controlling external LED arrays with Procyon

[electronguy]

Andy, et al.

I just contacted Janice to order two of the Procyon lightframe adapters and wanted to clarify one issue regarding the output of the Procyon.

To control the current through each LED, I'm making an assumption that the Procyon uses Pulse-Width Modulation of the full 3V/18 mA (whatever it is) drive signal in order to control the "average" current through each LED. This is opposed to using a digitally-controlled current source output for each channel.

If PWM is used, then I'm guessing that any external control circuitry should be fast enough to keep up with the switching rate and control the external brightness proportionally. Perhaps a simple interface would be to pass the Procyon output into an RC averaging filter and then run that into the DAC input of a microcontroller for further use--for example, to adjust color balance between channels for large color-wash effects or to mitigate the different power/thermal requirements of the large output arrays.

Thanks,
--Scott

[CraigT]

My inclination would be to use an opto-isolator on each channel to protect the Procyon from any misadventure and just use FET drivers (or even bipolar high speed switching transistors) to provide the required drive current. A pot to adjust the FET drive should provide ample colour balance tuning. The PWM output of the Procyon shouldn't need any further conditioning. Obviously a seperate power supply will be required on the LED array side of the opto-isolators.

Cheers,
Craig

[electronguy]

Craig, thanks for the info. I think the optoisolator approach is the best way to go, so that's what I'll implement.

I have a number of solid-state relays, but I don't think they're quite fast enough. I'll have to review the specs. Also, the ones I have on-hand have zero-cross detectors in them, so I don't know how that would affect the overall timing/control if I used a 60 Hz AC line.

I'll likely stick to DC only.

Best,

--Scott

[CraigT]

Good choice! I suspect solid state relays are a bit overkill anyway. I don't know what sort of current you're after, but I'm very fond of ULN2801 octal darlington driver chips - cheap and convenient.

Cheers,
Craig

[electronguy]

Hi--Just a quick update before I have to get to my homework

My current plans to interface to the Procyon are as follows:

OPTION #1
========
1. Use a high-speed optocoupler
2. Create an output voltage proportional to PWM input
3. Scale voltage and trim as necessary
4. Pass analog signal into RCD-24 LED driver

OPTION #2
========
1. Use a high-speed optocoupler
2. Scale PWM
3. Drive PWM input of RCD-24 LED driver directly with isolated PWM
4. Provide analog dimming/current control as necessary to adjust levels

I'm still unclear as to the exact LED arrays I'm going to use. The Titan looks great but may be way too much for my needs. Any ideas on cheap low-ish power LED spots/linear washes to hide behind a speaker array?

Here's a picture:


Thanks again,
--Scott

[electronguy]

I just placed and order for six RCD-24-0.5 LED constant-current drivers and two HL16-RGB light engines. The light engines fit in a standard MR16 fixture, which is ideal for my situation.

I will design and build an interface to the LED drivers and spotlight outputs that will allow me to control via the Procyon, DMX, color organ, sound, etc. This will likely have a microcontroller, 16x2 character display for info/programming, etc.

To get started, though, I'm going to just have a prototype with the RCD-24's and some terminal blocks to output to the RGB lights.

The HL16's were available from Newark for $53+ each and I got the RCD-24's from Allied Electronics for about $16/ea.

I still plan to use optoisolators and vactrols (for analog voltages) for the real-world interfacing.

I'm hoping for results mid-December.

BTW, the overall goal of this project was so that I could "open" my eyes and get a better color representation with the light reflected off a white surface (so I don't blind myself). With my eyes closed, I mostly see red colors with the Procyon and I would like to better enjoy some yellows/blues/greens.

--Scott

[Andy]

Very cool!

Keep us updated on the progress. This is a very interesting project.

-Andy.

[electronguy]

[edit: PCB artwork is here! Jump a couple of posts down for the .zip file]

Hello forum,

This is just a quick note to let you know that I'm creating some very general artwork for a PCB that will interface with the Procyon. For reference, this is the most basic switched current source using a linear regulator, so unfortunately it dissipates a bit of wasted heat and power but is the simplest and cheapest.

