Raspberry Pi and Powerbank

[zLagos]

Hi, I have a raspberry pi 3B + with a DAC, a fan and a 7 "touch screen connected to run the Volumio program.

When I connect a powerbank with a 5V and 2.4A output, I see the yellow ray and the "undervoltage detected" message, but the Volumio is still working perfectly, the music is heard well and there are no jumps or anything.

And the screen looks perfectly and responds well to touch.

My question is: can having the powerbank connected to the pi cause any physical damage? Can "undervoltage detected" damage the pi?

Thanks for the help.

[davidcoton]

I suspect you are overloading the powerbank. That won't damage the Pi, but will at least shorten the life of the powerbank. It is also possible (no idea of the probability) that the powerbank will fail in a way that then damages the Pi.

[W. H. Heydt]

You're using a powerbank with a maximum specified output of 2.4A to feed a system with a specified current requirement of 3.0A. Not surprising you're seeing the low voltage indicator.

Either of these https://www.monoprice.com/search/index? ... ame=DM7826 should meet the power requirement....but they aren't cheap.

[zLagos]

You're using a powerbank with a maximum specified output of 2.4A to feed a system with a specified current requirement of 3.0A. Not surprising you're seeing the low voltage indicator.

I knew that, that's why I asked if the raspberry pi could be damaged.

It also has a USB C output. Could you use it with a miniUSB-USB C adapter or will it also give me the problem of undervoltage detected?

[W. H. Heydt]

You're using a powerbank with a maximum specified output of 2.4A to feed a system with a specified current requirement of 3.0A. Not surprising you're seeing the low voltage indicator.

I knew that, that's why I asked if the raspberry pi could be damaged.

Undervolatage won't damage the *Pi*, but it may case SD card corruption. It may also cause erratic or operational failure of USB devices, further, it may damage your powerbank by attempting draw more power than the voltage converters can deliver--with the resultant risk of catastrophic fire.

It also has a USB C output. Could you use it with a miniUSB-USB C adapter or will it also give me the problem of undervoltage detected?

Doesn't matter what kind of connector the powerbank has. The Pi will try to draw the current it needs to operate plus whatever attached devices need. If the powerbank circuits can't deliver the current, they may overheat and fail. You need to determine what your *actual* current requirements are and then verify that the power source can supply--and supply in a sustained way, not just as a temporary surge--sufficient current and still maintain proper voltage. Because you are seeing undervoltage indicators (the "lightning bolt") this is clearly not the case with your current set up.

Powerbanks are really intended to recharge other batteries. That process isn't particularly sensitive to voltage, so long as the powerbank can supply current at or above the charging voltage of the target system, all is good. A Pi, by contrast wants no only enough current to power itself and any attached devices, but it wants the input voltage to stay between 4.75v and 5.25v. If the input voltage drops to 4.65v or below, you get the low voltage warning.

The best way to avoid these problems is to use a good quality power source that can be trusted to meet or exceed the specifications of the Pi. If you have to use power sources that you can't completely trust to supply their specified current and maintain proper voltage, your only recourse is to use ones that exceed the specification by a reasonably margin. Hence the powerbanks I linked previously.

I would not attempt to power a Pi4B with a powerbank that didn't have specs of at least 3A at 5v output, even with no additional devices attached. If I had to use a battery like the one you cited, I would drop down to a Pi3B, or even a Pi2B, to provide an operating margin.

[zLagos]

Could this be enough?

https://www.amazon.es/dp/B07L6WTYQK/?co ... _lig_dp_it

[zLagos]

As the powerbank has two outputs, I have connected one to the raspberry pi and another to the screen ... et voilá ... problem solved. The yellow ray has disappeared.

Gracias a todos por todo.

[Imperf3kt]

As the powerbank has two outputs, I have connected one to the raspberry pi and another to the screen ... et voilá ... problem solved. The yellow ray has disappeared.

Gracias a todos por todo.

Make sure the 5v between the screen and Pi is not connected (but ground must be)
in the event that powerbank has two regulators, they will end up "fighting" as neither will output exactly the same voltage
as the other. this may shorten the life of the device.

