[General] Flash Memory Corruption after Power Outages

[AlwaysRedE2Ring]

Dear Forum Members,

Our church uses three Rabbit 2000 embedded controllers to activate solenoids to ring our bells. They work fine until a power outage. Occasionally one of the processors will display code instead of showing certain programing options. Plus, the ring time settings, although they appear to set, doesn't function. The manual ring options work fine. I replaced the small coin type battery backup batteries. I believe they keep the time settings current, but the function settings are stored in flash memory. Any suggestions on preventing this problem. I had the processors on a power strip that, upon further inspection, doesn't appear to have a surge protector. We plan on getting a good electronics grade surge protector installed for use on all of the processors. We have had the system designer reinstall programing on one corrupted processor and it is since working fine. Again, the problems don't happen until a power disruption. Any suggestions are greatly appreciated.

P.S. We are even considering the purchase of a line interactive pure sine wave UPS (1500 VA at 1050 W) to maintain power. When the processors are not activating solenoids to ring the bells, they only draw about 5 W total. The 2 little bells upon activation use 250 W and the one big bell (two solenoids) use 500 W. The activation period is between 60 and 90 milliseconds. The only time more than one rings is when AC current is available and then manually.

Sincerely, AlwaysRedE2Ring

 

[BradtheRad]

Does the system have a startup/reset button? Do you know whether the system performs an internal reset/startup procedure when you start it normally...

As compared to when mains power comes back on without performing the reset/startup procedure, and hence sections of the circuitry assume unpredictable states?

I do not know for certain, but the solution may be as simple as installing a simple delay circuit at the reset terminal, to hold it low for a second or two after mains power is restored. Then it can execute a stable reset/start sequence.

 

[AlwaysRedE2Ring]

The system does indeed go through a startup procedure where it identifies the version, sets up receiver (to sync with atomic clock transmitter for time of day), states user functions are now complete. Could circuitry unpredictable states be caused by voltage surges associated with power disruption and subsequent resumption (sometimes multiple AC on/off cycles in quick succession)? One of my processors now shows code during part of the startup sequence. On another program setting when I attempt to setup receiver (i.e. atomic clock) I also see "code" as opposed to normal setup prompts. Consequentially, programmed ring settings do not function. I had a similar system in a small chapel in the country. One New Years Eve I set the bell to ring at midnight, but forgot to turn off the dusk to dawn light that was on the same circuit. I had a similar problem and had to have the program reinstalled by the designer. Plus, I thought my present three processor system was running off a surge protector/power strip, but upon further inspection now believe it was just a power strip. I plan on getting a good surge protector made especially for sensitive electronics. Maybe such surges could be causing the instability?

 

[BradtheRad]

1.

Yes, it is possible that when mains power comes back on, it is interrupted by several rapid 'switch bounces'. These might carry spikes/ hash/ dropouts through an attached DC power supply.

A power line conditioner could go a long way toward helping to solve your problem. Perhaps a surge protector.

(Note: The varistor, or surge absorber, is only able to absorb so many spikes at a certain energy level, and then its useful life is over.)

2.

Static charge is a hazard if your system is exposed to the elements or located high above ground. It may or may not contribute to your problem.

Anyway a static charge arrestor is typically recommended. (The same kind as for TV aerials.)

3.

Some pins on the microcontroller may need to be tied to a supply rail, particularly input pins. If they are left floating, they can adopt unpredictable volt levels.

Guidelines state that unused input pins should be connected to the power supply or to ground.

4.

The reset switch may need attention, either regarding item #3 above, or in regard to keeping it activated until the power supply settles.

(I have no direct experience with microcontrollers, but I did install a reset switch in my Commodore computers.)

There have been previous threads asking about a delay circuit at the reset pin, in order to ensure that a microcontroller will wait for a moment after power-up, before it starts running.

Simple method that will work in the case of a reset pin that needs to be pulled low to halt the microcontroller. The IC begins to operate when the reset pin goes high.

Component values are not critical. The high-ohm resistor is added so the capacitor can discharge during a blackout.