stations and restaurants that always show up on the corners.

After touting the interchange for several years with research spent on analyzing the design through traffic simulations, they found out that France already had a couple of these. But the first one in the US was just opened last year in Missouri and Georgia plans on doing several as well. You can click on the picture below for a bigger version. Notice from the crosswalks that pedestrians walk down the middle of the bridge.


Another similar concept, only without driving on the wrong side of the road, is a tear drop interchange where you put traffic circles (they are actually squished into tear drop shapes) on each end of the interchange. Here is an aerial view of one in Colorado. With traffic circles there are no left turns, everyone makes right turns. Another approach is to make the whole interchange a giant traffic circle: at the north end of the circle a bridge carries westbound traffic and the bridge on the south end carries eastbound traffic. You would need two (narrower) bridges this way but still avoid building left-turn lanes. Here is an aerial view of one in New York.

April 22, 2010

Ultrafire WF-502B with XP-G R5 LED

Overall Rating: ****

Ultrafire WF-502B with 1-mode Cree XP-G R5


Battery: 18650/2xCR123A
Switch: Reverse clicky
Modes: 1
LED Type: Cree XP-G R5
Lens: Glass
Tailstands: No
Price Payed: $17.24
From: DealExtreme


  • Bright

  • Cheap

  • Good battery life

  • Looks good


  • Rough threads

  • Not much heatsinking

Features / Value: *****

I have been looking for a light using a Cree XP-G for a while. I was able to resist the first round of XP-G lights from KaiDomain which turned out to have some problems with the LED set too far below the reflector. And I was able to resist the first round of DealExtreme lights including the Skyray and Uniquefire L2's (the L2 looks just like my Ultrafire WF-504B). But then DealExtreme introduced some older P60 hosts with XP-G drop-ins, including this 502B for only $17.24 which was less than $4 more than the drop-in by itself at $13.59. The drop-in was available in a 5-mode (with strobe and SOS and usually memory) as well as a 1-mode. The 502B with the 5-mode drop-in is $18.93, so I decided it would be a better value to get the 1-mode and order a 3-mode driver separately. From reviews posted so far the 5-mode driver has some problems.

Ultrafire WF-502B tail and clip

I like that it comes with a clip instead of a lanyard. I don't know that the clip is that useful to me, but otherwise there is nothing to keep this light from rolling. Plus it looks good. However if you want a lanyard, there are no holes in the tail so it would probably have to be attached to the clip. The clip is mounted to the tube with screws so if you remove the clip the light won't be watertight.

502B head, reflector, and LED

The light does not tail stand, but it is nice that it doesn't have any sharp edges at the back either. The crenelated head works out fine. Not too aggressive to tear pockets, but it allows light to escape if it is head down.

Build Quality:***

The 502B is only my second P60 host. My 504B is silver and I was very impressed with it. I ordered the 502B in black (partly because I don't like the glossy black that the 504B is offered in). It was smaller than I thought it would be, but it looks quite nice. The anodizing doesn't seem to have any flaws. The clip on the side is very sturdy, being held in place with two phillips head screws. I'm not sure if tapping those screws makes the light less water resistant or not, but they do go all the way through, including the top screw whose hole goes into the threads (but not the screw itself).

The reverse clicky switch is firm but the rubber boot itself is smooth instead of textured. It feels like it has been worn down. I like the textured one of the 504B better.

Ultrafire WF-502B parts

The body comes in 4 main pieces: the tail, body tube, throat, and head. There are o-rings in each location. The threads and o-rings came lubed, but they are rough. As I was cleaning off the lube and rubbing a rag along the threads with my thumbnail, the rag would tear. After working the threads taking the light apart over and over again, the threads were okay, but this was disappointing especially compared to the 504B which had very good threads right out of the bubblewrap. The o-rings themselves aren't all that tight against whatever screws in over them, but I don't doubt the light is watertight.

