Request for Help on a Simple Circuit Project

[Joel Balestra]

Hello all,

I took an Arduino class with you guys several years back which was super helpful and I was hoping to get some guidance on a small project I'm working on. The concept and circuit is really simple. I'm creating a drum triggered LED so that when a drum a struck, the LED(s) turns on instantly and then fades out. I've got a working proof of concept, but I need help scaling up. Currently I'm using a piezo sensor to detect the drum strike which is connected to a transistor that controls the LED. The LED is powered by a 9V battery with a capacitor wired in to handle the fade out. I was hoping I could get help on what components to use (or if what I'm doing is even the best way to do this) to power several bright LEDs and to make the LED brightness determined by the intensity of the strike. Your website says the best time to drop in are Tuesdays and Sundays at 7:00pm. Would anyone be willing give me a hand with my project this upcoming Sunday? Frosty beverages can be provided if anyone needs incentive and as a way to show my appreciation. :)

A little about myself: I'm a software developer by trade and I play around with soldering and electronics enough to be dangerous (what's burning?) I either understand the basic components or can easily grasp those I don't. I'm super eager to learn more and make more stuff. Thanks for your time.

-Joel Balestra

[nfolken]

For a purely analog circuit, you could just rectify and amplify the piezo signal, which would directly translate the intensity of drum vibration to a light (Not an LED though, as LEDs are digital and only on or off, and need PWM to vary intensity). The fade out would be diretly related to how long the drum vibrates after it hits, which probably isn't long.

I would recommend a digital circuit instead, using a ATTiny85. Condition the piezo signal as an analog input, and then have the ATTiny85 output a PWM to slowly fade out the LED. The ATTiny85 has 2 PWM ports, so you could even have both Red and Blue LEDs, and use a mix of them depending on how hard you hit the drum.

There will probably be a bunch of people to help out, as this Sunday there will be an arduino class. Come on by

[Keith Rogers]

Joel - Interesting project!  I don't make it down to the space very often but have a key piece of info that might help guide you.  LEDs are diodes and although their brightness may be somewhat dependent upon how hard you're driving them, they have a pretty small range of brightness that is controllable by current.  The trick to making LEDs fade is to pulse width modulate their on/off state (blink them quickly, resulting in the appearance of brightness control).  You will probably want to do this with an arduino.  Just look for the code for PWM (think this is analogWrite).  As for making the brightness a function of strike intensity, I suspect you can filter and read the piezo response into the microcontroller as a voltage.  You can use the built-in analog to digital converter to take in a voltage (between 0 and 5V for 5V arduino) which will provide you with a number between 0 and 255 within your code.  Good luck!  -Keith

--
You received this message because you are subscribed to the Google Groups "Arch Reactor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arch-reactor...@googlegroups.com.
To post to this group, send email to arch-r...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/arch-reactor/14ccbe25-0645-4007-8db1-95c6b445ff5f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[EschewObfuscation]

What you have done already is not bad at all, and can be scaled up pretty easily. You will likely find that people here feel that every project requires a microprocessor. I attribute that to "man with hammer" effect, and admit they're cheap enough the approach is no longer as insane as it would have been a decade ago. But there are many ways to skin a cat.

What you have done can work on a larger scale. Consider: take the trigger signal you are already generating, trigger a 555 in "monostable" mode. You will find that circuit in a thousand places on the net, also called a "one shot". (Just google "555 monostable".) That lets you generate a pulse of any duration you want. That longer pulse can turn on a transistor to charge up a big cap which feeds  a resistor and one or more LEDs. Or just multiple LEDs, if you run them in series, so no current limit resistor is needed. That will vary with your design. The reason for the one shot is it might take a while (several milliseconds, maybe.) for your transistor to fully charge your capacitor, depending on how big it is. You can get 5.5 volt supercaps in the 1 farad range (gi-normous) cheaply, if you like. That may prove excessive, but the point is you can get as much cap as you need cheaply nowdays, if you wish.

[Joel Balestra]

Thanks for the quick replies. Since it will be used during live performances, the device needs to be as simple and fail proof as possible. I'm also going to need at least nine (more as backup) of these devices so keeping the cost low is important. I'm avoiding Arduinos or other programable devices because of cost and size issues. Also because I like the challenge of making the design as simple as possible. My choice of LED was more a matter of availability and power efficiency than anything. They just need to by white and light up super bright under the drums. I'm not married to the design or components so I'm open to what ever will work best. They'll be used for about eight minutes at a time. The feedback has been great. I'll bring my proof of concept up on Sunday. Thanks!

Here are a couple links I've been using to base my design on if that helps.

