Looks like my tech is gaining some attention - the BBC's program "CountryFile" will be at the UK facilities for filming Monday.

This should be rather interesting. Shame I'm here in the USA - I guess I'll have to settle for a mere mention on the program instead of being physically present to explain things.

I've received another request for one of these, so I figured that I would put it up. Since they only last about two weeks, please feel free to track me down and ask for a new one if you ever have a question!

This weekend (I think, maybe earlier), Slashdot published some statistics about the most active people. Apparently I am in the top four most active commenters for the past month and the past quarter. This is quite depressing.

In happier, and unrelated news, my FreeBSD commit bit was approved this weekend, so I can now cause untold destruction on the Internet at large...

So my side porn shop job has these LED-lit display cases for Lelo vibrators. Of course, being an LED nut, I pay more attention to those cases than I really should.

That paid off. A couple of weeks ago, I had noticed issues with the display cases flickering on occasion. Yesterday, I could smell electronics burning, and my first instinct was to go check the Lelo Displays and their wall-wart power supplies.

At first touch, I got burned. So I kicked the warts out of the wall plug, waited a few minutes, and opened up the transformer housing.

This is what I saw.

The entire output side of the boards (bottom of the picture) are totally fucked. Resistors burned to a white ash, diodes burned black, capacitors bulged, transformer wrapping coming unglued from the heat.

Now, the design of the circuit is just fine, and the components appear to be quality. The major flaw I see is that this is 12V 1A being pushed down 20 feet of thin thin copper wire (WITH insulation, the wires are about 1mm thick.) DC hates being pushed any considerable distance (exception being HVDC) and this leads me to believe these power supplies were inadequately tested. We tried powering 6 meters of LED strips on DC back at the research facility in the UK, we measured 23.6V input at the beginning of the chain and 10V at the end, causing huge resistive losses and making heat sinks almost intolerably hot to touch. Lesson learned, don't drive DC power down any long length of wire, especially low voltage DC. I believe this is what caused the failure in these power supplies, and the fact that the failures are in the same area (on the DC output side of the board, essentially) would give extra strength to my theory.

That porn store should be happy they have me. I just saved them from a potentially expensive fire.

I'm really not sure how long these last as comment-on-able, so here's another one. I'll try to be a little more vigilant in making sure there's always a journal open for asking questions.

A little more elucidation: I'm in the fourth year of a combined bioinformatics/medical informatics degree. Most of my semester is from the CS curriculum, but I have gotten through biochemistry, genetics, molecular biology, physiology, and organic chemistry courses. (Mostly in the presence of pre-med students.) I've also taken genomics-specific courses and worked in a molecular biology lab studying C. elegans, and a medical lab studying Autism.

So ask away!

Where other major companies have failed, I have succeeded. I have pumped 300w of LED power into a package that is 30mm x 30mm before actual packaging material/power leads are concerned, and kept it cool with a copper-cored carbon-coated heat sink. I am about to push the current envelope 5x past what China thinks is possible according to my inquiries to major manufacturing companies.

Here's a photo of the original 100w prototype, modified from an RGB module.

Everything is about to experience a severe system shock, from HID to plasma to fluorescent. This much power, in such a small space, and cooled easily via air cooling with an advanced heat sink not using phase changing tech or peltier cooling, is a game changer.

Questions? Criticisms? Comments?

Current trials of a moving LED system to reduce the number of lights and power we require are going very well. Now, to design a single-light instead of a double light version, to try to cut power consumption down even further. I think I am going to work with a roving and rotating wedge design. This should allow for multiple angles of coverage and exposure across every plant. I will have to come up with a spirograph-inspired movement pattern, however, to make this workable.

Droop on the blue side of the LED equation really needs to be solved. I am tired of being unable to run these things at full power.

My current phone is a Nokia N80. I've had it a few years and I'm reasonably happy with it, but it has a fault with the charging circuit and it's pretty bulky, so I'm thinking about replacing it. Unfortunately, there seem to be about 3,000 different options with no competent way of way of working out which one is sensible.

I mainly use my phone as... a phone. So, the most important feature for me is the ability to make and receive calls. Because I am a cheapskate, this includes SIP (and WiFi), since my SIP provider charges a lot less than my mobile provider when calling landlines. I really like WebOS in terms of UI, but that seems to rule the Pre out because the only WebOS SIP client is alpha quality and doesn't integrate with the address book. This is something that Nokia does really well - the SIP client is fully integrated, so I can just select someone from my address book and select Internet Call to make the call. No extra skill required.

Beyond that, the only thing I really need is to be able to sync contacts via bluetooth and to use it as a modem via bluetooth - both pretty standard features, I'd assume, since my last three phones have had them.

In terms of smartphone features, I'm not that bothered. A programming environment that supports native code so that I can port my ObjC runtime would be nice - I have no interest in VM-based crap - but aside from that I don't have any strong requirements.