The lens seems well secured by an internal bezel (the 504B's is external) with an o-ring in front of the lens. The lens itself had a little dirt on it which was easy enoughto remove once I unscrewed the internal bezel.

One of the biggest problems with this design is the throat is quite a bit larger than the P60 drop-in itself. The drop-in is about 22mm and the internal diameter of the throat is about 24mm, so there is a 1mm gap all around and the only way to transfer heat is where the reflector bears on the top of the throat. With the 504B the throat is only 22.5mm and I was able to cut out some strips from a soda can and get a snug fit. I wound up using 4 layers of strips in the 504B. But in this one, I needed more like 8 (I doubt the heat transfers as well with so many layers but I scrubbed most of the paint off the aluminum to help, and the light did get warm during the runtime test). Another possible solution would be to get some 1mm wire and wrap it around the drop-in, but I don't have any wire the right size (it would need to be 19 gauge wire which isn't a common size; on a trip to Home Depot I couldn't even find any 20 gauge wire that was solid copper). With the 8 strips,the fitwas tight enough that I had to remove the Super Bright R5 label from the drop-in to slide it in place.

Heatsinking aluminum strips

The drop-in itself seems pretty reasonable. There is a textured OP reflector and the opening in the reflector seems to be correctly sized for the XP-G LED instead of an XR-E which would have a bigger hole. A piece of thick paper is stuck over the emitter mounting board with a square hole for the XP-G shape. This is good as it allows the LED to get up into the reflector as far as possible. I didn't notice a speck of dirt on the dome until I was taking pictures and it brushed off no problem. Taking off the paper shield, the LED seems securely mounted to the pill, though you should be careful if you try this because the paper is very sticky and it seems to have been applied before the adhesive around the LED could set up.

Bottom of XP-G drop-in

I didn't take the pill apart, so I haven't looked at the driver yet (I will once I finish testing and put a 3-mode board in), but the back has a red ring and supposedly the red circuit boards are better than the green ones, though maybe that only applies to the 5-mode version.

Ultrafire WF-502B 1-mode XP-G R5 driver

There were a few thin shards of metal, probably from the threads, in the head of the light that I removed. They were tiny, but they shouldn't have been there at all.

Overall, I can't complain too much about the quality of a $4 host.

Battery Life:****

For the amount of light, the battery life is impressive. Maybe this is as it should be since the XR-E R2 has an efficacy of 104 lumens per watt while the XP-G R5 is 132 lumens per watt (based on Cree literature). So even though you are getting a lot more light, the LED is more efficient. The 1-mode driver is drawing more current, but it doesn't shorten life by that much, giving 90 minutes instead of the 105 minutes I got with the 5-mode R2 using the same battery. Based on the declining current draw, the light doesn't seem well-regulated and is probably direct driven (to some extent, because the light can take 2xCR123A, there must be some way of bucking the voltage).

Here's a group shot of the Fenix L1D, AKOray K-106, Ultrafire WF-504B, Ultrafire WF-502B, and Brinkmann 2xAA Minimag:

Fenix L1D, AKOray K-106, 504B, 502B, minimag

Time Volts Current (mA)
0:00 4.21 1480
0:15 4.00 1250
0:30 3.88 1150
0:45 3.81 1030
1:00 3.74 950
1:15 3.66 870
1:30 3.60 810

With the aluminum strips in the head, the light gets warm, but definitely not hot at all. After 20 minutes it is less warm. It should probably be getting hotter.

After writing this I did a more elaborate test of driver effficiency. I ran batteries down to 4 different voltages and then took a voltage and current measurement at the battery and then at the LED. Multiplying voltage and current gives power which I then used to calculate efficiency with some impressive results. It almost makes me think the light is in direct drive (there is a LOT of solder on the positive and negative pads and it looks like it spills to some other pads, which may not have been intentional; you can see this on the right edge of the driver in the picture above) except that I accidentally put the reflector in the light without an isolator disk over the LED and the light was in direct drive, drawing over 2 amps and only one mode. So if it were direct drive, the draw would be much higher.

