… well, when I say “help me” I really mean “please do it for me”! I know my way around a soldering iron and how to read resistor values, but I can’t design an actual circuit to save my life.
The project is to mount 3 LED strips behind a small artwork. All 3 are to come on when triggered by a proximity sensor, and then turn off automatically in say a minute or so. Each installation to be independent of the one next to it.
The artworks are glass slides around 70x70mm and will be mounted in some sort of small frame. The circuitry must fit in the back of the frame, so be as small and light as possible, including batteries. And as there may be up to 100 of these things I have to make, as cheap as possible! As part of that it would be ideal if we didn’t even have to mount components on a normal circuit board, either perhaps just cardboard or even as a rat’s nest of free-floating components.
I wanted to use no programmable elements in the circuit if possible, to avoid complexity and cost.I am hoping that things like the timer will be possible with a few cheap components - say a capacitor charge decay - instead.
Even if asking for an actual design is beyond the pale, I’d truly appreciate any input or suggestions.
CS5555C timer can measure the minute, and can run from 1.2 volts.
Someone may have to change the battery in each light each day ???
and if they are rechargeables , how do they all get recharged ?
Maybe installing cabled power, or running off 6 volt lead acid , makes more sense…
if you use lead acid in each room, eg for 10 such slides each, hopefully each battery charge lasts the week, and there’s fewer to recharge…as they recharge at faster amps… its possible to do that way.
That proximity sensor draws 5mA – if you use 1200mA/h AAA batteries, that alone is going to cut you down to 10-days of idle time. A 555 timer sounds good if you want to avoid programmable parts, but if I were doing it I’d toss a cheap micro in there and do some hefty power management.
Also, if you’re going to do this in that kind of quantity, you might want to consider getting your own PCBs printed. It’ll make assembly much, much simpler. https://oshpark.com/ is pretty simple to use, but you might have to pay a dollar or two a board.
Thanks for the feedback, guys. So I have between 1 and 10 days of battery life So you think a programmable will cut down on power use? By, say, only sensing every 10 seconds instead of permanently, that kind of thing?
They can’t run off power; it’s not one installation of 100 slides, but 100 separate slides which will (hopefully!) be sold in unpredictable quantities to a number of people.
Of course I can’t get a board made until I can get a circuit designed! But I’m hoping it’ll be simple enough, few enough components, not to need an actual board.
Could small wires be embedded in wallboard compound (“mud” to those in the sheetrock/plaster business) and run between frames? At these tiny currents, bare 26 gauge applied to masking tape (adhesive side) and run along the level of the displays will save somebody a LOT of trouble.
120AC to 1.5 - 6vdc power supplies can’t be that hard or expensive to procure.
If I were designing this, a couple questions come to my mind:
How will I design the 1 minute delay?
How will I power the LEDs?
I can think of three approaches for #1:
555 Timer in monostable operation. This would be the simplest solution. But 1 minute is considered a long time for a 555. This is because the interval is based on an RC time constant, and long time intervals require high C and/or high R values. High values of C and/or R can be problematic due to component tolerances and leakage currents. If you use this approach, you will need to use a CMOS 555 chip, very stable components (e.g. do not use an electrolytic capacitor), and you’ll want to make sure the board and components are very clean to reduce leakage currents. This is probably an O.K. solution for what you’re doing, but I would never use it for a commercial product.
Clock and counter. This is probably the best compromise between reliability and cost. The clock can be derived from a 555 in astable mode, or (better yet) a crystal.
Microcontroller. This would the most reliable solution, but also the costliest.
As for #2, I would connect the strips in series, and then power the series using one of the following methods:
Constant voltage source with series resistor. This is the most common approach. It’s extremely simple and cheap. But it doesn’t offer any protection against overvoltage/overcurrent, and the load current will vary if the load impedance changes. I am not a fan of this approach.
Constant current source. This is my favorite approach. Configure an LM317 as a 15 mA constant current source, and use a 15 VDC to 24 VDC power supply. (Instead of connecting the light strips in series, you could instead use three separate LM317s – one for each light strip. With this approach, you’ll be able to power the strips using a lower-voltage power supply.)
LED power controller ASIC. The best approach, but also the most complex and most costly.
The first sensor linked in the OP has a really short range and the second, while a little longer, has an analog output.
The website calls the second one a distance sensors and they have an analog output with respect to distance which means that they have a variable voltage signal which changes as the detected item draws near. This probably isn’t ideal for a simple circuit since whatever is ‘reading’ this analog signal will need to be tuned for a certain threshold. Also, since there is a light source, it will draw more power.
For what I’m imagining the art installation (something like a picture frame which lights for a minute when someone approaches it, no?), the 12 inch range is way too short. What does slide mean in this context?
I also like LM317 in constant current mode for LED power, though they can be a little thirsty and drop 3 volts when on.
Depending on the power draw of the LEDs and expected duty cycle, I might consider something else.
Without-a-doubt, the minimum 3 V drop is the biggest disadvantage of using the LM317 in CC mode. So it’s probably not the best approach when P[sub]out[/sub]/P[sub]in[/sub] efficiency is paramount. If efficiency is high on the list, then he’ll have to do one of the following:
Use a relatively low power supply voltage and relatively low-valued series dropping resistor. But there are dangers in doing this.
Use a LED power controller ASIC. This is probably the best approach in terms of efficiency and reliability, but also the most costly.
So you think a programmable will cut down on power use? By, say, only sensing every 10 seconds instead of permanently, that kind of thing?