Also, when I stopped development of my prior circuitry, I was relying on some misinformation that I found regarding the Procyon's output--ultimately, my problem here lies in not having a schematic and a dead oscilloscope, but I recently purchased a cheap USB 'scope and have verified that the LED outputs seem to be open-collector, which accounts for the "noise" that I noted in a prior post when I used an IR LED (from an optocoupler) as the input load. Because the IR LED was not forward biased when the output of the Procyon was "off" there was an observed noise.

For those who have made prior circuits, simply place a resistor, say 620 ohms, in parallel across each optocoupler input. I will update the "cheap" schematic and will include these pull-up resistors in the circuit board to come.

I have created a full circuit board with constant-current DC-DC converters but their PWM frequency is about 200 Hz and it looks like the Procyon is around 26 kHz for its PWM frequency. As a test, I'm writing a quick program in an Arduino microcontroller board (using a Mega right now) that will scale the 26 kHz PWM output of the Procyon to a lower-frequency (but same 8-bit resolution) PWM. It is expected that there will be a one-sample latency between what the Procyon outputs and the modified PWM output to the external high-powered LED driver--this is necessary because I have to determine the "average" duty cycle over this period.

Over time, I will make an interface that behaves as quickly as the linear circuit but has some of the power-saving features of a switched-mode power supply.

I am not charging any money to make these circuits available and will see what the best methods are for posting and answering questions that would otherwise be outside the scope of this forum. If any administrators would like to contact me, please feel free to leave me some ideas so that the members who are interested in this can benefit.

I still love my Procyon and use it often--especially when I can't sleep because I have too much on my mind. It's quite fun to watch the Procyon sessions with high-power external LEDs being shown on a wall (more of a color organ feature, I suppose).

Thanks for your patience,
--Scott Thompson

[Andy]

This is all VERY cool!

Did you ever get to making the circuit boards?

-Andy.

[electronguy]

Update:

Started to draw a fully linear circuit board but had a hard time (1) keeping the size of the heatsinks to something reasonable that would account for possible power supplies (e.g., 12V battery), and (2) being "okay" with a significant waste of power.

While exploring switched power supplies in the simulator, I wasn't getting the fidelity of what could be had with a linear solution. I keep track of some of the latest LED drivers and finding a 350 mA-750 mA driver that utilizes PWM dimming at ultrasonic frequencies is near impossible. Unfortunately, I just haven't had *that* much time to devote to creating a PWM driver that operates at 30 kHz max.

For your reference, some of my design goals are to use components that are readily available to most just by going to their local electronics hobby shop (e.g., Radio Shack, Frys Electronics, and their variants). If I begin to specify obscure and hard to find parts, there are a couple of problems: (1) they are likely going to be surface-mount devices (convenient to reduce the size of a project but difficult for novices to use / difficult to prototype), and (2) they will be expensive.

The main drawback with a linear only system is the heat dissipation. I may force some of the design issues, such as limiting the output to 1 W devices and hard-specifying the power supply so the wasted power is reduced. Another option is to use a high-efficiency buck (step-down) converter as the front-end to a linear system. This will likely be the best compromise.

One other option is to utilize a small microcontroller and RGB LED driver "backpack" called a "Rainbowduino" (see rainbowduino.seeedstudio.com). It is an 8x8 RGB matrix that could be interfaced with the Procyon--not as powerful to display on a wall, though, but very fun.

I have an existing circuit board that uses constant-current DC-DC converters but the PWM frequency is limited to 200 Hz. For anyone interested in this design, I will happily provide you with the circuit board artwork and bill of materials. I have a number of microcontrollers and over the next couple of weeks before my fall semester (my last one!) I will program one with a simple program that will down-convert the ultrasonic PWM frequency from the Procyon unit to the 200 Hz maximum PWM that the Recon converters use. I do not know what the lower PWM frequency will mean in terms of actual entrainment. The expected latency would be one 200 Hz PWM period (i.e., 1/200 s).

I'll try to get a photo out of the current circuit board working with a Procyon unit via a small microcontroller that converts the signals. The output is a 2" PAR-16 RGB 1W LED light.

--Scott