Step 6 : How to hook up AC power

At the bottom of the board you will see 22 terminals as follows:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
8 7 6 5 H1 H1 4 3 2 1 8 7 6 5 H2 H2 4 3 2 1 (around corner) “unlabled” N

It is assumed that the “white wire” or neutral wire in the AC system will never be attached to the printed circuit board but they will all be attached together in a wire nut (not included in parts list) or a wiring block. The only exception to this is that a single neutral wire must be connected to the “N” terminal in order for the zero cross circuit to work. The reason there is an extra terminal next to the N terminal is because having a single terminal is bad for mechanical stability. Having two soldered to the board makes it sturdier. The unlabeled terminal is not connected to anything. I could have connected it to neutral as well, but I had visions of someone screwing a black wire in there, assuming that is where the power goes in… POOF.

When attaching the “black wire” or hot AC wire, in the simplest case, it would be connect to one of the H1 terminals, and then another (preferably black) wire would go from the other H1 terminal to one of the H2 terminals. In this way, you are supplying 120VAC to the “center” of both 8 channel groups.

All of the “hot wires” for your lights (the outputs of the SSR’s) are screwed down on terminals 8-1 on the left, and 8-1 on the right. These are your “channels”.

If it turns out that each side of this controller does not explode when you run 15 or 20 amps through it, (assuming you have installed good heat sinks), then you could potentially have two separate feeds from your circuit breaker box (Must be GFI protected for outside use) and provide a separate 15 or 20 Amp feed to the H1, and a separate 15 or 20Amp feed to H2. Having said that, I have no idea under what conditions this board will become a flaming projectile. Experiment at your own risk. I personally error on the side of safety and will only be running about 1/3A to 1/2A per channel, but it was fun to completely over design the board, and it will be novel and interesting to find out at what amperage the board ignites.

Step 7 : Cost to Build

The cost of the PCB itself will be between around $20 each, perhaps as low as $14 if enough people join in. I will make no profit, I will sell them at cost, plus shipping, and paypal fee’s. The parts are going to be $46 per board (not including shipping and tax if applicable) from Mouser. However, $13 of the $46 is for optional components. You can decide what to install if you would like to keep your costs down. Having said that, sockets for the DIP’s and LED’s really should not be left out of the design, you could cause a lot of pain for yourself by leaving those out. Read the comments in the parts list, I’ve indicated which optional parts are truly optional.

The parts list can be found here:
http://www.ritzfam.com/ChristmasMadness/Renardx16SSR/Release

Look for the latest revision of this document: 16cPicDimmedSSR_Parts_List_RevX.X.xls

Step 8 : How much power can it handle?

There are online trace power handling calculators, but they don’t agree, and neither can we. The traces are on the top and bottom of the board as well as being as wide as the design will allow. The power distribution traces are ¼ inch wide on the top and bottom. The traces out to individual triacs, and out to the screw down connectors are .08 top and bottom. ExpressPCB claims to have a 1.25 oz/ft^2 copper process. The triacs appear (From the experiments of those who do this regularly) good for up to 1A without a heat sink. And It seems reasonable that you might be able to get up to 3 or 4 A (each channel) with a good heat sink. I think someone (me?) will need to sit down and run a duration experiment until one of these things explodes, only then will we know the true maximum limit. For now, the parts list will ship with a 5A fuse (5mm x 20mm common glass tube type), and you can make your own decision about how to handle the power issue.

A template to create your heat sink is provided here:
http://www.ritzfam.com/ChristmasMadness/Renardx16SSR/Release

Look for the latest revision of this document:
16CPicDimmerSSR_HEATSINK_TEMPLATE_300DPI_RevX.X.jpg

Step 9 : How much power does it use?

(Despite all the calculations below, I find my entire built controllers under worst case conditions draw about 25mA each).

For this configuration
(May not be identical to yours)

5VDC Power Requirements in mA
LM7805 8mA
ST485EBN 20mA
MOC3023 - 16 * .008 / 256 = 0.5mA
H11AA1 - 5V / 27K = 0.2mA
PIC16F688 =(HFINTOSC mode = 8 MHz) = 10mA
Osc = 35mA (Optional, not currently used)

Total with option Osc 73.7mA
Total without Osc 38.7mA


The traces on the board are setup such that the orange and orange-white wires are spares. Should you require them for more current carrying capacity to the controllers, there are traces close to the related RJ45 connector pins that you can just drop a ball of solder on, and you can now have 3 VCC wires and 3 GND wires in your Cat5 cable.

Step 10 : How many channels can I use at once?


Phil Short and I have taken this design to 38400 bps with no problems. Thus at 25mS resolution you can drive 64 channels through a single serial port.

The math:
Theoretically, at 19200 baud it is possible to get up to 192 channels using 100ms resolution, but in the real world, with 100ms resolution 128 channels use 85-90ms of the 100ms cycle time. Thus you are at a 85% duty cycle when running 128 channels. In the picture above, Vixen is driving the serial port (RS232) with 100ms resolution and 128 channels. The width of the squares is 100ms, you can see what I mean by the statements above. In order to go beyond 128 channels at 19200 baud, you have a couple of options. The fastest and easiest is to just use another serial port. For each additional serial port you add, you can use another 128 channels. Vixen can natively support up to 4 serial ports, and some computers already have two serial ports on the back. Another option is to reduce your resolution from 100ms to 120ms for example. That would give you room for (theoretically) 20% more channels.

Phil and I continue to investigate using the external oscillator to push the frequency higher than 38400 bps, but that is a little bit harder, and will take some time and energy. I think we can have something running 57.6Kbps by April perhaps, maybe faster.