Fading LED

https://www.youtube.com/watch?v=Hil2EiXzXCM

Audio input reactive LED's

http://www.instructables.com/id/Music-LED-Light-Box/step6/Building-the-circuit/

[EschewObfuscation]

Well once you design and debug the hardware, and do the low level programming, recurring cost of an attiny design need not be high. The big cost is the programming effort to get to initial function. Once that's invested, then you can more easily accommodate "feature creep" in a software-based design than a simple analog design.

Analog will give you quick initial build, low hardware cost. Arduino will give you higher hardware cost, but you would do that only for the test prototype, then you would port it to a tiny, hope it fits, and do a production run of that hardware. So you could make a case for either approach. I'd say "quick and dirty, and pretty cheap = analog" in this case.

[nfolken]

A simple monostable 555 circuit would work well for an incandescent light, but not an LED. To dim an LED with 555s you would have to do a Diminishing frequency astable circuit, which is a bit more complicated. To get enough light to light up a drum you would need a lot of LEDs, or some very expensive ones. You might be better off using a 12v car light bulb.

What is your power source? If you are trying to run off of batteries, then I guess LEDs would be better, but then I would still recommend using an ATTiny85. Adafruit Trinkets only cost $8, and are easy to use. You could prototype with one of mine. Then if you are building a bunch of them on the cheap, the ATTiny85 itself only costs $1.15

I have a bunch to say on the matter because I have been thinking a lot lately about make 'smart' color fading LED throwies. Here is a project that just an ATTiny85, LED and a coincell battery, and lasts for days. You would just have to add your piezo sensor instead of light sensing from the LED. But you would need a bunch of them, or else a bigger battery so you can drive more LEDs at once.

[EschewObfuscation]

The 555 in the proposed circuit is just a pulse-stretcher, to lengthen the charge time of the cap. Intensity decay is simple RC time constant as the cap discharges through the LED(s)

People seem to think that LEDs are purely binary, on/off devices, but if you have observed your flashlight with weak batteries, you will have seen otherwise. At any rate, a pretty quick experiment will determine if that RC decay mechanism provides sufficient dimming effect. It's possible this may end up a microprocessor based project; certainly great versatility results, once one has the software platform basically working. But the 50 cent chip plus one transistor solution may prove sufficient too. Experiment is the way to tell.

[Joel Balestra]

I'm starting to get sold on the ATTiny85 idea. Especially if I want to do any advanced lighting tricks. I've got an Arduino I can program on to test that out. I also found a 555 circuit in my collection of odds and ends I can play with. I'll bring it all in and see what happens. Thanks again.    -Joel

On Friday, August 1, 2014 11:47:51 AM UTC-5, Joel Balestra wrote:

[Chris Weiss]

LEDs are diodes, so once the voltage drops below the LED's designed
drop then it's off, and above it is on. if there's less amps
available than its rating, it could be dimmer, depending on battery
chemistry and protection circuits. so the dim flashlight effect is
first due to reduced capacity, then due to the battery cycling below
the LEDs voltage then since the load is "off" it pops back up over it
and the load is "on" again causing the voltage to drop again, so it
really is just blinking very fast. also the capacity dimming affect
varies greatly between LEDs and even somewhat in the same batch, so
it's very difficult to control predictably in that manor.

[EschewObfuscation]

Doesn't matter; he said he's going mpu-based pwm for this project.

[EschewObfuscation]

The timing of this sale may be fortuitous for this project:

http://www.goldmine-elec-products.com/prodinfo.asp?number=G18580B

[Joel Balestra]

I'm planning on using ATTiny85 chips since I'm dealing with so many variables. I'll need to program the chips to work for four types of drums. There will be 13 drums in all.

On the topic of LEDs, I picked up a couple high power LED's yesterday from SuperBrightLEDs.com

Cool white

http://www.superbrightleds.com/moreinfo/high-powered/vollong-3w-white-high-power-leds/896/

RGB LED just for kicks

http://www.superbrightleds.com/moreinfo/high-powered/vollong-3w-rgb-high-power-led/899/2214/

I haven't played with the RGB yet, but the white one works great. I'm heading into the drum studio tonight to work on programming my Arduino and calibrate the piezo. I still need to apply a logarithmic fade on the LED too. The LEDs on sale is a hell of a deal. Using two gets close (lumen wise) to the single one I'm testing and if I add a third, it's still cheaper. I was hoping to have multiple lights per drum anyways. A set pointing down and a set pointing up reflecting off the bottom of the drum.

So a few questions:

-What kind of of wires should I use for the LED circuit? I want to make sure I don't go overboard or create a fire hazard.

