Hackaday Links: Sunday, April 28th, 2013

hackaday-links-chain

Another week has gone by and we hope you’ve been happily hacking away in your underground lairs. If not, here’s some inspiration that didn’t quite make it to the front page this week:

[Razr] used a CFL ballast to replace the mechanical one in his fluorescent tube light fixture.

To make the drawers of his workbench more awesome [Rhys] used the faceplates from some servers.

This week saw some changes in the hobby PCB market. Looks like BatchPCB is being sold to OSH Park starting May 1st. [Thanks Brad]

[Rich Olson] shouldn’t have any trouble getting out of bed now that his alarm clock literally shreds cash if he doesn’t shut it off.

We faced the same problem as [Kremmel] when we first got a Raspberry Pi, no USB keyboard. We bought one but he simply hacked his laptop to work. [Thanks Roth]

You may remember that post about a self-propelled snowboard. Here’s a similar project that uses a screw-drive system.

And finally, if you need help reading a quadrature encoder from a microcontroller this lengthy technical post is the place to look.

40 thoughts on “Hackaday Links: Sunday, April 28th, 2013

    1. Firstly it seem you’ve got it backwards, he’s actually replaced the ‘Magnetic’ ballast in an old florescent fixture with the electronic ballast from a CFL.

      Second while the main component in the old style of magnetic ballast was the very large iron core inductor, the very old magnetic ballast’s did also have a moving part called the starter, which was responsible for preheating the filaments and working in conjunction with the ballast to give a little inductive ‘kick’ to get the arc going.

      ( Some links I found to help shine some illumination on my explanation. )
      https://www.youtube.com/watch?v=hhUGPu3DgHw
      http://nemesis.lonestar.org/reference/electricity/fluorescent/preheat.html

      That is why old florescent fixtures used to flicker while starting, because the starter has to disconnect during the right part of the AC waveform when the current in the inductor was high enough to generate the required voltage to ionize the gas inside the lamp. If The starter disconnected in the wrong part of the sine wave it would just cool off and try again. Hence the flickering effect.

      1. Whoops… I missed the point I was getting at; which was even though calling a magnetic ballast a ‘mechanical’ ballast is a bit dubious, there is at least a little justification in doing so.

  1. Just wanted to point out that my server tool drawer actually uses the entire server and rail assembly. All the server parts are removed and then lined with plywood. It makes for a very practical drawer that can take a lot of wight. And just for fun the drives have been cut down and glued back in so that it keeps the server look.
    (I’ve already tried to post comment but for some reason it’s not showing up)

    1. Other than my wife’s wireless keyboard mouse combo, I can’t honestly say I have a single other USB keyboard. It’s been an issue from time to time, but even now, there are still a few older computers around which don’t seem to recognize USB keyboards in BIOS.

    2. I don’t think it’s a case of only having a PS/2 keyboard available. I had to go out and buy a USB keyboard to use with my RasPi and Pandaboard because until recently, I didn’t have a desktop PC at home (I’m not a gamer, so I made do with a couple of laptops) and no desktop machine means no separate keyboard at all…

    1. True, but you can do better than this link by treating every edge as an event (leading to 4x better resolution). That’s what the linked Arduino code does, and it’s what I’ve done previously on FPGA fabric too.

      1. This post does describe 4x resolution as it does treat every edge as an event. At least in the double interrupt case.

        I’m not saying this is THE way to read quadrature. Just explaining what I think is a very elegant, if not cryptic, way of doing it.

        I’ve never tried FPGA before. Should be worth a look at even with the extra hardware.

    1. Actually it is the entire server chassis and rail assembly. I’ve tried to leave comments to clarify that point but they just get stuck waiting for moderation. In fact the putting the face back on is just for fun, The main reason for using the server is that it makes a really practical drawer. Cheers, Rhys (Let’s see if this comment goes though!)

      1. I’ve been trying to comment here but for some reason they don’t get approved. Trying again from my phone…anyway I am using the full server and rail assembly. All the guts striped out and then lined with ply wood. I did think of using the drive bays as little drawers but the way it all comes apart meant it wasn’t practical. At leasr not on the HP server. So I just cut the front off the drives and glued them back in. They are just there for fun and only take up about 2 inches.

  2. the CFL “hack” is looking for a fire, and really is generally highly irresponsible. Packaged electronic ballasts meant for the task are really cheap, and have better dielectric withstand than DUCK TAPE! (not duct, thank goodness, which is aluminum!)

    1. Plus CFL lamps run at a different voltage across the terminals of the tube, They are smaller in diameter so the voltage is higher so a watt to watt replacement will not always be appropriate. Too much voltage into the lamp can cause the ends to overheat, short lamp life, damaged sockets, etc. There are some pretty universal ballasts like the Workhorse brand ones, try one of those before yanking the one out of a cfl.

  3. Discovered using CFL ballasts to power T8s back in 2005, and did a write-up on an aquarium site. They had MUCH better build quality back then, and functioned perfectly with T8s of roughly equal wattage. I still have some about ten years old I’ve saved in case I ever need them. But I don’t trust the modern ones. They fail if stressed, if the bulb fails or is disconnected, or for no reason at all – and they usually fail with smoke. So it’s a trick I no longer recommend. If you do try it, use the oldest ballast you can find.

    1. Yeah, I definitely understand that those features aren’t going to be important to everyone. My hope is that some of them will be important enough to some people to get or keep their business.

      Having the boards made in the US is important to me, and even if I got rid of all those extra features (the high temperature substrate, the gold finish, dropped the lead time back to a month, and loosened the specs to batchpcb levels), I still wouldn’t be able to match the prices BatchPCB got from China. They just don’t have the environmental or labor costs that US fabs do.

      Since I have no setup or shipping charges, any board under 5 square inches will come out cheaper with OSHPark than BatchPCB did (and you’ll get three copies of it). Looking through the BatchPCB database, 75% of people’s boards should come out ahead. Even a 10 square inch board (which covers 90% of boards ordered from BatchPCB) would cost only 30% more with my service, due to the setup cost.

  4. There are some distros that you need a keyboard for to do the initial setup. But after that, you could just ssh, vnc, or rdp to the pi. Unless he’s dead-set on using the Raspberry Pi as a stand alone desktop, I can’t think of a reason to do this.

  5. There are some distros that require a keyboard to be connected for the initial setup, but after that you can just ssh, vnc, or rdp to the Pi. Unless he’s dead-set on using his Pi as a desktop computer, I can’t think of any reason to construct a Serial to USB converter.

    Unless, of course he /is/ using it for the initial setup, or he wants to show off his mad skills, by which I am entirely impressed.

Leave a Reply to maconaCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.