At rest battery voltage3.563.814.044.12
Battery voltage with load3.463.723.893.97
Current draw at battery (mA)61084010801100
Power in (watts)
Voltage drop across LED3.
Current drawn by LED (mA)60087011101130
Power at LED (watts)1.822.693.453.51
Efficiency Pout/Pin86%86%82%80%

The efficiency is very high. Another multi-mode driver I tested at the same time with this LED was in the 60's. The current to the LED is actually higher going in to the LED than what is drawn at the battery. But look how the power at the LED drops off sharply between 3.81 and 4.04 volts.

Light Output:*****

Finally! This light is quite bright, significantly brighter than the XR-E R2 drop-in I have. The R5 should give off about 22% more light than a R2, but the driver seems to be driving the LED pretty hard as well with the draw of 1.4A to start with (the R2 driver draws 1.2A). Some of the measurements on CPF indicate 300 lumens for the first few seconds before dropping off to around 270 lumens. That seems about right.

The tint is quite nice. The hot spot is very white, but the spill is a little bluish. The R2 has more of a violet or gray looking tint whereas this is just bright white.

After about 50 minutes it is less bright, but still brighter than my R2. Even at the end, with the battery at 3.6 volts and drawing 0.8A, the light is as bright as the R2 on a fresh battery. That's impressive.

The beam quality itself seems very good with a tight hotspot and a decent amount of spill. I was expecting the hotspot to be bigger and more washed out than the R2, but actually the intensity is quite good. Given the lack of rings, I wonder if a smooth reflector wouldn't do this LED more justice.

XP-G on the left, XR-E on the right at 1/25th second:

502B vs 504B 1/25 sec

at 1/200th second

502B vs 504B 1/200 sec

at 1/1600th second. The P60 reflector does a nice job of focusing the hotspot, but you can see it is still bigger and it also seems more intense.

502B vs 504B 1/1600 sec

The last picture is at 1/200th second with the 502B on a low battery (3.63V) and the 504B on a fresh battery (4.20V). They light is about the same, but the 502B might have the edge since it has a bigger hotspot.

502B low batt vs. 504B high batt 1/200th second

Out on the backyard range, the 502B outperforms again. For comparison, here is the 504B with the XR-E R2. This light throws its little hotspot pretty far, so notice how the fence posts are illuminated. They are about 120' away. The stick in the ground is 25' away. This is a 4 second exposure:

Ultrafire WF-504B R2 outdoors

Now here is the 502B. You can see there is more flood, but the throw is also better. There's just a lot more light:

Ultrafire WF-502B XP-G R5 outdoors

Now for one more comparison, here is my Ultrafire MCU WF-1200L with P7 emitter, running off of 2x18650 batteries. This light has a big reflector for better throw and puts out at least 400 lumens:

Ultrafire MCU WF-1200L P7 outdoors

To compare the throw, here is a zoom-in of the fence posts for all three (504B R2, 502B R5, and 1200L P7):

Fence 504B R2 Fence 502B XP-G R5 Fence WF-1200L P7 LED


This is a pretty solid light and a very good value. If you want a 1-mode light, this one is great (though the Seraph clones or Uniquefire L2 hosts are better quality for only a few dollars more). But because the light is so bright, I do think lower modes are needed. Then the problem is finding a good driver for it. I will try the NANJG AK47 driver and see how that goes. I may also try to measure the current to the LED at some point.

502B in hand

April 12, 2010

Praise You (Like I Should)

I remember watching some kind of MTV or VH1 compendium of best videos and seeing one I really liked for Fat Boy Slim's song "Praise You". It's a pretty famous video, but probably a lot of people still have never seen it or even heard of Fat Boy Slim, who was in the 80's British band The Housemartins under his real name, before mixing electronic music as a DJ calling himself Fat Boy Slim (and using an Amiga computer, no less, at least up through doing the music for the movie Moulin Rouge in 2001). I put the song on Grant's MP3 player and he says it is one of his favorites.