-Is LED heat a concern? I lean towards it not being an issue since the LED's won't be on constantly.

-Stick with 2n2222 transistor for the high power LEDs? TIP31 and 2n2222 (which I believe was suggested) seem to be the popular ones for 9V circuits like mine. But most of those are using less powerful LEDs.

-LED and piezo connector suggestions. I'd like to be able to easily detach the piezo and LEDs from the "brain" of the device. Any suggestions on a good/cheap connector for this?

-Mounting recommendations. All of this stuff needs to be mounted to a Velcro strap that wraps around and attaches to one of four different drum stands. My thought was to have a single "board" permanently attached to the strap that houses the "brain": battery, resistor(s), transistor, ATTiny85 and two cheap connectors for the piezo and LEDs. The piezo gets taped to the drum and the LEDs would attach to the strap via their own Velcro backing to shine where they need to shine. If the LED Velcro idea doesn't work well, I'll just sew them into the Velcro strap. How good/horrible of an idea is this?

I'm demoing a single drum with the trigger tomorrow night. After that I need to work on financing the project. Thanks again for all the help.     -Joel

On Friday, August 1, 2014 11:47:51 AM UTC-5, Joel Balestra wrote:

[EschewObfuscation]

On Wednesday, August 6, 2014 12:04:59 PM UTC-5, Joel Balestra wrote:

...

    So a few questions:

    -What kind of of wires should I use for the LED circuit? I want to make sure I don't go overboard or create a fire hazard.

Operating in the 3 Watt range, you won't need to worry about fire hazard. The obvious choices would be 22 or 24 gauge hookup wire, and that would be fine. Stranded would likely be a good choice, as you will likely handle it more than in some other type of project. Just be careful about strands getting loose at the solder points.

    -Is LED heat a concern? I lean towards it not being an issue since the LED's won't be on constantly.

The 3 Watt will likely need some heat sinking, but not a great deal of it. A piece of aluminum 2 inches square would be my first trial, use the 5-ended temperature probe (your fingers) to see if that's adequate. If you can hold it comfortably while it's running steady state, you're good. An old 1/2 inch tall pentium 1 fan/cooler would be vastly more than required, as they were made to handle 60 Watts and you're dissipating 3. You could also stick it to the bottom side of a TO-3 fin assembly, if you want something factory made. Once you have chosen your heat sink approach, you can use two part thermal epoxy (there's a ceramique variant from arctic silver, gateway may carry it). That would be a quick and easy.

    -Stick with 2n2222 transistor for the high power LEDs? TIP31 and 2n2222 (which I believe was suggested) seem to be the popular ones for 9V circuits like mine. But most of those are using less powerful LEDs.

2n2222 would be fine for the 1 Watt LEDs, as long as you run then at saturation and cutoff, not in linear range. The 3 Watt will be drawing around 1 Amp at roughly 3 Volts, so something in a TO-220 package would be needed  if you go that route.

    -LED and piezo connector suggestions. I'd like to be able to easily detach the piezo and LEDs from the "brain" of the device. Any suggestions on a good/cheap connector for this?

Visit Gateway, ask him for a 2 pin connector pair. Something at .1" pin pitch would likely do just fine. 1/8" audio connectors would do also; your current requirements aren't extreme.

    -Mounting recommendations. All of this stuff needs to be mounted to a Velcro strap that wraps around and attaches to one of four different drum stands. My thought was to have a single "board" permanently attached to the strap that houses the "brain": battery, resistor(s), transistor, ATTiny85 and two cheap connectors for the piezo and LEDs. The piezo gets taped to the drum and the LEDs would attach to the strap via their own Velcro backing to shine where they need to shine. If the LED Velcro idea doesn't work well, I'll just sew them into the Velcro strap. How good/horrible of an idea is this?

Sounds plausible to me. Is there tensioning/adjustment required when changing a drum head? Would it be feasible to pop the whole assembly inside? It would be neat if it's feasible.

