pi3b+ needs a beefier power supply than the pi3b

[timrowledge]

I’ve had no problems so far using decent quality USB chargers like the Anker 4 or 6 port versions. They have plenty of power capacity , seem to regulate it well, have no problem running a Pi on each outlet, even look nice. And they’re not much different in price to rather plainer and less capable wall warts.

[mahjongg]

given n67 a well deserved week free from this forum, everywhere there is trouble he tries to escalate it. enough is enough....

Note that one reason the 3B+ can seem abnormally power hungry is that if you try to use older bootcode with it, it doesn't boot, and gives a "rainbow screen" in combination with the undervolt icon. This does NOT mean the power supply you are using is under-powered, it means you need to use a newer version of the boot code.

and yes, Doughy isn't always as polite to people who he thinks don't deserve it... sigh....

[rurwin]

I have tested "5V/2.4A" chargers that dipped well below 5V at only a 1A load.

Unfortunately, a power supply can be within USB specs and outside Raspberry Pi requirements. It's allowed to drop as low as 4.4V (USB2) or 4.45V (USB3) unless it is a smart USB3 Power Delivery device, which none of these things are.

That's where the Raspi power supply would have to go next -- get some intelligence into the front end and actively demand a higher power. That's not impossible but it would effectively mean that you could only use an official supply. Although the hackers would find a small number of phone chargers that were compatible with the demands -- my Samsung charger charges my phone five times faster than using my Raspi power supply.

[Imperf3kt]

I have tested "5V/2.4A" chargers that dipped well below 5V at only a 1A load.

Unfortunately, a power supply can be within USB specs and outside Raspberry Pi requirements. It's allowed to drop as low as 4.4V (USB2) or 4.45V (USB3) unless it is a smart USB3 Power Delivery device, which none of these things are.

That's where the Raspi power supply would have to go next -- get some intelligence into the front end and actively demand a higher power. That's not impossible but it would effectively mean that you could only use an official supply. Although the hackers would find a small number of phone chargers that were compatible with the demands -- my Samsung charger charges my phone five times faster than using my Raspi power supply.

Thats because your samsung charger uses the data pins of the USB to check some things, detect a Samsung phone, and bump the voltage to power a different charge circuit.

[rurwin]

Thats because your samsung charger uses the data pins of the USB to check some things, detect a Samsung phone, and bump the voltage to power a different charge circuit.

Yes, but the way to do that is defined by the USB3 specification, all the way up to delivering 100W with 5A at 20V .
https://en.wikipedia.org/wiki/USB#Power_Delivery_(PD)

[HawaiianPi]

Unfortunately, a power supply can be within USB specs and outside Raspberry Pi requirements. It's allowed to drop as low as 4.4V (USB2) or 4.45V (USB3) unless it is a smart USB3 Power Delivery device, which none of these things are.

As I recall, 4.4V is for low power USB devices. The high power USB 2.0 spec is 5V +/- 5% (4.75-5.25V), so a PSU rated for 2.4A that dips that low under load is out of spec.

4.75 to 5.25V is also the recommended range for our little Raspberry Pi computers, although I recommend a PSU that tilts to the high side (5.1-5.25V) because there is some voltage drop through the Pi, so 5V in at the micro-USB power port means less than 5V at the USB peripheral ports.

That's where the Raspi power supply would have to go next -- get some intelligence into the front end and actively demand a higher power. That's not impossible but it would effectively mean that you could only use an official supply.

Only true if RPF created their own proprietary power spec. If they used an established USB power protocol then most "smart" USB chargers should work. Then again, the "next" Pi computers might be more power efficient and need less power (I think current models are 40nm, so there is a lot of room for improvement).

[tkaiser]

Only true if RPF created their own proprietary power spec.

If the SoC of next gen Pi can deal with USB-C even without the need to implement more expensive USB PD (switching voltages) then this should be sufficient since USB-C defines two primitive 5V power modes with 1.5A and 3A. At that time dedicated USB-C chargers limited to 5V/3A and not implementing the USB PD voltage modes should be cheap enough too...

[royhenderson]

I have been using BitScope Pi-Quattro boards to try and reduce the whole mess of individually powered boards ... I presently feed each BitScope board from a 24V 4A PSU and that comfortably runs 4 x RPi3B 24x7. Have not yet tried the 3B+ but hopefully it will take a block of 4 with no issues.

I have also bought a couple of Mean Well PSUs - one is 5V 26A and the other is 12V 12.5A. The 5V unit is to directly power a block of 10 RPi and the 12V unit is to power a block of 3 Quattro boards.

The reason for the farm is that my current project is to build a small hosting configuration purely as a technical exercise, not as a commercial venture, but I want reliable 24x365 power.

[jbudd]

Edit - oops, should have read page 2 before I posted this!

I think some devices - eg iphones? have a way of telling some "Smart chargers" that they can accept high charge current and / or voltage? Qualcomm Quick Charge, etc(?)

I have a "Smart" USB power pack 2.1A / 2.4A that could fast charge a recent iphone but it supplies < 1A to a Raspberry Pi.
Is it possible for a current or future Pi to negotiate with these smart chargers to get a high current?

Would it be fatal to the Pi if more than 5V was supplied to the USB power in socket?

[davidcoton]

Would it be fatal to the Pi if more than 5V was supplied to the USB power in socket?

The spec limit is 5.25V.
At 6V, the protection diode will conduct, and the polyfuse will trip. Should not be fatal but could put your Pi in hospital overnight.
At a high enough voltage there could be permanent damage before the protection kicks in -- usually the protection diode fails short circuit, so the Pi is still protected. Tthe wise guys resusitate the Pi by amputating the protection diode. A repeat of the overvoltage while the protection diode is absent will be fatal. So should you have to remove a failed protection diode, rest your Pi until a new one is installed.

