Can undervoltage hurt a circuit?

[alank2]

Hi,

A friend of mine was looking at a broken board with me and he was using a variac to bring it up slowly instead of just hammering it with full strength mains all at once.  One question I have (and I probably should have asked him) is, are there circuits that will not be happy with being fed 10V, 20V, 30V instead of 120V and while I understand they likely won't work properly, could under voltage cause damage to components?

Thanks,

Alan

[Zero999]

Depends on the circuit. AC motors don't like undervoltage because they stall and can draw excessive current. Some switched mode power supplies also draw more current when subject to undervoltage.

[bd139]

Incidentally the switching power supply failure mode is a function of watts usually. If you have to say deliver 10W to a load, at 10V, that's 1A at the input assuming 100% efficiency (ideal case here!). If you then half that to 5V and the switching supply doesn't shut down and the source can deliver the current then that's 2A going through instead. Any non conservatively rated components tend to go to silicon heaven then.

MC34063's are notorious for this as the switching transistor built into them is fragile.

[PA0PBZ]

Yes, we once had a product that when used with almost empty batteries made the boost converter release the magic smoke. Turned out that the battery monitoring was set a bit too low...

[alank2]

It sounds like the "short" of it is that it depends...  Is there something that you can use to limit current after the output of variac?  Obviously a fuse comes to mind, but something more like a DC bench power supply that would let you say, 500mA maximum...

[rsjsouza]

Some older processors that require power sequencing may be locked up or in a brownout state as well, depending on how the different voltage regulators start up in undervoltage conditions. I don't recall anything permanently locked up in these conditions, but I am not so sure if devices that are flash/EEPROM programmed at power up  can be damaged permanently (or recall a great deal of effort to come out of depletion mode).

[Ian.M]

Its anecdotal, anonymized and not mine, (Reddit: Tales From Tech Support), but here's the tale from a chip manufacturer's support tech.  Parts:-

• That Time I had a Lift Commandeered
• At Thine Own Peril

I've summarized the technical issues below.

<spoiler space>

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.
</spoiler space>

TLDR:  Customer ignores chip manufacturer's datasheet power sequencing requirements - rash of premature field failures occur: Increasing tendency to crash, requiring a hard reboot with every power cycle + its (long odds) Russian roulette  with every cycle if  it will 'brick' it or let the holy smoke out! 

[fourtytwo42]

Hi,

A friend of mine was looking at a broken board with me and he was using a variac to bring it up slowly instead of just hammering it with full strength mains all at once.  One question I have (and I probably should have asked him) is, are there circuits that will not be happy with being fed 10V, 20V, 30V instead of 120V and while I understand they likely won't work properly, could under voltage cause damage to components?

Thanks,
Alan

Definitely YES, particularly power switching circuits using power mosfets, if there is insufficient gate voltage available to sustain saturation the device can fail due to excessive power dissipation. Of course if the circuit as a whole has been properly designed there should be an undervoltage detector to prevent such a situation but then properly usually comes at a price

[babysitter]

The first version of the great little yaesu ft-817 ham radio transceiver had a bug concerning undervoltage:

IIRC, the final amplifier stage of the transmitter is connected to battery voltage, and if the battery runs down to a certain voltage level it started oscillating which lead to a early demise of the expensive final MOSFETs. Possibly a voltage regulator going nuts near their gate was a culprit, too.

BR
babysitter

[babysitter]

And it happened to mine this year. I am just taking a break after removing the old final amp and soldering the new one and searchbrought up my own post.

[Psi]

You can also damage things in some systems that need +ve and -ve supply if you power the circuit up with one of them missing.

A long time ago i wanted to run a pc motherboard in the car from 12V. I knew I would need to generate +5 and -5V but was curious if I also needed -12V too, so I powered it up without the -12V ATX wire connected.
Everything worked except for the soundcard which wasn't happy and blew a transistor, which i then had to fix. 

[ifonlyeverything]

SMPS without proper under-voltage lockout can blow up if you spend too much time at low voltage. MOSFETs won't fully switch on and off due to insufficient gate-source voltage.

[p.larner]

i seem to recall the mel prc-320 radio unit 5 psu was common to fail when the nicad pack was low due to more amps being drawn to makeup the wattage?.

[Terry Bites]

Been there with cheap ebay DC-DC that caught fire as an AA stack went into retirement.
A shocker.

[shapirus]

Not precisely the same scenario, but I have damaged a few 40106's by hovering the input (not VCC) signal peak level very near the triggering threshold, just a tiny bit below it. At a sufficient repetition rate the IC begins to draw a lot of power and eventually dies. And that's while the input voltage stays well within the spec!

[radiolistener]

yes, if circuit has DC/DC converter its efficiency drops down with rise voltage delta between input and output, so lower or higher voltage leads to higher heating. Step-up DC/DC also drain more high current from source with lower voltage. As result, undervoltage leads to higher heating and higher current flow and it can damage circuit components if it don't have protection.

Also, as mentioned above, MOSFET may not fully opened with undervoltage, so it leads to higher resistance and as result higher heating. So, it can be burned out if there is no protection.

Overvoltage also can damage circuit, because it can lead to higher current and higher temperature and can lead to breakdown of capacitors and semiconductor elements, such as transistors or diodes.

Note, that both - undervoltage and overvoltage can lead to higher temperature which causes transistor and capacitors ageing much faster. So, even if circuit is still works with undervoltage or overvoltage it still leads to shortened lifetime of circuit.

Overvoltage or undervoltage also can lead to a higher current in the inductors which can leads to inductor core saturation and as result high heating which can damage or melt inductor core or degrade its magnetic properties.

[David Hess]

SMPS without proper under-voltage lockout can blow up if you spend too much time at low voltage. MOSFETs won't fully switch on and off due to insufficient gate-source voltage.

And even if the transistors switch fully, at lower voltages the input current becomes higher so current related losses are higher and parts may overheat from IR^2 losses.  Off-line switching power supplies usually have a specified minimum input voltage.

Usually switching power supplies can operate over a 2:1 input voltage range, however wide input range units commonly support a 4:1 input supply voltage range, and some may be higher.  For this to work, the switching transistors have to both support double the voltage and double the current, making them larger and more expensive, and lowering efficiency because of increased switching losses.

[liaifat85]

Yes. It is possible. Some power supplies and voltage regulators may not operate properly below a certain input voltage. This can lead to unstable output voltages and potentially affect the components connected to them.