[Chris Weiss]

what's the duty cycle of the LEDs? 3W is a fair amount of power, and
the hotter it runs the less efficient it is, which mean less light,
and less overall lifetime. you won't need a lot of heatsink though, a
strip of aluminum sheet should do well with the LED thermal epoxied to
it, or with the star LEDs a couple screws could do if you're careful
not to short it on the top side. you could have slots for the straps
in the sheet taking care of the mounting as well.

instead of transistors you might consider dimmable constant current
LED drivers, this will not only take care of the dimming, but will
also take care of the higher-power side of the circuit, and give you a
more consistent brightness and a longer LED life. might be cheaper to
have one bigger non-dimmable LED driver to power everything and then a
power transistor for each set of LEDs. you'll likely want to move up
to a FET, such as is used for motor control, to handle the greater
power, assuming 6 LEDs and 18W, that's a lot of power.

if you double up the conductors, cat5 cable would work well at 15W+
(based on POE specs, 100mbit POE uses the 4 unused pins for power, 15W
max. 1Gbps specs can go more watts, but also use all 8 pins on
heavier cat5e or 6 in a phantom power mode so it's not really
comparable), so 2 wires for the piezo and the other 6 split off and
combined for the LEDs. RJ45 stuff is pretty cheap too, just make sure
to well label it as "not ethernet".

[EschewObfuscation]

I don't think there's any thought of running 6 x 3 Watts; if you check the data sheets for 9 Volt batteries, they're not even spec'ed beyond 350 mA, which would be about 3 Watts total. Short circuit current is undoubtedly higher, but then we would be looking at considerable internal heating, possible device failure, and likely very low terminal voltage. A 9 Volt should be able to sustain 3 Watts for a significant fraction of an hour, which is IIRC the design goal.

Current source LED drivers are linear devices, which will add cost and burn power. I thought we were discussing pulse width modulating the LED(s). If we wanted to use an analog current source, that can be done with an LM317 and a resistor, at a parts cost of around 40 cents. But still a high power cost, as above.

If you cut up a cat 5 cable for wire, you will find you have 22 or 24 gauge wire. So you can cut up a cable, or take wire off the spool of 22 or 24. same-same.

[Chris Weiss]

I assumed the 9V cell was just for the proof of concept. multiple LEDs
we're clearly spec'd, so a stronger power source will be needed either
way.

[Joel Balestra]

So a bit of clarification. Pardon the bullet points.

-There will 13 independent devices created. Each with it's own battery, LEDs and "brains". They will not interact with one another.... for now.

-I'd either use two of the 3 watt LED's (one up, one down) or three of the 1 watt LED's (one up, two down). I'm leaning towards the 1 watts since those will be brighter for the wattage and are cheaper.

-The goal was to go with a single 9V battery to keep the size down. I'm open to alternatives though.

-The  duty cycle (if I understand this) will be ~250 ms followed by a fade out for ever strike. LED brightness depends on intensity of the strike so not all strikes will peak.

-How thick should the aluminum be? I assume this is using sheets/strips you can buy from the hardware store. I have the tools to cut heat sink squares.

-The devices will wrap around the stand below the drum. They can't go in them. These are sealed drums that use cowhide heads so light wouldn't get through anyways. The drums will be used for an hour show and the devices will be attached only for the final song so they need to be relatively portable and easy/quick to attach.

-Joel

On Friday, August 1, 2014 11:47:51 AM UTC-5, Joel Balestra wrote:

[EschewObfuscation]

Okay, 6 Watts total is most likely beyond the capacity of 9 Volt batteries (definitely outside expected range), though it's possible you might just get away with it in a "mad scientist lashup". I'd say you're looking at a 4 or 6 cell AA pack to reliably drive that load. Even that's pushing it a little bit, but I think you'll get away with it.

There is 1/8" aluminum bar stock at any of the home centers in the hardware department. Try cutting 6 inches, if that seems more than sufficient, cut off an inch, try again. Empirical engineering! (Or go with a small fan style cpu or gpu cooler, if you have a source of those cheap.) Or try a TO-3 fin assembly, those are available off the shelf. They'd likely do the job okay.

If you go with 1 Watt LEDs, and PWM them, a 2n222 will do the job, one transistor fro each LED. That's a 1 - 9 cent transistor, depending where you get them. If you go with 3 Watt LEDs, you need something a bit heavier, as you will be conducting 1 Amp, maybe a little bit more, during on time.

So, 3 x 1 Watt = 2n2222's sufficient, and you can get away with 9V battery (I think) for 3W total.

2 x 3 Watt = bigger transistor (one for each, or one for both, depending on the transistor chosen), and bigger battery pack, probably 4-6 x AA. Bigger transistor probably runs you 50 cents, plus or minus.

[EschewObfuscation]

BTW, a while ago I ran across a surplus offer of lead clipped and formed, house numbered 2n2222's at 1.3 cents each. They work fine in a bit of perf board, would require a bit of care if soldered random-wiring style. I posted the offer on the old board, but I don't think anyone was interested.

In any event, I could offer you the outrageously marked up offer of 50 of those for a whole federal reserve note. That'd cover your requirements, with 11 left over for spares and/or prototypes. Might even be talked into tossing in a couple extra, if you felt that too pricey.

