LED board, dremel'ed

I guess, the next time we move the factory, I should make sure thet I have a compressed air line in my office.

I disagree.

Surplus stores still exist and modern online stores are incredible. Take a look at Radio Shack's web site for a look back in time. $1 for each 1/8w surface mount resistor. $8 for a s***ty power transistor. Local parts dealers were marking up parts even more because they figured they were to fix something important. Mail-order took over a month and cost another $30 in shipping. Surplus stores weren't much cheaper. I couldn't afford anything.

Now you buy a bunch of inexpensive variety packs from Digikey/Mouser/Future/Jameco/etc. and have your prototype done as fast as you can think your way through it. They sell chip-to-solder-pad PCBs for prototyping the old fashioned way.

I remember surplus stores as cheap. I used to buy aircraft radar displays from Fair Radio Sales for a few dollars, photomultipliers for $1. An aircraft under-wing radar pod was $70.

Before real estate got so expensive in Silicon Valley, there were gobs of surplus electronics places. There even used to be one on Market Street in San Francisco.

Kids just don't seem to be very interested in real electronics much any more. They mostly write code.

It'll be a lot more expensive by then, but that's inflation for you.

Best regards, Spehro Pefhany

I vividly remember having to spend two weeks' allowance to buy a ferrite-core inductor, and having nobody to ask for circuit advice.

The web has pretty well fixed both problems. (Having more money helps too.) ;)

Cheers

Phil Hobbs

Is compressed air banned in SF? It's not impossible to retrofit and not all that expensive.

One thing that strikes me is that the PCB is visible to the mixing cavity. I would have added a piece of satin aluminum with strategically placed holes so most of the LED light would pass through on its way out of the LEDs, but returning light would mostly miss the holes and be diffusely reflected back into the cavity. This makes the cavity more like an integrating sphere.

Joe Gwinn

Don't pressure me.

We have a huge old compressor, left behind when we bought the building from the fortune cookie people. The cookie machines had a lot of pneumatics.

There's an air line up to the engineering lab, where I use it to test NMR temperature controllers. The zillion-RPM dentist drill thing does sound interesting.

We have a small bead blaster cabinet too, that uses the air supply. We used that on Friday to bead blast the insides of my optical cal fixture, to diffuse the light from the four LEDs. Makes a beautiful flat finish on aluminum.

We're shopping for a new soldering machine. The choice is between vapor phase and a 6-zone conveyer reflow oven. The 6-zone needs nitrogen, which means either huge dewar deliveries or a nitrogen generator and storage tank, both a pretty big hassle. I like the vapor thing, but it's an ongoing debate. Lead-free soldering is somewhat less terrible if done in an oxygen-free atmosphere.

One thing I don't understand: you can mix oxygen and nitrogen, or dissolve salt in water, or mix uranium isotopes, with near zero expenditure of energy. But it's a huge amount of work to separate them back out. Doesn't seem fair to me.

Welcome to the second law of thermodynamics: A mess grows by itself, but you need to put huge amount of energy to sort it out. A good example is the kids' room.

When I was a kid, my mom opened a revolving-credit account with Allied Electronics so I could order parts. A transistor cost more than a steak dinner. Luckily, I also had an uncle who had his own TV repair shop, and had a shed in his back yard full of probebly-stolen Military surplus electronics.

The LEDs themselves are already diffused. I suppose I could paint the PCB flat white, but it's mostly gold anyhow, so a lot of any downward light gets recycled into the system now. I have to test some more diodes this weekend, so maybe I'll paint it white and do a before/after measurement. It might reduce light loss a little. The overall CTR (current from the PD over current into the LEDs) is about 0.0005, and I wouldn't mind a little more.

We could coat the inside of the tube with that super-flat-reflective barium paint stuff, but that would be messy.

I just discovered that these green LEDs make nice square light pulses at high currents, but have a significant optical overshoot at low currents. We'll have to fix that somehow.

With an integrating sphere (which need not be spherical for non-precision applications), it does not matter how well diffused the light source is. A laser beam coming in through a tiny hole in the wall will work.

..

For measuring the optical output of photo flash units, I've used a cardboard "china barrel" painted white with ordinary wall paint on the inside as an integrating sphere. The main requirement is to arrange things so that there is no direct path between source and sensor, so all paths include at least one diffuse bounce.

I think that a sand-blasted satin aluminum surface will work just as well.

Feed a negative turn-off pulse in at the end? We used to implement that with a capacitor and diode in the transistor drive circuit, to manufacture leading and/or trailing current pulses to add and subtract charge as quickly as possible.

Or it could be the population inversion decaying. If this is the cause, it may be difficult to change.

Both effects can be happening at the same time. The LED maker will have an opinion.

Joe Gwinn

The compressor is the cheapest part. Depending on the construction and aesthetics, plumbing it to the rest of the building can be a hassle.

We had them (slower) in IBM. They were great tools but air-powered solder-suckers are the cat's rear.

I'm considering one of those for home use. Harbor Fright sells them for cheap money.

At the PPoE, we an a 4-zone. It was a RPITA to set up correctly for RoHS. They switched to a 12-zone (they were giving them away in '09-'10) and everything went incredibly smoothly. I'm sure they didn't use nitrogen.

If it were the other way, you'd be bitching because they were so hard to mix. The universe is that way - TANSTAAFL.

You don't think kids supply energy to the system? That's the problem, it's all wasted on them.

Overshoot? Physics?

InGaN? 525nM?

Etching in NaOH also give aluminium a nice flat light grey finish

-Lasse

I don't have the Avago part number handy. It is 525 nm.

We have seen similar effects in laser diodes, microsecond turnon effects as if some internal layers were acting like PIN diodes.

We'll eventually use different numbers of these LEDs for the low and high intensities. For low intensity, we could put a filter in front of an LED and run it at higher currents, where it seems to behave better.

I really like those LEDs. Nice color and very efficient. I even got them to work well on 3V. Not much current but they don't need it. Expensive, though.

Interesting. The charge has to sweep out of the junction(s)? Why doesn't it show this at high currents? ...just to a lesser degree (constant after-glow?)?

More LEDs and more diffusion?