I've recently finished a "Mad Scientists Lamp" (http://www.instructables.com/id/Mad-Scientists-Light/) that uses 3 25W incandescents and an off-the-shelf dimmer. Yep, it's pretty cool, but I have that dimmer turned way, way down, so it's only putting out about the same amount of light as a candle (<2VAC).
What I want to do now is introduce some random flickering, and was going to do so with an arduino and some clever programming, such as adjusting the duty cycle of a digital out to control the brightness.
I can't seem to work out why 10VDC barely lights the light, yet 2VAC (so says my fluke) lights it much brighter. Even if RMS is taken into account, I can't figure it out why 10VDC is so dim.
Obviously, I can't drive 3 lights with the arduino, so I was going to use a transistor driven by a digital out to "blink" the 120VAC to control brightness. It's been, perhaps 10 years since I've messed with transistors, but it IS possible to choose a high-power transistor, and drive it with the output of the arduino, yeah?
Being all retro-looking in the first place, why not use a vacuum tube instead of a transistor, for the same effect. But, since arduino is digital, the vacuum tube would be acting like a relay, which has limited life. Should I perhaps look into preamp tubes, like they use in guitar amps? I've got no vacuum tube experience, i just think it'd be neat.
The bulbs in question are 25W tall bulbs they sell at home depot.
The relationship between voltage and light output is not linear. A bulb only glows because the filament is hot. At the lower voltages the filament is heating up but it is only emitting light in the IR region (and that is being absorbed by the glass), when it starts to get hotter you start to see it.
so I was going to use a transistor driven by a digital out to "blink" the 120VAC to control brightness.
Do not connect this transistor directly between the mains and your arduino, it is dangerous and it would not work unless you connected one side of the mains to your arduino. Instead use an opto isolator. Where are you thinking of putting this transistor.
tube would be acting like a relay, which has limited life.
No more limited life than if it were in the liner mode. A vacuum tube is very much like an FET in it's operation and can be used in a linear or switched mode. Problem is that the grid (like the gate of a FET) needs a higher voltage that the arduino can put out so you need to drive it with a transistor or FET. In any event I would also use an opto isolator for this.
You need to opto-isolate (i.e. 4N25 for low speed switching) or for fast switching, best use magnetoresistive-isolator (i.e. IL610)
Indeed for saturating a mosfet or a vacuum tube, you need to bang it with higher voltage & peak current totally incompatible with arduino TTL outputs drivers.
Indeed for saturating a mosfet or a vacuum tube, you need to bang it with higher voltage & peak current totally incompatible with arduino TTL outputs drivers
Not entirely true; many FETs have a "logic" version, where saturation is achieved with logic level voltages.
Not entirely true; many FETs have a "logic" version, where saturation is achieved with logic level voltages.
I would be interested to have this reference or P/N of such FETs.
Please note my application uses IRF540 & IRF840 to pulse extremely sharp on-off up to 1Mhz with very low PW which requires external FET driver IC.
P.S. The main point being stressed in this thread being best to use an opto-isolator or other tech isolator between TTL arduino and any FET tech or vacuum tube otherwise boom arduino or computer via USB link
Thx Grumpy and AWOL about bringing me awareness of "logic level mosfet".
Please note my projects involved Solid State Tesla Coil and Rife healing plasma tube devices with direct frequency and pulse width software generated by arduino. This is why i'm using specific Tesla Coil or Ruhmkorff driver including digital-isolator and IC driver to control mosfet to protect my arduino board and Macintosh.
It might not be relevant in this thread because the initial question does not involve same high voltage and fast frequency switching as found in Tesla engineering but do you think "logic level mosfet" has an advantage compared to "classic mosfet" i'm using so far such as IRF family and power IC drivers ?
Well, with tesla engineering, low Ron value is necessary but it is also crucial to have very low rise on - fall of fet switching, very low rise on - fall off fet's body diode along with a capacity to accept high voltage power supply (i.e. 500V) pulsing primary coil & very fast frequency with low duty cycle.
Once you hit 50KHz - 200Khz, PW between 5% and 50% as in my projects, i don't think arduino TTL output drivers have so much sharp edge plus cannot provide peak local current to charge fet's internal gate capacitor. From high voltage & back-emf protection point of view, I rather use digital isolator then high speed power mosfet driver to really make pure PWM and saturate my actual "classic mosfet".
As I said before, i don't know if "logic level mosfet" has internally the equivalent of high speed power mosfet driver (i.e. TC4423 or MCP1403) which would be neat to simplify PCB, remove additional power supplies and lower $ project costs.