One correction: if you are running a 9 Volt supply, those 3 1 Watt LEDs could all be run in series, if about 3 Volts each lit them up sufficiently. (They'd not be running at quite 100%.) In that event, no current limiting resistor would be required, and one 2n2222 could light up all 3. 9 Volts minus 0.7 Volt Base Emitter drop = 8.4 Volts / 3 = 2.8 Volts each. Alternatively, you could run one LED with one 2n2222 and two off another 2n2222. They'd require different values of current limit resistor, but could then be run at max. Two transistors required per unit, rather than 3.

[Chris Weiss]

9V are really made for low draw situations. you can likely run of a
9V, but it's not going to last more than an hour, and probably quite a
bit less. if it's just 1 song and long term is just a couple shows it
would be fine, but if you want to do a whole tour it's probably worth
getting some small lipo packs that can easily be recharged. 6xAA
would work as well, but be much larger.

[Keith Rogers]

I'll suggest another LED option that I see as having a lot of potential now that you're going with a mcu.  I recommend playing around with one of these WS2812 programmable RGB LED's.  I used one in a project recently and was surprised at how bright they are, especially with all LED's full on.  They have software libraries for implementing with arduino and there's no need to PWM the output, just command what you want.  And then down the road, I could see using a strip of these that send that drum pulse down a line, intensity of strike captured in a wave of color or brightness or both. 

--
You received this message because you are subscribed to the Google Groups "Arch Reactor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arch-reactor...@googlegroups.com.
To post to this group, send email to arch-r...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/arch-reactor/CAEhO07OYNHTwVCiP%2BRbUdz_NwRv2P%2Bs35POc4Z6vL9BrBcsmzw%40mail.gmail.com.

[Chris Weiss]

I have a 1M strip of 60 WS2812's I can show you at the next meeting if
you like. they are seriously bright, I usually run them at about 20%
brightness for close range indoor use. nice thing is the long strips
can be cut between any pixels and wires soldered on either to
reposition the chain or to make a new chain. an attiny85 should be
able to run plenty for your project, and you just need a good 4-5V
supply.

[Joel Balestra]

I'll try to make it this Sunday for the open shop at 7:00 with my updated prototype (not just a proof of concept anymore!) and sketch code. There's a slight change I might be bringing an "assistant" or two, but I've seen kids at the space before so I don't think that'll be a problem. The LED I'm working with is bright enough so I know what my target is when comparing other LEDs. Brightness is more important than RGB capabilities, but if a cheap bright RGB option presents itself I'll take it over a plane Jane white LED. I would be grateful to see other LED options to make sure I pick the right ones. EschewObfuscation, if that's the transistor I end up going with I'll be happy to take those off your hands. Thanks.

A few more questions/concerns for this project.

-LED flickering: My Arduino sketch code smooths out the value being sent to the transistor to avoid spikes and dips, but I'm still seeing some flickering in the LED. Having not worked much with LED's I'm not sure if this is just an LED thing, something in my hardware or something software/Arduino related. Any thoughts on causes for this?

-Serial VS parallel LED: It sounds like I should wire up the LEDs in serial, but is there a benefit for running them in parallel? More of a general curiosity question. Peak brightness is very important for this project.

-Four/Six AA VS Four/Six AAA VS One 9V: After picking the right LEDs, batteries are my next biggest concern. The battery pack can't be huge and I'm looking for LED intensity over longevity. The strap will be about 8-12" long and wrap around a stand that sits 1-2 inches from the bottom of the drum. If I need more batteries, I can work something out. But the less bulk the better.

Thanks again for everyone interest and help.    -Joel

PS: Here's the song I'm hoping to light up. The camera audio doesn't do it justice. Each drum would have a device under it lighting up the ground like a rocket taking off. The stage lighting would be adjusted to allow for the drum lighting to be more effective. I think the effect would be great for the solos and the speed up at the end.
http://youtu.be/te7o63j-TVY?t=1m25s

On Friday, August 1, 2014 11:47:51 AM UTC-5, Joel Balestra wrote:

[Chris Weiss]

series vs parallel is a factor of voltage drop. you can over-drive
the LEDs by using a smaller than ideal resistor which will make them
brighter but also shorten the life. you can use a LED calculator to
get the ideal values. http://led.linear1.org/led.wiz I woulnd't go
too much lower since you can burm them out. and unless the power
source has effective current limits I wouldn't go completely without
since that can damage the LED and the power source.

[EschewObfuscation]

On Thursday, August 7, 2014 12:57:49 PM UTC-5, Joel Balestra wrote:

... if that's the transistor I end up going with I'll be happy to take those off your hands. Thanks.

