| Electronics > Projects, Designs, and Technical Stuff |
| All-Tube Automatic LED Lamp/Night-light |
| (1/2) > >> |
| Cyberdragon:
You already know I'm nuts. Gone and done something really weird again...:wtf: I made an LED lamp with a light sensor controlled entirely by valves. Well the circuit anyway, still need to solder it up, connect the transformer, and add a case of some sort. There are no transistors, the LEDs are driven directly by a valve, everything runs on 12V. It took me a very long time to figure out how to get the tubes to turn off in space-charge mode. (the circuit still acts up a bit, I'll have to see how it behaves once assembled) I originally intended it to be a night-light, but it's a bit bright unless I use a diffuser of some sort. Do you think that little transformer (on top of PSU) will be able to handle 300mA continuous (500mA startup surge) without getting hot? Parts List (minus power supply, any 12V power supply capable of at least 350ma will work): 12AT7 (or 12AU7 or 12AV7) 12BA6 7427 LDR tube (any LDR sensor will work but you might have to add some bias, you could also use other sensors or triggering devices) three 4.5v (20mA?) ultra-bright white LEDs (IDK what the max current output of this device is) 1K 1/2W resistor LED current limit resistor (optional, depends on your LEDs and how hard you drive the input) 500K Potentiometer (switch ganged or seperate) Trimmer Potentiometer (optional, to inrease max brightness/sesitivity) Tube sockets Tube shield for 7-pin socket (12BA6 must be sheilded for thermal stability) You could use different LEDs with a different current resistor. Just keep in mind that space-charged electron tubes don't play around, they are totally capable of blowing up your LEDs (mine luckily survived 50mA through the tube with a 10ohm resistor). I'll draw up a schematic some time or other (probably in MS paint :P). I also note that the tubes respond to touching the wiring, so my next project could be a tube touch lamp. |
| technotronix:
I think it want handle 300mA without getting hot. BTW what happened? did you tried? |
| SeanB:
Those little transformers only will handle 200mA pulsed, they run fine under 100mA. You can underrun the heaters in the valves though, and if you get a 9VAC or 6VAC one and run the heaters in series off that they will barely glow, but should have enough emission to work for your use. Bonus is the turn off will be easier, and they will be constant current switches. |
| T3sl4co1l:
Strange, I would've thought a 60FX5 (or two, for more current capacity and no filament supply resistor!) would be enough. Phototube and an output, wired for hysteresis (schmitt trigger). Uhh, and something to invert it, maybe some screen-grid trickery? Also, isn't "all tube LED" contradictory? ;D You could go with fluorescent or neon, instead. Fluos aren't so good at DC (reliability issue), so you might use some diodes (whether junction diodes count, is debatable; selenium and copper oxide rectifiers were standard for some time..) to hang a tube inside a FWB so it effectively controls AC. Or you can use a P-P transformer winding to control AC with a pair of DC-only devices. Or thyratrons, which has the advantage of controlling quite a bit of power from not much tube (a pair of 2D21s will do 100W easily). Tim |
| Ian.M:
In an alternate universe where Oleg Vladimirovich Losev didn't starve to death in 1942 during the blockade of Leningrad, LEDs would probably have been used in period valve equipment. He was publishing research papers on point contact Silicon Carbide LEDs throughout the second half of the 1920's, and had come to the attention of international wireless magazines. If he'd survived and fled to the USA, blue SiC LEDs (and white by phosphor conversion) could have been commercialised 40 years earlier than their actual development by Cree in the late '80's. A schematic would be nice, but I certainly wouldn't want to use MSpaint! - Circuit Maker 6 Student edition (which can be found by googling circuit maker student edition zip site:.edu - hint 'Bobsclass') includes schematic symbols for vacuum diodes, triodes, tetrodes and pentode, and can simulate triodes and tetrodes. Sorry, no direct link - I don't want to give Dave any legal hassles as the 'Circuit Maker' brand name is now owned by Altium Otherwise hand draw it on paper and scan it. You shouldn't have any real problems getting the 12BA6 to cut-off cleanly but you do need a negative grid bias. Assuming you are running its filament on 12.6V AC, as this isn't an audio application you can ground one side of the filament supply and run a voltage doubling rectifier from the other side tio give you a -35V bias supply. Another approach, as the control grid current is negligible, would be to level-shift the 12AT7 output through a grid bias battery - use three 12V A23 batteries in series for -36V of shift, which should be enough to get the 12BA6 into hard cut-off. |
| Navigation |
| Message Index |
| Next page |