[W. H. Heydt]

Edit - oops, should have read page 2 before I posted this!

I think some devices - eg iphones? have a way of telling some "Smart chargers" that they can accept high charge current and / or voltage? Qualcomm Quick Charge, etc(?)

I have a "Smart" USB power pack 2.1A / 2.4A that could fast charge a recent iphone but it supplies < 1A to a Raspberry Pi.
Is it possible for a current or future Pi to negotiate with these smart chargers to get a high current?

Would it be fatal to the Pi if more than 5V was supplied to the USB power in socket?

I have a 10.4Ah powerbank that has one 2.4A output and one "QC 2.0" output. I connect a Pi2Bv1.1 to the 2.4A port and an RPF 7" display to the QC port. It all runs just fine...for up to about 5 hours at a time.

[marcosbong]

Hello.
Anybody here knows why I have an official Raspberry pi power supply made by Stontronics, but rated 2A ?
Doesn't look like fake.
I got it second hand, in a blanked cardboard box and several plug adapters.

[davidcoton]

The Stontronics supply was changed from 2A to2A5 about the time the Pi3B was launched, IIRC. 2A was enough for previous models but 2A5 is recommended for the Pi3B and Pi3B+ to allow the full 1A2 for USB devices.

[W. H. Heydt]

The Stontronics supply was changed from 2A to2A5 about the time the Pi3B was launched, IIRC. 2A was enough for previous models but 2A5 is recommended for the Pi3B and Pi3B+ to allow the full 1A2 for USB devices.

Yup. Pi2B is spec'd for a 2A PSU. IIRC, the polyfuse on a Pi2B is actually 1.8A.

[marcosbong]

Thanks guys.
I'm fighting to upload images, but I give up for this time.
I used the PSU with a Rpi2, now I'm using this PSU with a Rpi3+ I've just bought
Do you recommend replacing it, or give it a try?
I've using it for a few hours without any issue, but I only have plugged a USB keybrd an a mouse

[Imperf3kt]

It should work fine unless you use high power devices like a hard drive and even then I suspect it should cope under most circumstances.

[marcosbong]

Thanks man.
If I need to plug a HDD I will use a powered Usb hub I got some time ago for my first Rpi that was trusted with no gnd loops issues with Rpi

[HawaiianPi]

If I need to plug a HDD I will use a powered Usb hub I got some time ago for my first Rpi that was trusted with no gnd loops issues with Rpi

The recommended 2.5A includes 1.2A for the USB ports, which leaves 1.3A for the Pi computer itself. So your 2A PSU should have enough power if you have no more than 700mA of load on the USB ports (which should be enough for a keyboard and mouse, with a little left over).

[marcosbong]

If I need to plug a HDD I will use a powered Usb hub I got some time ago for my first Rpi that was trusted with no gnd loops issues with Rpi

The recommended 2.5A includes 1.2A for the USB ports, which leaves 1.3A for the Pi computer itself. So your 2A PSU should have enough power if you have no more than 700mA of load on the USB ports (which should be enough for a keyboard and mouse, with a little left over).

Thanks, dude

[386sx]

This was mentioned in the launch video, but struggling with current 2A random adapters which have always worked for me with pis in the past.

Official raspberry pi adapter is a lot of dollars shipped to Canada. On the lookout for something local... any ideas? 2.5A seems rarer than 2A.

From my experience with different power supplies, I end up buying the official cause every one I had seems to not be really stable not necessary on the current itself but on the voltage. I tried a 3A-5v not official but well built power supply and still even if wattage wasn't that high at the plug (5-7W) sometimes the low-voltage indicator showed on-screen not to mention with other supplies I had.
I suppose/imagine that many generic phone oriented power supply go lower than 4,8v or similar with an heavy and variable power load beside how much current they could give? I had also a 5 USB ports power supply charger that have 2 USB ports @ 2.5A and 3 USB ports @ 1A. Even with that and the Y double cable for a total of 5A, the voltage still went lower than needed.

[marcosbong]

I agree with 386sx.
The problem is that most of the market available products are power adaptors for charging batteries.
When you charge a battery, you dont really mind if you're feeding it with 5v or 4.8v. You just need a wattage flowing to the battery, so if volts goes down a little, the adapter just rise amps a little more and that compensates the total power.
But Raspberry needs an stable Vcc supply, so most of the adaptors in the market won't fit its requirements.
They're wrongly called PSU, when they must be called adaptor.
Most manufactures just want to grab a wider sector of the market without any effort, neither money on it, just saying " PSU for Raspberry Pi "

[jamesh]

Worth noting the Official PS is actually 5.1v at the wall end, and uses thick copper wires, so by the time the power gets to the Pi itself, its dropped only slightly to 5v.

[bensimmo]

Also to add, Amazon and Google use 5.1/5.2V wired PSU's for their devices too, most likely the same reason Raspberry do.

(That's Home mini's and FireTV's etc)

[FTrevorGowen]

The Stontronics supply was changed from 2A to2A5 about the time the Pi3B was launched, IIRC. 2A was enough for previous models but 2A5 is recommended for the Pi3B and Pi3B+ to allow the full 1A2 for USB devices.

FWIW, having finally acquired the newer version, and had the time to update my webpages, here are my (updated) measurements of the (static) current loading of both versions of the (Stontronics) PSU:
http://www.cpmspectrepi.uk/raspberry_pi ... _Models.29
Trev.

[386sx]

Worth noting the Official PS is actually 5.1v at the wall end, and uses thick copper wires, so by the time the power gets to the Pi itself, its dropped only slightly to 5v.

And also it should have only the two big power wires and no other communication wires.