Low voltage warning with 5.18 V / 10 A PSU

[FalCT60]

Hello,
I am facing this nuisance with a DIY PSU.
The source is a 14.8 V / 10 A PSU + adjustable 40~3 V / 30 A buck converter set to 5.18 V + PD USB-C cable taken from a 90 W USB-C PD PSU.
RPI5 8GB + Pimoroni NVME board + Samsung 970 EVO SSD.
I (randomly) get this warning after boot though the load is not that high. Restarting the PI and the warning disappears, even if I add some extra USB device.
In my testings, I could power a 5 V / 6 A motor for more than 30 minutes with no noticeable heating.
My question : is the warning due to the fact that the CC wire is not connected ? Would adding the 5.1k pull-down resistor between the CC wire and ground solve this ? I could not understand if it is used for the PSU to know what it has to deliver to the device, or if it makes the device know the PSU is good for it. I'm a little confused. I couldn't understand how it actually works.
Rgrds,

[FTrevorGowen]

Hello,
I am facing this nuisance with a DIY PSU.
The source is a 14.8 V / 10 A PSU + adjustable 40~3 V / 30 A buck converter set to 5.18 V + PD USB-C cable taken from a 90 W USB-C PD PSU.
RPI5 8GB + Pimoroni NVME board + Samsung 970 EVO SSD.
I (randomly) get this warning after boot though the load is not that high. Restarting the PI and the warning disappears, even if I add some extra USB device.
In my testings, I could power a 5 V / 6 A motor for more than 30 minutes with no noticeable heating.
My question : is the warning due to the fact that the CC wire is not connected ? Would adding the 5.1k pull-down resistor between the CC wire and ground solve this ? I could not understand if it is used for the PSU to know what it has to deliver to the device, or if it makes the device know the PSU is good for it. I'm a little confused. I couldn't understand how it actually works.
Rgrds,

In your "testings" did you consider these factors:
1) How long is your "PD USB-C cable taken from a 90 W USB-C PD PSU" cable? (You may find these tests/measurements of interest:
https://www.cpmspectrepi.uk/raspberry_p ... plies.html )
2) The low-voltage warning can be triggered by short-duration voltage droops that may be observable with an oscilloscope but not via a multimeter.
Trev.

[pidd]

What current is the buck converter? If its 5A or more, have you told the Pi5 to override the 3A non-PD limit.

Your NVME makes load high in terms of a small SBC,.

[FalCT60]

Hello,
@FTrevorGowen : I have shortened the cable to ~60 cm - half the length of the official PI PSU.
I own no oscilloscope, unfortunately.
@kip_the_elder & pidd : this buck converter is to be capable of 30A, so let's assume 15 is a pessimistic actual value.
@pidd : whether the limmit is set or unset makes no difference.
I forgot to mention : I have used a USB multimeter to monitor the PI's consumption, and the power never reached more than 1.5 A, with voltage from 5.11 and 5.14 V. Which mlakes me feel really puzzled.

But all this s till doesn't answear my question : will a 5.1k resistor between CC wire and GND tell the PI it is powered by a PD capable PSU and let it draw all the power it needs, or will it be useless in my case ?
I don't want to dismantle the official PI's PSU to see the way it is built. Not even sure it would help me.

[bensimmo]

Hello,
@FTrevorGowen : I have shortened the cable to ~60 cm - half the length of the official PI PSU.
I own no oscilloscope, unfortunately.
@kip_the_elder & pidd : this buck converter is to be capable of 30A, so let's assume 15 is a pessimistic actual value.
@pidd : whether the limmit is set or unset makes no difference.
I forgot to mention : I have used a USB multimeter to monitor the PI's consumption, and the power never reached more than 1.5 A, with voltage from 5.11 and 5.14 V. Which mlakes me feel really puzzled.

But all this s till doesn't answear my question : will a 5.1k resistor between CC wire and GND tell the PI it is powered by a PD capable PSU and let it draw all the power it needs, or will it be useless in my case ?
I don't want to dismantle the official PI's PSU to see the way it is built. Not even sure it would help me.

Useless, you just put the line mentioned in the documentation to bypass PD checks.
As long as your PSU 'setup' doesn't use PD to deliver more than 3A...


The Pi5 can tell you the Voltage it is seeing itself (see documentation), though it cannot show you the current.

[pidd]

Are you sure the "PD USB-C cable" actually contains the PD source electronics? If so, did the"90 W USB-C PD PSU" have a 5V@5A mode, if not then it is pointless going down the PD route. Ditch the PD stuff altogether, use a plain cable of suitable thickness and apply the config override.

As far as the current you measured, if that was with a meter then it is an average current reading and will not indicate short surge currents such as when your NVME does an erase or write cycle.

[MikeDB]

What current is the buck converter? If its 5A or more, have you told the Pi5 to override the 3A non-PD limit.

Your NVME makes load high in terms of a small SBC,.

I've often found that the claimed output of some of these DC/DC converters can be a little optimistic too. I always go for devices that are supposedly much more powerful than I need.

I find that no matter how much you overrate them, quite a few of them always blow up after some time

[drgeoff]

But all this s till doesn't answear my question : will a 5.1k resistor between CC wire and GND tell the PI it is powered by a PD capable PSU and let it draw all the power it needs, or will it be useless in my case ?

USB-PD protocol is much more complex than a resistor or two on CC wire(s).

[MikeDB]

But all this s till doesn't answear my question : will a 5.1k resistor between CC wire and GND tell the PI it is powered by a PD capable PSU and let it draw all the power it needs, or will it be useless in my case ?

USB-PD protocol is much more complex than a resistor or two on CC wire(s).

Actually it wasn't bensimmo who said that.

But I agree it's more complex. To be standards compliant the pair of resistors only guarantees the PSU will deliver 500mA, but most do deliver an ill-definied lot more.

[cleverca22]

But all this s till doesn't answear my question : will a 5.1k resistor between CC wire and GND tell the PI it is powered by a PD capable PSU and let it draw all the power it needs, or will it be useless in my case ?

USB-PD protocol is much more complex than a resistor or two on CC wire(s).

also, a resistor to ground, is a sign that you are a sink, not a source
the source has different resistor to 5v

but as drgeoff said, that cant signify 5A, you need the active protocol for that

[stevend]

I find that no matter how much you overrate them, quite a few of them always blow up after some time

I haven't experienced that. I've found them to be reliable. Maybe your environment is more challenging than mine.

Seems to depend on manufacturer for us - we have a board using a 5watt dc-dc converter where one manufacturer's parts start to fail after about 7 years. Most go on for ever (25 years is about the record!). Maybe it was a "bad" batch - we have two on the board, and its always the 5V output one that fails, despite being less heavily loaded than the 3.3V output part.

[FalCT60]

Thanks to everyone who replied there.
I hoped it would be as simple as adding a resistor, but since it isn't the case I'll just ignore the warning.
As I wrote, the total load never exceeds 1.3 A and the source can deliver way much more.
Thanks again.

[bjtheone]

Thanks to everyone who replied there.
I hoped it would be as simple as adding a resistor, but since it isn't the case I'll just ignore the warning.
As I wrote, the total load never exceeds 1.3 A and the source can deliver way much more.
Thanks again.

Note you are striving to deliver 5.1V at the Pi, not at the source. The PD cables will do nothing for you. Go with a nice thick/short cable between the supply and the Pi.

[namlow]

I had been using and working on this low voltage warning for one and half year and still using it occasionally as a backup or a trivial
power supply for RPi 4B4GB.

When the system is overloaded, some intermittent problems may occur.