1

Typically, is the likelihood higher for a chip/module with Vcc = 3.3V to break if it gets switched on and off frequently over a prolonged period? For instance, on for 3s and off for 3s, continuously. That's almost 30k switches a day.

Are there (subtle) technical properties that could render the chip/module less durable, e.g., the silicon, the metalisation (if any)?

I'm asking because I wonder if it's a good idea to switch off a module whenever the MCU goes to sleep; the MCU can sleep and wake frequently.

EDIT: The part in question is CC2504.

Kar
  • 1,517
  • 1
  • 15
  • 35
WKleinberg
  • 173
  • 1
  • 9
  • 3
    would depend entirely on the component and what it does and what switching it causes. If it e.g. causes some flash write each time, check the flash datasheet for mtbf – PlasmaHH Sep 10 '15 at 21:08
  • 1
    I think you might have more luck getting an answer if you mentioned what the exact part was. With just the supply voltage to work from, it's hard to come up with a meaningful answer... – Matti Virkkunen Sep 10 '15 at 21:14
  • Sorry, it's the [CC2504](http://www.ti.com/lit/ds/symlink/cc2540.pdf). It's not clear to me if the it causes flash writes each time. – WKleinberg Sep 11 '15 at 07:04
  • Do you plan to switch on off Vcc totally? Do you plan to switch it to energy save mode? I think with a bluethooth radio there is a lot of things to keep in balance. Do you plan to use the radio in those 3 seconds? – DThought Sep 11 '15 at 09:34
  • Yes，the VCC of the bluetooth module will be switched off totally. That's because the even in power-down, the module I have is using a bit too much power. Whilst it's on, the radio will likely be used. – WKleinberg Sep 11 '15 at 09:46

2 Answers2

1

Failures due to power cycling are most often associated with Thermal Shock. If the unit is cold compared with its normal operating temperature, then the act of powering it can cause damage, but just how much is somewhat difficult to discern.

There are other failure mechanisms, such as electrostatic stress, but those tend to be associated with (generally) passive components.

If you have a unit that is being switched on and off quickly, but is in thermal equilibrium due to the average power (such as the regulator Andy mentioned because the thermal time constant is very long compared with the switching cycle time) then no thermal shock occurs.

Thermal shock depends on the rate of change of temperature of a device and the gradients it can cause (although the absolute temperature is also used to compute device failure rates normally using the Arrhenius equation). Long term thermal failures and thermal shock failures are different mechanisms.

In your case, the 3 seconds between power cycling is unlikely to change the temperature in any significant way for any given power cycle, and the unit will eventually achieve thermal equilibrium after many cycles in all probability, so regular cycling of a low power object might well have no problem.

A clue would be to find the effective self-heating of the device for each power-up and understanding the thermal time constant of the various parts of the device.

Peter Smith
  • 21,923
  • 1
  • 29
  • 64
  • What does the O.P. mean by "module" and "flash memory?" That SoC can sleep down to 400nA, just use that as opposed to turning it completely off. The datasheet doesn't mention flash endurance, but as long as the 8051 CPU core isn't writing data to the flash each power-up, that should last 20+ years. Most flash seem to have endurances of 10k to 100k write/erase cycles per bit. – rdtsc Nov 10 '15 at 13:32
0

Well, think about a high power synchronous buck regulator - it might be operating at 1MHz and be switching several amps every cycle for years. It all boils down to selecting the right components to do the job i.e. stress management of components.

Switching on and off a lightweight MCU every 6 seconds seems trivial in comparison but the devil lies in the undisclosed detail.

Andy aka
  • 434,556
  • 28
  • 351
  • 777