3
\$\begingroup\$

Our high school theater is hosting a production, and a big part of the set is a giant turntable. It's going to be cool and all, but I've been challenged with controlling LED strips on the turntable over DMX, and I can't use any wires. Any excess wiring cannot be stretched across the turntable, either, and the lighting will be in three generalized areas.

I've already looked into standardized LED controllers combined with wireless DMX, and then connecting it all to deep cycle car batteries, but our theater isn't too hot about spending the money for all the equipment. We've also looked into renting, but it seems like what we need is too custom, and we'd be buying the LED strips so it would be great to be able to reuse them at any time.

So, after much thought (and still contemplating) I've decided that we could build three boards that have some form of PWM driver on them to switch MOSFETs. The three boards would all need to be controlled digitally, through SPI or I2C. I'm open to other methods of control. This digital line would be mastered by a raspberry pi, and it would connect to another raspberry pi over a standardized wireless protocol (maybe wifi?). I can script. The second raspberry pi (in the booth) will connect over DMX to our lighting board. The lighting board is an ETC Ion, and I'm almost certain the DMX is completely standard.

The lighting strips that we'll be dimming are from amazon, http://www.amazon.com/gp/product/B00W0VEOP4?ref_=cm_sw_r_awd_XlWawbR5CFVDD. Each strip is RGB. In each of the three general zones, we'll have three subdivided zones. So on each board that I make for each general zone, I'll need 9 channels of PWM/9 MOSFETs.

As far as PWM drivers go, I've been searching through different DIY websites like adafruit and through different manufacturers, like NXP and Ti. I'm trying to find a driver that can get a decently high frequency. All I've been finding are SMD chips. I found a SMD chip already attached to a breakout board sold by adafruit, meant to control servos and it would probably work for this application, but is was $15 and the chip on it was only a couple bucks. I need a DIP chip because there is no way I'd be able to solder the SMD with any method. I'm not good enough to do that. Besides a specified PWM driver to do the job, I've also considered using a couple of microcontrollers, like the ATmega328P-PU which has become standard through arduino, and ordering them with the bootloader already programmed in and using a couple arduinos I have lying around to program them. However, it would require extra components, more work, the board would be more complex only to waste the full potential, and it nears the cost of the adafruit breakout board. Not to mention, I'd need two of them because each controller can only power 6 (hardware based) PWM outputs.

My first question: what should I use as a PWM driver? Where can I get one?

For the MOSFETs, I'll need some pretty high powered devices. Each generalized area will have 15 amps worth of LED strip, at 12v each. I'm not knowledgeable on RGB leds, so I don't know if it'll pull full current with only 1 color on. But just to be safe, It'd be nice to have MOSFETs that can handle 6 amps each. I'm hoping it'll also help to overshoot with the MOSFETs with the whole heat situation thing, where switching accuracy decreases. I'll probably be able to add a standard 12v computer fan to the enclosure, maybe two, and bolt some washers onto the MOSFETs for extra heat dissipation.

My second question: what MOSFET should I use? Switching side doesn't matter, but I think I'd prefer high side. I'd prefer a common MOSFET.

My third question: what kind/what size capacitor should I use/should I use one to balance/average the PWM output from the MOSFET? The show will be recorded using high quality cameras, and I don't want the camera capturing a frame where the LEDs are off and a computer to pick that frame while downsizing the framerate (whatever that is called). It'd be bound to happen more than once.

My fourth question: since the MOSFET needs to have such a high amperage, would it be correct to assume I'll need a transistor to switch it on and off? Also, should I use some form of isolation? I read a thread in the this forum, http://www.avrfreaks.net/forum/how-use-opto-couple-pwm, and it said that octocouplers generally are not fast enough for high frequency PWM.

For power to the entire setup, each area is going to have to have deep cycle car batteries. We'll be using standard car battery chargers, and they'll have to last for 4 to 5 hours. We'll probably be getting huge 80-100 AH batteries, and I'm hoping that'll cut it. On top of the batteries, I've been scrounging ebay for a decent buck/boost DC DC converter to get as much time out of them as possible and make sure we don't burn out the LEDs.

