High voltage hardware question

I have a PIC that I have interfaced with a 74VDC system using a voltage divider that is working very well with one exception. One of the inputs comes from a buzzer and it must create a lot of eletrical noise. It's one of those big, beefy buzzers with a coil that takes a good amount of current to make it work. I can't get an oscilliscope to the box anytime soon but I am assuming that the electrical noise is feeding back to the power supply on the PIC because when I activate the buzzer, the PIC will recycle the power. I do have some capacitors right at the input to the PIC but I guess that's not enough. Is there anything I can try without going into a bunch of formulas or trial and error. I guess I am not trying to reinvent the wheel here. As long as the buzzer doesn't turn on, the system works perfectly.

Chris

Are you going straight from the voltage divider to the PIC, or do you have a regulator in there someplace. If no regulator I would guess that the buzzer is causing enough of a voltage drop to cycle the PIC. If you are using a regulator, what voltage is feeding it?

Hi,

For the power supply (3V) for the PIC, I am going from a voltage divider, to a regulator with a filter then to the PIC. For the digital inputs, they are going from the voltage divider directly into the PIC. I guess I am not sure if the input for the buzzer is causing the PIC to reset or if the buzzer is somehow feeding back into the regulator, which is causing it to reset.

Chris

Actually, a schematic will help here. However, you can try to clear the supply line as shown here. It usually works in noisy enviros like automobiles.

Jerson

Hi Chris,

It is likely that the buzzer is generating a lot of spikes. If so it is possiable that the spikes will get onto the regulated supply through the internal clamping diodes in the PIC's inputs.

I would look at suppressing the buzzer input.

good luck!
Bob

To suppress the input, do you think I can just get a 100VDC cap and put it on the buzzer input or is it more complicated then that? The problem is that all of the inputs and the buzzer use the 74VDC for power. Maybe I should have used an optoisolator but I was trying not to add additional components. Any ideas?

Quote:

---

Originally Posted by Christopher4187

Hi,

For the power supply (3V) for the PIC, I am going from a voltage divider, to a regulator with a filter then to the PIC. For the digital inputs, they are going from the voltage divider directly into the PIC. I guess I am not sure if the input for the buzzer is causing the PIC to reset or if the buzzer is somehow feeding back into the regulator, which is causing it to reset.

---

Like Jerson said, a schematic would really help.

A big cap at the buzzer might help... but.

1- 3 volt regulator fed from voltage divider. Input voltage I would guess is higher than 5 volts?
2- Voltage divider directly feeding inputs. What is this voltage?
3- If the divider is set for 5 output with 74 input a spike would over volt the inputs, and if 5 volts is enough to run the regulator, a voltage drop would also drop the regulator out.

Try not connecting the inputs of the PIC. What happens? If works when the buzzer is on try clamping the inputs.

You could try and fix it in the software, Make the pic trigger with a known source and then wait a set period of time (eg 8ms) and see if the source is still there. DC spikes are only about 4 ms long and looking for a solid signal could help.
Snap

Just a thought,

You say the buzzer has a big coil in it. Maybe you need a diode on the pic pin to supress any back emf from the buzzer?

Squib

I have had some really horrible spike problems. A lot of them have been totally cured by fitting a 100n ceramic capacitor across the PIC's power pins, as close as you can get.
Hope this helps.

Here is what I have done.
Bypassing, bypassing, bypassing. Make sure you have plenty of bypass capacitance on your pic power pins.
In really bad situations I have had to put a small inductor on the +5V line between the regulator and the pic.

Good luck

Ground plane, ground/power routing, splitting power/digital/analog ground, PSU bypassing, PCB layout... tons of possibilities here.

Overloading PICoutput/PORT is still possible. The whole schematic and actual layout could be helpful.

Try adding an inexpensive small relay, or optocoupler, in parallel to the coil, then use the contacts as your signal source. Simple easy and I use this type of isolation in industrial welder controls.