Aw man, I was planning my household budget based on that federal reserve note! :)

-LED flickering: My Arduino sketch code smooths out the value being sent to the transistor to avoid spikes and dips, but I'm still seeing some flickering in the LED. Having not worked much with LED's I'm not sure if this is just an LED thing, something in my hardware or something software/Arduino related. Any thoughts on causes for this?

If you are looking at the LED to see what the mpu is doing, you really don't know yet what's happening. Eyeballs are poor test equipment. Hook up a scope to the LED pins to find out what's really going on.

-Serial VS parallel LED: It sounds like I should wire up the LEDs in serial, but is there a benefit for running them in parallel? More of a general curiosity question. Peak brightness is very important for this project.

All that matters is correct voltage/current through the LED. 9 Volts is more than enough to run two LEDs with 3.something forward voltage drop, so a single transistor can be used there. The more power you are dissipating in current limiting resistors, the more power you are wasting.

-Four/Six AA VS Four/Six AAA VS One 9V: After picking the right LEDs, batteries are my next biggest concern. The battery pack can't be huge and I'm looking for LED intensity over longevity. The strap will be about 8-12" long and wrap around a stand that sits 1-2 inches from the bottom of the drum. If I need more batteries, I can work something out. But the less bulk the better.

If these are to be used repeatedly, then economics favor rechargeables. bgmicro currently offers a kit of 10 Sanyo 18650's and a charger for $20. Call that $2 per battery with a free charger. Checking the datasheet, those are spec'ed at 2.5 Amps discharge for 1 hour, start voltage 3.7, end voltage 3.0. Two such batteries could run your project (each 1 W LED would have its own transistor, no headroom for series connection), at a cost of $4 for the battery pack. You can pay that much for a 9 Volt alkaline battery if you're not careful.

[EschewObfuscation]

These would work the same as my 2n2222's and give you more lead. Cheaper, if you were already placing an order, so shipping would not be an issue: Suitable for 1 W LEDs, but insufficient for 3 W.

http://www.goldmine-elec-products.com/prodinfo.asp?number=G19973

[EschewObfuscation]

PS A pair of 18650's would *easily* support a pair of 3 Watt LEDs, considerably more, if you wanted to push it. (Remember that 2.5 Amp nominal current draw at 1 hour to exhaustion.) 2.5 Amps would support more than two 3 Watters, or several 1 Watts. No rule they have to all be the same color. 2.5 Amps x 6-7 Volts is 15 to 17.5 Watts.

Your Arduino may poop out as the batteries approach 6 Volts, if its regulator is a conventional model. A low dropout regulator swapped in would extend your run time by some incalculable amount, but likely have no effect until you passed the 1/2 discharged point, ie much longer than your current design goal.

[Joel Balestra]

Someone snagged a video from last night's short demo. The room wasn't as dark as the video looks, but the single LED worked marvelously. The LED was facing outward (unintentional) so I'd still want to use multiple LEDs facing up and down to get a better spread. Sunday we'll do some stress testing to see how a piezo holds up to that much stress and if the tape holds.

https://www.youtube.com/watch?v=Vk0FphQ71NQ

I've managed to easily break some of my piezo sensors so I need to look at a few more that are sturdier and figure out a way to reinforce the back where the wires connect to the sensor. I was thinking about attaching a thin disk of plastic to the back to protect the connections or using some sort of paint on rubber.

I'm also going to look into mini microphones (such as http://www.adafruit.com/products/1935) as another option for drum pick up. The huge benefit here is nothing needs to be attached to the drum head making set up quicker and greatly reducing the chance for failure. The obvious draw back being it won't be as accurate as a piezo sensor and may require some Arduino programming magic to reduce cross-talk from the other drums.    -Joel

[Chris Weiss]

plastidip works well

[Diane Sudduth]

This might be a good spot for some Sugru....

[Keith Rogers]

cool effect!!!  these just came out https://www.sparkfun.com/products/12999

--
You received this message because you are subscribed to the Google Groups "Arch Reactor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arch-reactor...@googlegroups.com.
To post to this group, send email to arch-r...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/arch-reactor/94cca815-f5da-4415-a1ad-d0d425b95c59%40googlegroups.com.

[EschewObfuscation]

So I'm curious, what design decisions did you ultimately make?

[Joel Balestra]

Sorry to disappear. Had a bit of a family emergency come up. The wife ended up in the hospital a week ago, but is home now and on her way to a full recovery.