[Zwartoog]

I know this is an old thread, but: the Pi registering undervoltage means in my experience nothing more than that the supply voltage is less than its reference.

I experimented previously with a Pi4B, put ~5V adjustable power supply to the +5V rail directly. The "under voltage" message still appears with 5.00V (according to my voltage meter), turning it up a bit to 5.05V or so removes the message. Not sure where the 4.65V fits in.

So, my guess is that your power bank supplies a little bit less than the 5V reference of the Pi. Nothing to be worried about, it's just annoying.

[Paul Hutch]

I experimented previously with a Pi4B, put ~5V adjustable power supply to the +5V rail directly. The "under voltage" message still appears with 5.00V (according to my voltage meter), turning it up a bit to 5.05V or so removes the message. Not sure where the 4.65V fits in.

IIRC (it's been a long while since I saw the info and verified it on my test bench), the 4.65V limit is what the official RPi engineers said is the limit that triggers the warning icon.

Since you where getting a warning icon at a much higher voltage either there is problem with your Pi or your voltage measurement. The most common problem with the voltage measurement would be making the measurement at the power supply terminals instead of the 5V rail on the Pi's PCB. At 2A current draw from a Pi and only one milliohm of resistance in the power wiring (cable & connectors) you'll get 2 millivolts of voltage drop (E=IR). So 0.1 ohm of cable & connectors resistance, a very common situation, @ 2A draw yields a 0.2V drop. The official power supplies have a 5.1V nominal output to help this situation.

So, my guess is that your power bank supplies a little bit less than the 5V reference of the Pi. Nothing to be worried about, it's just annoying.

Nothing to worry about as long as you don't mind the potential operational problems including a corrupted uSD card. The engineers who design the Pi have very sound reasons for the low voltage warning.

[Zwartoog]

I experimented previously with a Pi4B, put ~5V adjustable power supply to the +5V rail directly. The "under voltage" message still appears with 5.00V (according to my voltage meter), turning it up a bit to 5.05V or so removes the message. Not sure where the 4.65V fits in.

IIRC (it's been a long while since I saw the info and verified it on my test bench), the 4.65V limit is what the official RPi engineers said is the limit that triggers the warning icon.

Since you where getting a warning icon at a much higher voltage either there is problem with your Pi or your voltage measurement.

Then I guess my Pi is faulty. I just did the measurements again to verify. When supplied with 4.96V on the USB-C-input, the thunderbolt appears:

piundervoltage.jpg (126.54 KiB) Viewed 1093 times

I previously had a similar observation with a fixed regulator which measured 4.99V. I can turn the voltage down all the way to at least 4.55V without the Pi going down (did not try further).

Another perspective appears when I measure at the +5V rail directly. There is a 0.20V-drop compared to the input to the USB-C and what is delivered to the +5V rail.

So, my guess is that your power bank supplies a little bit less than the 5V reference of the Pi. Nothing to be worried about, it's just annoying.

Nothing to worry about as long as you don't mind the potential operational problems including a corrupted uSD card. The engineers who design the Pi have very sound reasons for the low voltage warning.

Only if there is a real problem with the supply, rather than reporting a bolt. If there is a slight undervoltage, but not because the power bank running dry, I do not see how this risks corrupted uSD cards.

[emma1997]

Far more likely those cheapie powerbanks and LM2596 modules are not putting out stable enough voltage. Something which can only be checked with a scope & shunt. Even the best meter can deceive. In fact attaching a meter of any kind will make things worse due to 'burden voltage'.

On at least 2 occasions I traced SD corruption to those flaky converters.

PS I found that model usually worked for Zero but about 5% failed with Pi3 and only half the ones tried worked long term for Pi4. Some users claim success but I suspect they either got lucky or received selected units and paid a lot more. All the $1 XL4015 type with toroid works with all my Pi and over the long term too. YMMV

[Paul Hutch]

Then I guess my Pi is faulty. I just did the measurements again to verify. When supplied with 4.96V on the USB-C-input, the thunderbolt appears:
piundervoltage.jpg
I previously had a similar observation with a fixed regulator which measured 4.99V. I can turn the voltage down all the way to at least 4.55V without the Pi going down (did not try further).

