Looking for feedback on mains triac circuit design

[that_guy]

Hi there,

I've created a simple schematic for a 240VAC mains-powered triac dimming circuit and I'd like to get your feedback on whether I'm likely to blow anything up or whether anything can be done better. The relevant part of the schematic is attached. The circuit is based on Figure 13 in the MOC3020M triac optocoupler datasheet. Some points and questions:

• Input protection. The lamp is 1200W and can consume around 6A. I have a 10A fuse in my wall-plug. I'm assuming I don't need to double-up on the over-current protection by putting another one on the board.
• The heat-lamp has a built in AC cooling fan of unknown specification, which introduces an inductive element to the load, so I've included the triac snubber (R4/C2). Does this mean that an additional X2-class filtering cap across the mains is pointless?
• The BTA16-600SWRG triac package is a TO-220 and one of the main terminals is the center leg which is really tough to route a wide high-current trace to on my double-sided board. I'm thinking of creating a custom footprint spaced a little wider and bend the legs a little to fit. Has anyone done this before? What about increasing current capacity by routing a narrower trace on both sides of the board and connecting them up with a grid of vias?
• The Triac's TO-220 back-plate is not connected to any pin so I assume it's safe to screw on a large heatsink and then attach that to an earthed metal chassis. If I use metal. I'd really like to use clear acrylic if I can.
• The low-power side of the board needs 5V and 3V so I've run the mains through a 240/6V isolation transformer and a full-wave rectifier. This also gives me the opportunity to use a bog standard optocoupler (U1) to do the zero-crossing detection instead of one of those back-to-back AC things. I assume this will work but there may be a phase shift from the transformer to take into account?
• The input to the LM1117's should be (6*sqrt(2)) - 1.4 = 7.08V so there's not too much energy wastage in the dropout.

Thanks in advance for taking the time to read this. It's really important to me to get feedback on handling the mains part of this design correctly and safely.

[IanB]

The low voltage part of your circuit and the use of the optocoupler seems like overkill unless you are looking to provide remote dimming control from another system (but you did not mention this in your post)?

If the lamp cooling fan has an induction motor (is it small and quiet?) then it won't take kindly to being dimmed. You may want to power it separately from the dimming circuit.

[that_guy]

The low voltage part of your circuit and the use of the optocoupler seems like overkill unless you are looking to provide remote dimming control from another system (but you did not mention this in your post)?

Yes the dimming control will be handled by an MCU that monitors the ZSENSE zero crossing signal and switches the triac at the optimum point in the cycle via the ACTIVE signal.

If the lamp cooling fan has an induction motor (is it small and quiet?) then it won't take kindly to being dimmed. You may want to power it separately from the dimming circuit.

Good point. Pretty sure it's not an induction motor. It's certainly not quiet. Might be a good idea to separate it out though as there's no benefit to dimming it.

[IanB]

The low voltage part of your circuit and the use of the optocoupler seems like overkill unless you are looking to provide remote dimming control from another system (but you did not mention this in your post)?

Yes the dimming control will be handled by an MCU that monitors the ZSENSE zero crossing signal and switches the triac at the optimum point in the cycle via the ACTIVE signal.

No, that's not what I mean. It doesn't make sense to use an MCU to perform cycle by cycle monitoring and phase control like this. An MCU is best used for high level supervisory control, not low level loop control.

Remote dimming control means you might have some kind of external 0 to 5 V signal (for example) that you want to use to control the brightness remotely.

In general you seem to have a very complicated circuit with more parts than necessary. Does it need to be so complicated? Without context it would be hard for someone to give review comments.

[oldway]

Primary of T1 must be feeded directly by mains, not after the 1200W lamp.

[wraper]

Yes the dimming control will be handled by an MCU that monitors the ZSENSE zero crossing signal and switches the triac at the optimum point in the cycle via the ACTIVE signal.

How it is supposed to detect zero crossing if there is 470uF in parallel to emitter of the optocoupler? This is not functional circuit. If you are using transformer, there is no need for optocoupler to detect zero crossing, it is just wasting parts. If using SMPS, then better to use optocoupler with AC input such as H11A1 at the mains side.

[wraper]

There is another error. T1 primary winding is connected in parallel with the triac and in series with the load. Once you open the triac, your circuit will just shut down because of power lacking. There will be virtually no voltage on the primary of T1.

[that_guy]

Primary of T1 must be feeded directly by mains, not after the 1200W lamp.

That one is embarrassing,  thanks for spotting it. ?

[picandmix]

Why not simply use a commercial SSR  thats  all safely enclosed in a strong housing.
It can be driven directly from your micro via a small series resistor
http://uk.farnell.com/omron-industrial-automation/g3na-210b-5-24dc/ssr-10a/dp/1181259

Again as Ian has said, you will need a separate power line for the fan, as many do not like being driven that way.
However some do, the easier way to test them is to try a simple household light dimmer  and see if it works ok.

That said, for its power use, is there any point in running the fan at a lower setting, the cooler the lamp is the better.