I am working on a project that will drive high loads (120AC) from a photon. I know that the best way is to use solid state relays. And this is how I plan to work it out with some SSRs for safety… (I know it is very dangerous to work with high loads).
However for learning porpoises I want to make a schematic and see if I got things right to hypothetically integrate my own SSR in my own PCB. On normal projects I get a proto board and try things out, but since this will never be made, I ought to squash my curiosity and learn from the experts…
So any opinion / correction / suggestion will be super welcomed! And I understand that whatever opinions I get it will be my risk to use, etc, etc, etc…
This is how the schematic looks like:
And this is how the (Theoretical) board looks like:
You need a lot more clearance between the 120V components/traces and the low voltage stuff.
Try to lay out the board with a clear area that has the 120V components & traces - and another clear area with the low voltage components & traces. The only things crossing between the two areas should be the SSRs. Rotate them as required to achieve this.
Do not extend the ground pours across the two areas. I’d avoid ground pours in the 120V area completely.
Is this better?
Question. Those bottom layer routes you see to the mounting holes are grounds… Should I avoid the grounds to the mounting holes since it crosses the High Voltage Side or is it ok to leave?
If this works, then the only remaining worry I have is the resistor Wattage as mentioned by @peekay123. Any suggestions in that regard?
You can consider thinning the low power traces between the opto and the power triac, they only carry a few mA, and that will help with your clearance. Extra credit for keeping the traces from pin 6 of the 3041M, and R6/R7 to the north (in the drawing above) of the keepout area. You want to maintain as much separation as possible between the two sections.
Whether to ground the mounting holes depends on what you are mounting the board to, and what with. However, I would not ground them using a thin trace back to your low voltage section, that is not a good safety ground. That thin PCB trace will look an awful lot like a really bad fuse, when you need it for safety reasons.
If the thing you are mounting to is conductive, then you can rely on that to ground the fasteners.
If it is a non-conductive mounting surface and your fastener is metal, then yes ground it, but consider adding a safety ground connector and not doing it through a thin pcb trace. You can consider eyelets connected to grounding wire, or thick PCB traces to another pheonix terminal block just for safety ground.
Or you can use non-conductive fasteners.
Lots of options to consider.
WRT the rating of the snubber components, the boiler is probably ~10A, but mostly resistive, the pump will be much lower, but inductive. If you err on the side of higher wattage, there’s little to no downside for you, because you are not pushing this into production, so the extra cost of the part is meaningless. Make sure you use flame proof/retardent high voltage resistors, and X2 caps in the snubber circuit (from the footprint, it looks like you are already using X2 caps.)
I recall reading once that, “Motor controllers are strange creatures, with stray currents running everywhere.” With my experience making a 0.5HP motor controller, I have to heartily agree. Perhaps the most frustrating bug I had to deal with was my microcontroller (PIC16F747) randomly resetting after the motor was turned off. 2 days later, I discovered that stray current from turning off a secondary small induction motor was somehow spiking the reset pin! The reset pin was tied with 1K to the 5v rail, plus a 0.1µF capacitor, etc. I ended up with 10µF on the pin, which solved the problem. And yes, the hand-soldered board clearly separated 120v from the logic.
Be prepared for some “delightful” tracking if things aren’t working quite right, and don’t be too quick to blame your code for unexplained actions!
If someone else has more experience with TRIACs, they’re welcome to speak up . Unlike a MOSFET, a TRIAC is a current-controlled device: I would say that R9/R10 are optional. Here’s what I used on my 0.5HP motor controller:
R4 was 1/4-watt. (Admittedly, I was “chopping” the AC wave like a light dimmer…probably needed snubbers to the moon and back )
I’d recommend at least testing your circuit in a (solderless) breadboard on 12vAC from a small transformer make sure the optotriac is biasing it correctly. As I recall, the TRIAC wouldn’t turn on if the optotriac was pulling the gate the wrong way.