So good news and bad news for the drum/light project. Bad news is I've decided not to do the project for our upcoming show at the Japanese Festival this Labor Day weekend. I came to my senses and realized that considering my experience, getting 13 devices built and stress tested in less than two weeks was a bit unreasonable.

The good news is we are still going to do the lighting project for a performance in October. That will give me more time for testing and tweaking to create a solid product. I'm going to continue researching and designing until this project takes off again in early September. At that time I'll start picking up some components like the ATTiny85s, batteries, transistors and LED's to play with on multiple drums at the same time. I'm also going to experiment with different trigger methods for the drums. All this may end up leading to a much more exciting and involved show next year, but I don't want to get ahead of myself yet.

Given the flexibility I need to handle multiple types of drums and the control I need over the LED's I'm going to go with ATTiny85s. For now, I like the idea of using using three of the 1W LED's (one up, two down) EschewObfuscation recommended for lighting the drums, but I want to take a closer look at the WS2812s. I've seen a ton of projects using those and I have a bunch of ideas for them. I like the idea of using the Sanyo rechargeable batteries instead of the 9V as well.

There's still quite a lot to be figured out and that'll happen once I start building again in September. For now I'm just drumming. I do have one question about housing the components. I've never built something with a nice enclosure. Is it feasible is design and 3D print a "box" for the circuits and to mount the LEDs (heat issue?) in or is it better to just find and customize an existing box? This box would be attached to a velcro band that straps to the drums stands. 3D printing is another area I'm super curious and excited about, but have no practical experience with. Thanks again.    -Joel

[EschewObfuscation]

3D printing will likely be a popular option, same "man with a hammer" principle, and it could make a slick looking assembly. You might cope with the "no undercuts" design constraint by making two identical or near-identical halves, held together by a through bolt, possibly secured with a wing nut?  I'd observe that PLA is chosen for most 3D printing mainly for its very low melting point, and 1 Watt LEDs will require a little bit of heat sinking, though not a crazy amount. 2-4 square inches of sheet aluminum would likely be plenty, so my feeling is you might well manage a hybrid design, PLA box/enclosure with grooves in the end to slide in bits of sheet metal heat sinking, something like that. Shouldn't get warm enough to soften the PLA, particularly in a 15 minute run time, but do keep the concept in mind, just in case.

As only about dozen are being made, and they don't need to withstand close inspection, I'd suggest also woodcrafting would serve your purpose. If you were interested in a bigger production run and slick cosmetics, another possible option would be hand crafted wooden model, then make a silicone mould for resin casting from that? Any of those approaches would likely do. Too bad, but premade 2XAA battery holders won't fit 18650's, and I've never seen similar units sized for 18650's offered, or something based on such would do nicely.

[Travis Megee]

https://learn.adafruit.com/gemma-powered-neopixel-led-sound-reactive-drums/overview

Travis

On Tue, Aug 19, 2014 at 12:51 AM, EschewObfuscation <google...@mailfilter.33mail.com> wrote:

3D printing will likely be a popular option, same "man with a hammer" principle, and it could make a slick looking assembly. You might cope with the "no undercuts" design constraint by making two identical or near-identical halves, held together by a through bolt, possibly secured with a wing nut?  I'd observe that PLA is chosen for most 3D printing mainly for its very low melting point, and 1 Watt LEDs will require a little bit of heat sinking, though not a crazy amount. 2-4 square inches of sheet aluminum would likely be plenty, so my feeling is you might well manage a hybrid design, PLA box/enclosure with grooves in the end to slide in bits of sheet metal heat sinking, something like that. Shouldn't get warm enough to soften the PLA, particularly in a 15 minute run time, but do keep the concept in mind, just in case.

As only about dozen are being made, and they don't need to withstand close inspection, I'd suggest also woodcrafting would serve your purpose. If you were interested in a bigger production run and slick cosmetics, another possible option would be hand crafted wooden model, then make a silicone mould for resin casting from that? Any of those approaches would likely do. Too bad, but premade 2XAA battery holders won't fit 18650's, and I've never seen similar units sized for 18650's offered, or something based on such would do nicely.

--

You received this message because you are subscribed to the Google Groups "Arch Reactor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arch-reactor...@googlegroups.com.
To post to this group, send email to arch-r...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/arch-reactor/03a735ca-f70b-49af-a5bc-55e9258c2714%40googlegroups.com.