My fifth question: does anybody know of any good buck/boost converters to make sure the batteries will last as long as we need them, and will we destroy the batteries if they drop to too low a voltage?

That's pretty much it. If you read this all, THANKS A TON, because it was pretty long-winded for a forum. If you can suggest anything or answer any of my questions, that would be GREAT. Please remember, I'm just a high school sophomore, so I won't understand high level electrical engineering terms.

\$\endgroup\$
  • \$\begingroup\$ I think your RaspberryPi's GPIO would make for a fine PWM controller, provided you'd be willing to put in the time to learn a bit about driver programming. \$\endgroup\$ – Adam Sep 27 '15 at 21:02
  • \$\begingroup\$ I've considered that, but I need 27 channels and they're split into 3 groups of 9 that are going to be physically separated. \$\endgroup\$ – Aaron Fuller Sep 27 '15 at 21:06
  • \$\begingroup\$ You could still use multiple RPis to manage each of those groups. Set up a cross compiling environment and create a character device interface that allows you to change the duty cycle at which the driver toggles the GPIO hardware, then have a script/application that receives settings from the master. \$\endgroup\$ – Adam Sep 27 '15 at 21:10
  • \$\begingroup\$ I could definitely do that. Unfortunately, even with the cheapest RPis, it would be my most expensive option. It's also going to be more complex, and the RPis aren't going to be as good at PWM like an IC specific to PWM will be. \$\endgroup\$ – Aaron Fuller Sep 27 '15 at 21:14
  • \$\begingroup\$ does this help in any way? ti.com/lit/ds/symlink/tlc5940.pdf \$\endgroup\$ – Wesley Lee Sep 27 '15 at 23:44
1
\$\begingroup\$

The problem with everything that you have listed here is that your borderline to a full custom solution and you have left very little to compromise with. But i'll try to fill in all the blanks.

  1. The PWM driver i would recommend is this one https://www.sparkfun.com/products/10615 It can support the voltage directly up to 17v. This would eliminate the need for opto couplers.

  2. You want a N-Channel MOSFET and to switch it on the low side. N-Channels are cheaper and have lower ON resistance. Find one that can support the current and voltage that you need.

  3. Don't use a capacitor. The capacitor looks like a short to the MOSFET when it's not charged. This can over current the MOSFET.

  4. You are partially correct. The gate of the MOSFET looks like a capacitor. You will need a resistor in series with the gate to make sure you don't over current the driver. Beyond that as long as the gate can be pulled sufficiently high enough the current of the driver only effect the switching times.

  5. Your math doesn't add up here. You state that each channel will draw 3A and at 27 channels you get roughly and hour of run time from the 100Ah battery. Using a boost converter may actually lessen that amount run time because there will be efficiency losses related to the converter. Also a converter that's capable of the kind of current that you want would be way over your budget. Run straight from the battery.

\$\endgroup\$
  • \$\begingroup\$ Thanks for the suggestions. I think I'll use the TLC5940 in the NT package to save some money. Also, for the capacitor, I understand what you're saying, but what if I run a small resistor in series with the capacitor? Would that fix the problem with shorting at no charge? Other than that, I think I'm pretty much set. Trust me on the math though, I kind of summarized it incorrectly in my OP. Sorry about that. \$\endgroup\$ – Aaron Fuller Sep 29 '15 at 13:28
  • \$\begingroup\$ @AaronFuller Adding the resistor gives wattage problems. If your system is 12v and you have a 6A MOSFET, for example, you would want to limit the current to 6A that requires a 2 ohm resistor, sound simple. In reality that 2 ohm resistor would be dissipating 72 watts of power (V^2/R). If increasing the resistance until the wattage was under control then the resistance makes the capacitor pointless. You have the right idea of forming a low pass filter with a resistor and capacitor but the fact that the dropping resistors are built into your strips makes that difficult. \$\endgroup\$ – vini_i Sep 29 '15 at 16:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.