R11/R12 snubber circuit values…I don’t have a meter that can measure AC amps right now, or I’d just plug a properly-rated 0.1µF into the wall, and measure the current draw! However, from the mathematics, a 0.1µF capacitor @ 60Hz should appear as a 26K impedance (= 1 / (2 * PI * 60 * 0.0000001) --note the capacitance in farads). Plugging 26.525K into Ohm’s Law for 120vAC gives us 4.5mA current (= 120 / 26525). 4.5mA current @ 120v turns out to be 0.54W (= 120 * 0.0045). You may want to go with a 1-watt resistor there. (Calculations from http://www.electroschematics.com/5678/capacitor-power-supply/ and my little “Ohm’s Law wheel”)
Of course, there’s the “easy” way: breadboard your circuit, try it out, and if something gets too hot, try reconsidering the value .
AHA–that circuit’s straight from the MOC3041 datasheet! As the MOC3041 is a zero-cross-detect optotriac, I don’t think you need a snubber network at all (C2, C4, R11, R12). All that it’ll do is add a negligible amount to your electric bill, a little bit to your BOM…
Basically, a zero-cross optotriac will wait for the AC line to reach 0v before turning on. The TRIAC (both inside the optocoupler and the BTA26) will wait until the next zero crossing to turn off (that is, if not still triggered). As far as I’m concerned, there shouldn’t be any spikes, etc., 'cause there’s no current to spike with!
Technically, there is a minimum current flow required through a TRIAC for it to remain on–and when the voltage reaches 0, the current disappears, and it turns off. Obviously, for my “light dimmer” style of motor control, I could not use a zero-cross detect optotriac!
EDIT: one more comment on the PCB. (Well, you asked for it !) I notice that most of the “turns” in the traces are 90° angles. Unless technology has changed from when I was reading about this, it’s preferred that these be comprised of two 45° angles, more like this:
Why? Because in the etching process, the etching acid tends to attack sharp corners, and you could end up with an open circuit where there was supposed to be a connection–or at least an inadvertent fuse location! Technology may have improved since then (these were hobbyist etch-it-yourself books), as I do recall seeing a VCR mainboard entirely done with 90° right-angles. Then again, KiCAD’s PCB software automatically inserts 45° bends in all trace corners.
Awesome insight… Although it got a bit complicated for me!
So what you suggest (And I agree) is that I should try this on a breadboard with low voltage (12volt or so)… Which I think it’s a great idea and I like that try first approach!
But… I don’t understand what you are suggesting… Should I remove resistor R9/10? Of which board layout are you referring to? (Some R names have changed from the first to the second, and I don’t seem to find R10 ).
Thank you very much for your help… Your schematic looks interesting.> What’s it for if I may ask?
What’s it for? Mhm, the fun of an elevator for my treehouse …a project that sourced more smoke than any other. I very quickly learned that when 600 watts was misdirected, smoke was a near-instant result. Dumb me, it took an instant triple-blowup for me to add a fuse to the power input! I forgot that an optoisolator (PC817) would break down at 60v…and from my computer keyboard, I inched the power up until a chain reaction resulted. At the time, I was using a 90vDC motor from a walking treadmill.
In the end, the elevator worked pretty good; I probably rode it over 2,000 times without mishap…although consigning myself to my code during debug wasn’t exactly my favorite job! There were so many control limits in the (assembler) source code that I was sometimes amazed that it worked! Overvoltage, overcurrent, overspeed, reverse direction…!
Anyway, I was just showing a portion of the schematic as an example of how I drove the TRIAC. You don’t need the transformer, bridge, current sense resistors, etc. Just the TRIAC, optocoupler, and a single resistor to the TRIAC’s gate.
I’m suggesting that the snubber circuit is not necessary, as your optocoupler is a zero-cross-detect (perfect for loads that you’re just switching on and off). The snubber circuit parts on your most recent PCB picture are (whoops–got a little confused!): C2, C4, R11, R6. I also submit that R10 and R9 are optional: unless you’ve got the world’s strongest EMI around your project, the TRIAC is unlikely to turn on without a couple of milliamps of gate current from the MOC3041.
P.S. updated my user profile photo to the elevator controller, housed in a repurposed computer power supply case. The MAC12M TRIAC is in the heatsink at the lower center. The large orange capacitor is the 4700µF 100v can-style main filter capacitor.