Another perspective appears when I measure at the +5V rail directly. There is a 0.20V-drop compared to the input to the USB-C and what is delivered to the +5V rail.

The only measurement that is useful is the one with the multimeter on the 5V/GND GPIO pins. According to what you've written you are getting a lightning bolt with 4.76V (4.96V - 0.20V). Now subtract the measurement accuracy of your meter, you'll need to look it up. A rough rule of thumb for DMM's that cost < $150.00US is +/-2 to +/-5%. Calculating shows that the actual values worst case minimums are -2% = 4.66V, -5% = 4.52V.

Running that close to the minimum can lead to uSD card corruption which is why there is the warning (not will but can, most specs in electronics are to guarantee proper operation, not guarantee failure). If you choose to ignore the warning that's fine, it's your data.

[Zwartoog]

Far more likely those cheapie powerbanks and LM2596 modules are not putting out stable enough voltage. Something which can only be checked with a scope & shunt. Even the best meter can deceive. In fact attaching a meter of any kind will make things worse due to 'burden voltage'.

On at least 2 occasions I traced SD corruption to those flaky converters.

PS I found that model usually worked for Zero but about 5% failed with Pi3 and only half the ones tried worked long term for Pi4. Some users claim success but I suspect they either got lucky or received selected units and paid a lot more. All the $1 XL4015 type with toroid works with all my Pi and over the long term too. YMMV

Thanks for the warning! The 4.99V I measured previously was with a custom build LM2576T-5V version, with a toroidal inductor and caps, each near the size of the prefab buck converter. Also displayed a bolt. Nevertheless, my aim is now to use a LM2596-ADJ followed by an LDO.

The only measurement that is useful is the one with the multimeter on the 5V/GND GPIO pins. According to what you've written you are getting a lightning bolt with 4.76V (4.96V - 0.20V). Now subtract the measurement accuracy of your meter, you'll need to look it up. A rough rule of thumb for DMM's that cost < $150.00US is +/-2 to +/-5%. Calculating shows that the actual values worst case minimums are -2% = 4.66V, -5% = 4.52V.

Ouch, that's a terrible accuracy. My DMM is a fairly old one, but its readings are consistent. Anyway, I have a LM4040 lying around to build a custom voltage meter. Would be fun to compare.

Thanks for the reflection!

[davidcoton]

The only measurement that is useful is the one with the multimeter on the 5V/GND GPIO pins. According to what you've written you are getting a lightning bolt with 4.76V (4.96V - 0.20V). Now subtract the measurement accuracy of your meter, you'll need to look it up. A rough rule of thumb for DMM's that cost < $150.00US is +/-2 to +/-5%. Calculating shows that the actual values worst case minimums are -2% = 4.66V, -5% = 4.52V.

Ouch, that's a terrible accuracy. My DMM is a fairly old one, but its readings are consistent.

Don't confuse consistency with accuracy. Many measuring systems give results that are consistent but inaccurate (for some interpretation of accurate). Accuracy requires calibration -- that is, comparison with a known standard reference. Maintaining accurate standards is a whole new world of pain.

[drgeoff]

Far more likely those cheapie powerbanks and LM2596 modules are not putting out stable enough voltage. Something which can only be checked with a scope & shunt. Even the best meter can deceive. In fact attaching a meter of any kind will make things worse due to 'burden voltage'.

A meter's burden voltage needs to be considered when it is measuring current. It has no significance when measuring voltage.

[emma1997]

Actually I was referring to the USB monitor in Zwartoog's setup which does measure current. The extra resistance inline often drops enough voltage to prevent operation as will many DMM. I can't count the number of times couldn't figure out what was failing until removing the ammeter got it going again.

Sort of like how half the problems caused by too thin wire or lossy connectors.