False Triggering from Fluorescent Lamps?

Thread Starter

acfcordero

Joined Jul 1, 2021
4
Hey all!

I'm working on a circuit that uses an MSP430F5529 microcontroller to control a stepper motor and 4x 40W fluorescent lamps. Here's a link to my circuit schematic.
https://drive.google.com/file/d/1XfEEVYC24WPUxfFD6geXqGXVXRaZSKL3/view?usp=sharing

I am using I2C to control a 16x2 LCD, a TB6600 stepper driver for the stepper motor, and then GPIO-controlled optoisolator(MOC3041)/triac (BT136) combo to supply power from an electronic ballast to the set of lamps. The I2C lines and push button are on 3 ft wires because my application requires it. I have an audio player on the device too (with button controls, also on 3ft wire), but I've had it disconnected for troubleshooting. I have a AC/DC adapter supplying 12V to my stepper driver, and a 12V-5V converter to power my uController.

The device goes between an idle state (motor and lamps off) and then a run state (motor and lamps are on) from a push button. My issue is when I press the button to go into its run state and the stepper/lamps come on, after a few seconds, sometimes immediate, it switches to idle state without button press (as if the button was pressed). This happens about half the time, and works just fine the other times. When the lamps are disconnected (motor still connected), the circuit works 100% of the time. The issue still occurs when lamps are connected and motor is disconnected.

My current guess is that in-rush current or noise on the line from the lamps is causing my button to trip. When the fail happens, I can see on my scope that the I2C signal is sent, so I'm guessing something is making the uC send the signal. I figured there was EMI getting picked up by my button or its wires, but I don't see anything weird when I probe the button output. Am I neglecting any decoupling caps anywhere? Playing with the capacitors of the triac snubber changes the frequency/speed at which it fails, does the capacitor type I use matter? What about troubleshooting on the triac, any suggestion for how to determine if it's operating correctly?

Any help would be appreciated!
 

Ian0

Joined Aug 7, 2020
3,754
Have you tried it without the snubber networks? 1Ω/560nF is an awfully low-impedance snubber. With that impedance there may be enough power going through it when "off" to keep the ballast slightly running. Try it without the snubber, or maybe with a 100Ω/47nF snubber with a Class-X capacitor.
Where are the capacitors that are shown connected across the pushbutton? Are they nearer the pushbutton, or nearer the MCU - nearer the MCU would be better, and maybe with a resistor to limit the current from the capacitor through the switch when the switch is pressed.
Just curious - but did you buy a job lot of 560nF capacitors? Are they all mains rated? I've never seen a 560nF class-X capacitor!
Is it the type of ballast that starts the fluorescent lamps immediately, or is there a slight delay? Does the spurious pulse happen at the point the lamp strikes?
 

Thread Starter

acfcordero

Joined Jul 1, 2021
4
Have you tried it without the snubber networks? 1Ω/560nF is an awfully low-impedance snubber. With that impedance there may be enough power going through it when "off" to keep the ballast slightly running. Try it without the snubber, or maybe with a 100Ω/47nF snubber with a Class-X capacitor.
Where are the capacitors that are shown connected across the pushbutton? Are they nearer the pushbutton, or nearer the MCU - nearer the MCU would be better, and maybe with a resistor to limit the current from the capacitor through the switch when the switch is pressed.
Just curious - but did you buy a job lot of 560nF capacitors? Are they all mains rated? I've never seen a 560nF class-X capacitor!
Is it the type of ballast that starts the fluorescent lamps immediately, or is there a slight delay? Does the spurious pulse happen at the point the lamp strikes?
I did try without the snubber before, but ran into an issue where the triac wouldn't turn off and the lamps would stay ignited, even after cutting power to the optoisolator. I'll try those RC values out and see how that works out.

The push button caps are on a protoboard next to the MCU, but I've found that adding the caps hasn't made much difference to alleviate the problem.

The .56u caps on the triac snubber came from a set of polypropylene film caps that I got off of Amazon, they are rated for 400V+. The ballast is an instant start ballast, so the light comes on right away. The false trigger happens between 0.5-2 seconds after ignition. If the lamp and motor work just fine for a few seconds, it doesn't usually fail after it gets going.
 

Thread Starter

acfcordero

Joined Jul 1, 2021
4
Thanks all for your help. I think I was able to figure it out. After I dropped the debounce capacitor value on the switches to around 50nF and added a resistor in between the switch and GPIO, my issue seems to have disappeared! Hopefully it stays that way when I put it all back together!
 

Ian0

Joined Aug 7, 2020
3,754
Thanks all for your help. I think I was able to figure it out. After I dropped the debounce capacitor value on the switches to around 50nF and added a resistor in between the switch and GPIO, my issue seems to have disappeared! Hopefully it stays that way when I put it all back together!
I think the large capacitance and lack of current limiting caused enough current to flow in the 0V connection to cause a problem.
 
Top