I would, however, like decent battery life and a small size, and ideally a nice camera. The bulk and poor battery life of my N80 means that I quite often leave it at home.

So, any suggestions?

Thanks to the power of my silly sig, I've been getting a lot of biology questions lately. Most of these are fun to answer, but occasionally they pop up in totally inappropriate threads just out of the blue. Since Slashdot supports commenting on journal posts, it seemed like the best thing to do would be to make one and encourage people to ask here instead. So do that!

And I really will phrase things as car and computer analogies when possible. Although computer analogies are way more common.

For clarification, I've taken physiology and genetics courses, and can answer most geeky things about the human body and fundamental biology. I don't know much about pharmacology or ecology (because it is very, very, dry), and I am not a doctor. But I'll try to answer those questions, too.

So now not only am I a research director, light designer, and consultant in the horticultural field, I am now a porn store clerk.

I needed the extra money. Stupid R&D eats up so much profit.

Well, had City of Riverside come by to inspect the house I'm renting. As suspected, it's worse than the landlord let on, and this house does not meet the minimum Implied Warranty of Habitability. Also, the house is in pre-foreclosure, which means he's not been using our money to pay the mortgage as he stated in e-mails, which puts me in the position of needing to find a new place, soon.

Time to call up Fair Housing, find a lawyer, and withhold rent, and maybe see if I can't squat on this place via court order.

Well, I've made incredible progress with LED. 19 days from dry seed and at a full crop, with all of the characteristics and content of sun-grown lettuce. The sun-grown lettuce takes 17 days, however that is in the summer, winter takes almost 90 days in our test solar area. This will be incredibly useful for winter production.

Now I've uncovered something a bit more dastardly in the same field and it's disturbing because I got victimized by it as well, though I've been lucky enough to not incur any damages of any sorts.

It turns out that almost every LED panel out there is nowhere NEAR pulling it's rated power. This presents several problems. The first one is pretty simple - when you advertise a lamp of any sorts to be of a certain wattage, you are promising that plus or minus a small amount, your light is consuming that much power per hour. If you have a 120w LED panel and it is only drawing 60wH, you are NOT getting what is advertised. If an advertised-as-50w panel is drawing 45wH (10% less than advertised) that is an acceptable deviation. If that advertised-as-50w panel is only drawing 25wH (50% less than advertised) that is NOT an acceptable deviation.

The second problem is just as bad, you're not delivering maximum photon flux to the plants, thus you will likely never get your advertised results.

I've found that many of my own panels are having this same problem. Certain lines (specifically my older ones) stay in specified tolerances, others are simply so far out as to be unbelievable. All of my 15w and 50w panels are well within tolerance. Everything else after that, it's a crapshoot.

In the sheds, my LED bars are pulling exactly what they should pull, thank goodness, although it would be neat to be able to say that I'm using even less power than anticipated and still competing with solar-grown crops.

This is actually a huge story I'm sitting on involving this fraud. China appears to regularly do this, and as a result many consumers are being taken for a ride while some of the more aware resellers continue to perpetuate lies to keep the product flow going. I've seen it already - "240w panel only draws 120w due to line purity."

It's very widespread. Am I the only one with a conscience? How do I handle this? I've contacted the Department of Commerce, I guess I should also contact the local or state DA. Any other people I should contact?

I'll be doing a quick run out to the UK in just a couple of days to examine and test the new prototype bar I specified for my parent company and their NFT systems. Remote testing so far isn't giving me much of the information I require (primarily because the numbers are near-meaningless to anyone but myself, and I'm not getting all of the numbers I need from the people I'm asking to collect the data.) I am hoping these OSRAM Platinum Dragon diodes are going to do the job. I don't like underdriving 3w diodes to 1w but this might be the easiest thing to do for thermal management. We shall see.

This should prove to be an interesting, informative, and entertaining trip, provided my current head cold is dismissed by the time I must travel out there.

Oh boy, the TSA is going to love me and my metal leg!

Customers are coming to me, understandably quite disappointed in prior purchases of equipment.

I've been noticing a very disturbing trend recently as more and more come to me - it seems advertisers are selling panels saying "1200watt panel, uses only 540 watts!" and these people are expecting 1200w of LED performance.

This isn't good for them, nor is it good for me. I'd love to help them out but I'm not entirely sure how. The best advice I can give so far is to tell them to return it if within the return period, or else try to resell the panel to recoup their losses.

How else am I to help them? I don't want to say 'lawyer up' though I have the feeling that's going to have to be their next step for a resolution to this dishonest advertising.

I want to conduct an experiment but I need to find proper CCFL tubes. They need to hit certain peak wavelengths. So far the only directly targeted one of four wavelengths I need that I have found is an Actinic CCFL. That handles 420nm, now I need something for 460, 630, and 660.

Hopefully they're all the same 3.6W tube at 300mm.

This might lead to something not as long-lived as the LED bulb but I won't need to worry about heat-induced failure, nor cooling. I'm not trying for massive plants, this will be more for multi-stack NFT systems.