Anyway, I was writing a review of the movie Where the Wild Things Are which I had very high hopes for since I always thought it was a cool children's book and it has Spike Jonze directing. He also directed one of my favorite movies, Being John Malkovich and starred in a lesser favorite, Three Kings. Well, Wild Things didn't turn out that great, but I looked up the movie on Wikipedia and wound up reading more about Spike Jonze. He got his start making commercials and music videos, which I knew. One of the videos he made, well into his career (after a number of videos and around the time of Malkovich), was for Praise You. And what I did not realize was that the main guy in Praise You is Spike Jonze himself as a nerdy community-oriented B Boy dancer. The video itself looks like something someone would shoot and put on YouTube, but at the time there wasn't quite so much material out there like this. But it was certainly shot the same way, and when the manager of the theater comes out to turn their boom box off, that is real, since they didn't get permission beforehand.

You can watch the video on YouTube and read more about the song on Wikipedia

April 11, 2010

Mom's New Laptop

Mom dropped off her new Dell Inspiron 1545 for me to set up this weekend. Her old laptop was a Frys store brand, but this is a Dell and the price was the same (except that she got an extended warranty on the first one and this one she is just doubling Dell's one-year warranty with her credit card). Moore's law says that the number of transistors on a computer processor doubles every two years. It has been five years so everything should be at least 4 times better and maybe 6 times. Let's see:

The old processor was a 1.3GHz Celeron and this is 2.2GHz Dual Core Pentium. Since there are two cores (processors) now I'm going to say this qualifies, but I have no idea how many transistors are on a Celeron or a Pentium. She had 256MB RAM and now she is getting 3 GB of RAM (1 GB was added to the old laptop later on). So that is 12 times. The old hard drive is 40 GB, the new one is 250 GB (6 times).

There are some notable changes between laptops of yore and today. Most of the ports, including the USB ports, were on the back of the old computer. Now they are on the sides. Her laptop didn't come with a built-in wireless card (though the computer that replaced it when her hinge cracked did) but it was 802.11g and this one is also 802.11g (I guess that's more of a notable lack of change). In the meantime they have come out with 802.11n, and most of the higher end laptops support that, but this one does not. There is a SD card reader in the front for transferring pictures from a digital camera. Also, while the old laptop had a microphone, this one has a microphone and a web cam for internet video calls. Also this screen is wider in relation to the height to mimic HDTV's. That changes the overall shape of the laptop a little, but this one seems about the same size and maybe a pound lighter than the old one (and noticeably lighter than my 2-year old Dell), which was pretty heavy. Lastly, this computer has Windows 7 and the old one had Windows XP, so she skipped right over Vista. I like Windows 7 so far. It seems pretty snappy and hasn't given me a ton of warning messages.

April 4, 2010

Many Mini Daffodils

I planted a bag of mixed daffodils several years ago in the little plot in front of my house. It gets a lot of sun and the daffodils come back every year. The regular ones come up first and are followed a week or two later by these little multi-daffodils where several smaller flowers are on a single stalk. So all of the regular flowers have died back and all of these little ones have just bloomed. Time for the macro lens.


April 3, 2010

I'm a Flashaholic

No, not just because I have too many flashlights now, or because I have done driver swaps, or gotten lithium-ion batteries. It is official now because at CandlePower Forums if you post over 100 messages your status goes from "Enlightened" (people with less than 20 posts are "Unenlightened") to "Flashaholic". The next step as far as I can tell is 500 posts when you just get an asterisk after Flashaholic.

This started last May when I was at Fry's and bought a LED bulb for my Versapak Snakelight, so not even a year ago. But I know so much more about flashlights now.