Raspberry 3B+ undervoltage?

Hi,

I just bought a Raspberry 3B+ and now I am setting it up.

Nice little box and it's really fast in comparison to my other RPis like Zero/W, A, B... It's fun to work with it.

But, besides this, one thing makes me really wonder:

Upper right on the screen there appears a yellow flash all the time and one of the boot messages reads something like "undervoltage detected".

The new box is working fine, though, but I connected a USB voltage measuring device and saw that the power input on micro USB connector of RPi3 never goes under 5.01 Volts, but, the USB power on those 4 USB port (I only checked 2 of them with identical results) is only 4,55V (!).

Total input current is at 0.6 A in average, so I doubt that the 5V/2A power supply is too weak.

Am I doing something wrong?

Best regards,

Markus

--
Please reply to group only. 
For private email please use http://www.dipl-ing-kessler.de/email.htm
Reply to
Markus Robert Kessler
Loading thread data ...

where are you measuring the 5.01V at the PSU itself or on the PI Cable resistance can easily cause a Voltage drop & many PSU's are of poor quality & drop voltage even under light load.

--
A political man can have as his aim the realization of freedom, 
but he has no means to realize it other than through violence. 
 Click to see the full signature
Reply to
Alister

Are you actually measuring the voltage at the RPi end of the cable, or are you measring at the other end and assuming that the same voltage arrives at the RPi?

Many people have had trouble caused by USB cables that have too much resistance, and thus a significant voltage drop between the ends of the cable. The shorted and fatter the cable, the better!

Dave

Reply to
David Higton

Am Sun, 03 Feb 2019 16:53:55 +0000 schrieb David Higton:

Hi Dave, hi Alister,

I am using such a device

formatting link
dp/B013FANC9W

between power supply out and raspberry power in.

So, I see exactly what RPi gets. With the same voltmeter I then checked the (2 out of 4) USB ports on the RPi.

The funny thing is, meanwhile I found a power supply which was part of a "Raspberry 2 starter kit" from Vilros. This one delivers 5V and 2A also, but in reality, there are 5.37 Volts without load, measured with the device above. Surprisingly RPi did not complain about "over voltage".

Best regards,

Markus

--
Please reply to group only. 
For private email please use http://www.dipl-ing-kessler.de/email.htm
Reply to
Markus Robert Kessler

capacity/

But where is it in relation to the power cable?

Since it has regular sockets I'm guessing you've connected

Wall Power plug ---> supply -> Powerjive --cable to RPi--> Pi 3B in which its (correctly) showing that the PSU is giving 5.01 volts but will not see any voltage drop in the cable connecting it to the RPi.

This is the cable that people ate telling you to check or just replace.

To measure the voltage into to the RPi you need to put the meter between the cable and the RPi but, as this uses microUSB connectors which the Powerjive does not have, you'd also need USB B -> microUSB adapters on either side of the PowerJive. Since yoyu don't mention them, I assume you don't have any.

Nope - because you are ignoring voltage drop in the cable.

Thats quite normal. Many PSUs show an overvoltage when only tiny currents are being output. Even pulling 100 mA from such a PSU will reduce its output voltage to under 5.1v

--
Martin    | martin at 
Gregorie  | gregorie dot org
Reply to
Martin Gregorie

Am Sun, 03 Feb 2019 18:42:52 +0000 schrieb Martin Gregorie:

Hi,

yes, you're right, the voltmeter comes directly behind the power supply. But, the cable I use is tested to have around 0.1 .. 0.2 Ohms.

B.t.w., is there an upper limit, what RPis tolerate as input voltage?

At least one of my USB power supplies has an open circuit voltage of 5.5 V. Where's the limit, where damages may occur?

Best regards,

Markus

--
Please reply to group only. 
For private email please use http://www.dipl-ing-kessler.de/email.htm
Reply to
Markus Robert Kessler

This will never give an accurate result as it is in line with the very thing it is measuring.

The only way to do this reliably is with a proper battery operated test meter. These days you can get them for about a tenner - and once you've got one you'll wonder how you ever managed without it :P

--
W J G
Reply to
Folderol

Logic ?5v? means anything between 4.5v and 5.5v.

--
-michael - NadaNet 3.1 and AppleCrate II:  http://michaeljmahon.com
Reply to
Michael J. Mahon

That's not normally true - lots of (most?) 5V stuff requires at least

4.75V these days. There is no generally applicable upper limit - it could be as close as 5.25V, or could be higher, possibly substantially so. The higher values are expected if the "5V" goes into a regulator and nothing else. In such a case the upper limit depends on the performance or heat dissipation of the regulator.

But I repeat, what matters is what voltage arrives at the destination end of the cable, which is often substantially and surprisingly less than what was launched into it. It's no use pontificating about what it must be, bearing in mind the current consumption and the cable's resistance. Just measure it at the consuming device. Only then will you know.

And if there's an undervoltage symbol on the screen, the consuming device is already saying there's a problem.

Dave

Reply to
David Higton

Depends on which logic system. Not true for TTL. And it has nothing to do with power supply voltage.

