Hey /diy/ers,
I'm looking to build a circuit to control and drive high-power single LEDs. I'm looking at LEDs with stated 1A maximum currents, but looking to build a circuit that can provide sustained operation at up to 1A, as well as pulsed operation at up to around 10x that (pulses of a few ms, so with an adequately low duty cycles I should be able to operate well over max specs for the LEDs without burning them out).
The problem is, I have no real knowledge of electronics beyond basic V=IR fundamentals. I'm trying my best here, but it's quite possible I'm approaching this problem in a very stupid way.
To control this, I'll be interfacing everything with a scientific instrument that has a number of active-low TTL output pins. I'm looking to use one pin to trigger the pulses, and (ideally) a few others to directly control the driven current. The CAT4101 LED control chip I'm looking to use pulls down current related to the current through the RSET pin, so is normally controlled by changing the resistance between RSET and GND.
Question 1: Given I have a bunch of output pins to interface with and no real way to output serial communications, is there a simple way of controlling a variable resistor (as shown) with parallel digital inputs? If not, what should I be doing instead?
Question 2: I'll be needing a 5V power supply that can do up to 10A. I'm not having much luck finding these; am I an idiot using the wrong search terms, or should I just find an old computer PSU and run things through the 5V rail?
/a poor confused scienceanon who needs help making technology work
1. Digital potentiometers
2. Yes, a 'real' 5V 10A supply is pretty pricey. An old PSU would be most economical.
>>950983
1a. There are things called digital potentiometers.
1b. If you just need few bits worth of control, you could use some mosfets to pull one (or more) of several RSET resistors to ground.
2. Yeah, a computer PSU would be an easy solution. Dunno what's wrong with your search terms, but you could check some big distributor's like Farnell's catalogue to get an idea what is available - assuming you aren't happy with PC psu. And then optionally buy shit from elsewhere.
>>950989
>>950990
Thanks guys. I've looked at a few digital potentiometers, but they all seem to either have increment/decrement control, or are controlled by serial protocols that I don't really want to code into my instruments by hand. If I want to be able to call $SomeResistance from parallel outputs (4-6 bits would be fine), what's my best option? I could probably use an arduino to translate to I2C, but are there better options?
>>950993
I think you're greatly overestimating how difficult it is to emulate a serial bus by twiddling bits. You can make things easier by using an SPI interface potentiometer, though.
Then there's that mosfet option. 4 mosfets, sources grounded, drains connected to the RSET pin via 4 different resistors.
>>950996
Do you have any information at hand about emulating serial buses? I'm willing to try, but I'm also trying to build equipment that others (with even less electronics knowledge than me) can use. One line of code to set the power level is fine; a dozen lines for each of the pulses in a serial commas train will probably ensure that no-one else ever attempts to use whatever I end up building.
The resistor bank with MOSFETs is definitely a good idea; I'll sketch something up as soon as I can. Thank you!
>>951037
Read the datasheet, produce the waveforms it expects. Or google Arduino projects.
But for Analog Devices' AD5160 (one of the simplest interfaces) it would go like this, assuming Value is holding the 8b pot value:
drive the CS pin low
for i = 1:8
if the highest bit of Value is true, drive the SDI pin high
else drive the SDI pin low
shift Value left by one bit
drive the CLK pin high
drive the CLK pin low
end
drive the CS pin back high
I'm assuming your DAQ card or whatever producing the digital outputs is slow enough that the digipot timings are totally irrelevant.
for converting digital outputs to resistance you can use a vactrol, either store bought or home made. if you have, for example, 4 outputs, you can make one with 16 levels of resistance by using weighted resistors to drive the LED or lamp.
>>950983
What is the meaning of the line labeled Intensity ctrl coming out of the variable resistor?
It's some electric input? Or it's there just to indicate you can vary the current ajdusting the value of the variable resistor?
If it is the later, and since you said you had some digital outputs, you could replace the variable resistor with as many fixed resistors in paralel as spare digital outputs you have. You can control which resistor will be connected whit some switching device (as BJT or MOSFET) drived by the TTL outputs and so, change the equivalent value of the array. More (paralel) resistors connected, the less the equivalent resistance.
Just check what happens if no resistor is connected at all (ie: the variable resistor gets "open" or "infinite"). If something really bad happens under this condition, you can leave a permanent (non switched, allways connected) resistor to ensure the equivalent resistance never gets "open".
That, or fiddle with some digital pot, as some anons sugested.
>>951250
The ctrl line represents a few pins of digital output from the controlling device (probably 3 or 4 bits); I included it as an abstraction of what I want to do.
Looking at the spectrometer specifications, I've attached some info on the TTL outputs. I'm having trouble interpreting it (what is U in the context of a circuit?), so some more assistance there would be amazing.
If I was to use MOSFETs, I'd be effectively shorting resistors with the transistor. Having not used transistors before, what sort of resistance would I see through an active MOSFET or BJT?
>>951426
a readable re upload of that file...
>>951426
>>951428
U means voltage.
Resistance seen through a satured transistor may (mostly on MOSFETS, never seen in BJT) be stated in the datasheet, but under some given condition (some load current and gate/base biasing). For both BJT and MOSFET, the voltage drop across the saturated transistor is given in the datasheet.
>>951426
>what sort of resistance would I see through an active MOSFET or BJT?
For mosfets... something like tens of ohms for the tiniest signal fets, around ohm for small fets and much less for power devices. 2N7000 would probably be fine for your application (6R or less at Vgs=4.5V).
If you use BJTs, you can expect relatively constant voltage drop at the applicable current range. BC547 would have around 40mV.
I'm not too sure what you mean by U, but if you omit R1/R2, you can expect the output voltage levels to be around 5V (high) and 0.5V (low), which would be ok for 2N7000.