[Joel Balestra]

That's actually the video I got the idea from using a mini mic for the drum trigger instead of a piezo sensor taped to the drum head. Unfortunately the drums don't open so I can't place the mic (or lights) inside where the mic would work best. The likelihood of cross talk or picking up another drum would be greater placing the mic just under the drum, but it would make installing the device easier and faster and greatly reduce the risk of failure. I'm really starting to dig those NeoPixels the more I see them in use. Fun stuff.    -Joel

[nfolken]

I thought the use of the neopixles was overkill for that project. Neopixles are great for controlling individually, but if you are lighting them all up identically like in that video, it would be a lot more cost efficient to just use a regular rgb led strip.

[EschewObfuscation]

These LEDs just went on sale again, short-run sale though. Decide quickly, and strike or don't; goldmine is not the most trustworthy supplier (ie they could well declare out/stock).

http://www.goldmine-elec-products.com/prodinfo.asp?number=G18580A

[Joel Balestra]

Hello, hello. Looks like this project is on again with a completion date of the week before Halloween for a performance on November 1st. I'm feeling pretty confident about the power supply and control aspects, but I'm still wondering about the lighting. I've been playing with a single 3W LED, but I'd like to see multiple less powerful LED's like the 1W discussed before. Are there any available at the space or would anyone have any I could see demonstrated? I saw some NeoPixels last time I was there, but they wont work for this application. Thanks.    -Joel

[Joel Balestra]

I got my first set up parts in. 15 ATtiny85s (two extra) and three sets LEDs to test including the 1 watt ones suggested earlier. Hoping to get those tested over the weekend and pick a set to use. I'm also probably going to go with 9V batteries for now. I found a hell of a deal online. I'll look into using rechargeable later. If all continues as planned I'd like to drop in again and get thoughts on it.

http://www.harborfreight.com/4-piece-9v-heavy-duty-batteries-68383-8114.html

[Keith Rogers]

Heads up on those batteries... "Heavy Duty" has a much lower capacity than the alkaline 9V bats you're probably used to.

--

You received this message because you are subscribed to the Google Groups "Arch Reactor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arch-reactor...@googlegroups.com.
To post to this group, send email to arch-r...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/arch-reactor/8cf5a33d-8e76-4baa-90bd-739440e10689%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

<TriggerParts.JPG>

[Joel Balestra]

Thanks for the heads up on the batteries.

I'm got another circuit question. I've been looking into more LED control projects and in some cases I've seen an adjustable voltage regulator used (specifically the LM317). My current design seems to work just fine without one, but are there any benefits of using one? It sounds like it's used for cases where a strong current is being fed in and needs to be limited. From what I can tell using a transistor to control the flow does the same thing. Sorry if this is an obvious question. I've much to learn.    -Joel

[Keith Rogers]

It looks like the LM317 is an adjustable voltage regulator with built-in current limiting capability.  The LM317 can be adjusted between 1.25 and 37 V DC and apparently you can choose it's max current by placing a resistor between the adjust and output pins (cool!).  LED's are usually spec'd to function at an optimal current and voltage drop.  As mentioned earlier, they aren't really intended to be dimmable via modifying the current but there is usually a small range of brightness control capability.  If you exceed their max current, they'll burn probably burn out.  Generally if you drive them too hard, they'll burn out much more quickly, and so you typically will add a current limiting resistor in line with an LED so that whatever the max voltage drop is that your application can apply to the LED, it cannot exceed the max current of the LED in series with that resistor (I = V/R).  I assume with this chip you could limit the current without needing an inefficient current limiting resistor.  At an amp or so for some higher power LED, the size of current limiting resistor could be huge.. P = I^2*R or... 1amp * R ohms equals something a lot more than 1/4 or half watt resistor.  However, I'm not sure I understand how this would eliminate the need for a transistor or some means of turning on / off the high load with the microcontroller.  Your transistor is probably being used more like switch to control the voltage at, and current flow through the LED.  If you don't use a current limiting resistor, and you exceed the max voltage drop of the LED, you're likely to push more current through the LED than it can handle and burn it out (or overdrive it and burn it out sooner than you'd like).  I'm not really an electronics expert so if I'm off on this hopefully someone will correct me.

To view this discussion on the web visit https://groups.google.com/d/msgid/arch-reactor/458ab113-48df-4986-b9c6-6c2a33255a00%40googlegroups.com.

[EschewObfuscation]

Short form: for your project, disregard the 317. Not really what you want. Use a current limit resistor to keep current within spec for your LEDs. Also, "heavy duty" batteries are zinc chloride, which is a major step down from alkaline. It may make sense for you to use them anyway, depending on price, just be aware you may have to parallel them for your application.

[Joel Balestra]

Ok, that makes more sense. My assumption was that the adjustable voltage regulator could be controlled by the ATtiny85 much like the transistor currently is. Thanks for clearing that up.