Reply to
A. Dumas

Just believe what it says. You need either a:

- power supply with higher voltage

- better quality power supply with more constant voltage

- shorter usb power wire

- thicker usb power wire

Or all of them. Check it on the command line: "vcgencmd get_throttled". It'll reply like: "throttled=0x50005". The first 5 means it has at one point during this run been throttled due to undervoltage, the second 5 means it is currently being throttled due to undervoltage (a different reason could be overheating). So sort it out or you will get lousy performance.

bit 0: undervoltage (voltage < 4.63 V) bit 1: capped (temp > 80 C) bit 2: throttled (temp > 85 C or undervoltage)

From

formatting link
: "Throttling removes turbo mode, which reduces core voltage, and sets arm and gpu frequencies to non-turbo value. Capping just limits the arm frequency (somewhere between 600MHz and 1200MHz) to try to avoid throttling. If you are throttled and not under-voltage then you can assume over-temperature. (confirm with vcgencmd measure_temp)."

Reply to
A. Dumas

I forgot:

- higher rated power supply that isn't near its limits at a current of ~1 A.

But the most likely reason by far is the quality of the power supply, where quality means constant voltage for any current draw in the rated range. Unfortunately commodity parts like usb power supplies don't come with datasheets where you could see a response graph. It's all a bit hit & miss, unless you go for "the official power supply". Btw, I never heard of the 1 or Zero having this trouble, maybe sometimes a 2 and a very shit power adapter, but it got big with the 3.

Reply to
A. Dumas

Am Mon, 04 Feb 2019 06:16:17 +0000 schrieb A. Dumas:

Hi,

sure, you're right -- just tested also with a very small PSU just for fun since this one delivered 5.5 V open circuit. RPi3 ran fine and throttled was 0x0. But as soon as a more power consuming command like iwlist was executed, RPi was somehow "switched off"...

Well, since it is hard to find a PSU with exactly 5.2 .. 5.3 Volts to make RPi3 happy -- what about using a normal stable 7.5 Volts power supply?

Will this make the power regulator melt down?

Best regards,

Markus

--
Please reply to group only. 
For private email please use http://www.dipl-ing-kessler.de/email.htm
Reply to
Markus Robert Kessler

What about using a shorter and/or thicker cable between the PSU and RPI?

Dave

Reply to
David Higton

I've had no problem with a bog-standard USB power supply (I think it was for my Amazon Kindle), even powering a hard disc (spinning, not SSD) that was plugged into the Pi. I only bought a Pi-certified PSU

formatting link
and a powered USB hub (for the HDD) to be extra sure that the Pi wouldn't crash.

Reply to
NY

Before you do anything else, buy a voltmeter and check the voltage on the Pi's GPIO interface. Pins 2 and 4 (end nearest microSD slot, side opposite HDMI) are the 5 volt bus, ground can be found at pin

39 (diagonally opposite on GPIO).

The pins are small and easily cross-shorted with typical test probes. It's wise to de-power the Pi, make the connections and then re-connect to test voltage. If you've got old motherboard connectors lying around you can probably extract pin sockets and wires to make test pigtails.

IIUC, the 5 volt bus isn't internally regulated, but the 3.3 volt bus is derived from 5 volt by series regulation.

I believe 7.5 volts will fry the Pi 8-) if I'm wrong surely somebody will correct me.

Making it easy to verify the Pi's supply voltage will come in handy when you report problems to others. One of the first questions I get when seeking advice with a problem is "What's the supply voltage?".

HTH,

bob prohaska

Reply to
bob prohaska

There are several PSUs on Amazon that claim to be 2.5 amp and suitable for the Pi 3B at prices ranging from GBP 7.99 to GBP 11.45 with at least one claiming to be "Official 5V 2.5A Power Adapter for the Raspberry Pi

3" - this has a permanently attached 5v cable, which may be A Good Thing because that's one less set of contacts to go high resistance.
--
Martin    | martin at 
Gregorie  | gregorie dot org
Reply to
Martin Gregorie

This. This a million times.

Reply to
mm0fmf

Particularly so in this case, as the thing is teling him clearly that it is receiving undervoltage!

And, yet again, I remind the OP (and everyone else) that what matters isn't what's sent, but what arrives. The only place that makes any sense to measure the voltage is within the RPi.

Dave

Reply to
David Higton

5.25V is the official max, and 4.75V the official min. The low voltage is detected by an APX803-46 chip, which trips at 4.63V. The screen flash is latched on for 3 seconds, so it only needs quite short and occasional dips below 4.63V to keep the screen flash on.

A good guideline for max resistance of USB lead is probably

0.25ohm. I measured a lot for a presention I gave last week, and very few random USB leads are within that range.

The best commercial one I had was a 1m Poundland PowerGeeks

That was beaten by one I had made which came in at 0.12 ohm for the same 1m, and another very short one I had.

If you want to see the presentation I gave, you can find it here. The first 10 mins is the relevant part for this.

formatting link

(I'm not responsible for misspelling the URL ;-)

--
Andrew Gabriel 
[email address is not usable -- followup in the newsgroup]
Reply to
Andrew Gabriel

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.