ThePunisher1024
Posts: 11
Joined: Sat Apr 14, 2012 11:00 am

Re: Raspberry Pi Power Requirements

Sat Apr 21, 2012 5:26 pm

As much as we have been instructed on the the power requirements being 5v 700ma, I am curious as to exactly what context this power requirement is taken.

Is this the maximum power the Pi will draw assuming USB ports drawing full power, ethernet being used, and the RPI operating at 100% of its capacity? I know we are meant to used a powered hub for devices requiring additional power but obviously some components will work well being powered by the Pi.

My question really is will a 5v 700ma power supply be enough to cater for the above situation? Will there be a situation in which the Pi will need to draw more than 700ma of current?

User avatar
abishur
Posts: 4477
Joined: Thu Jul 28, 2011 4:10 am
Location: USA
Contact: Website

Re: Raspberry Pi Power Requirements

Sat Apr 21, 2012 5:41 pm

If you're not using a powered hub and have two usb devices using the full 100 and some change mA available to it then I've heard reports of it going as high as 800 mA at full tilt, but those are some fringe data points
Dear forum: Play nice ;-)

ThePunisher1024
Posts: 11
Joined: Sat Apr 14, 2012 11:00 am

Re: Raspberry Pi Power Requirements

Sat Apr 21, 2012 6:39 pm

I am asking these questions just to determine if getting a power supply with more than 700ma would be reasonable, the 700ma official requirement left it somewhat ambiguous as to where exactly the amps would be going.

So we can assume 700ma basic for the Pi running full whack with two basic USB devices (keyboard and mouse), the USB ports however can give out up to 100ma (peak power for charging, so not recommended to be used continuously? use powered hub instead) of power each.

Although the hard limit of this is that if the Pi draws anything above 1A of current it will automatically shut down.

User avatar
jbeale
Posts: 3641
Joined: Tue Nov 22, 2011 11:51 pm
Contact: Website

Re: Raspberry Pi Power Requirements

Sat Apr 21, 2012 6:57 pm

If you're worried about current draw requirements, I'd recommend a 1A rated charger. That is sure to be enough, and I see them on Amazon for about US$7.  It's only a few dollars more than the 700 mA rated ones.  search for "micro USB charger 1A"

By the way, don't be tempted to go ultra-cheap, we know that many low price ebay parts are not suitable, due to high resistance cables causing a voltage drop under load. They get away with it, because cell phones can still charge with them (just more slowly) but the RasPi will just quit if the voltage drops too far under load.

ThePunisher1024
Posts: 11
Joined: Sat Apr 14, 2012 11:00 am

Re: Raspberry Pi Power Requirements

Sat Apr 21, 2012 7:21 pm

jbeale said:


If you're worried about current draw requirements, I'd recommend a 1A rated charger. That is sure to be enough, and I see them on Amazon for about US$7.  It's only a few dollars more than the 700 mA rated ones.  search for "micro USB charger 1A"

By the way, don't be tempted to go ultra-cheap, we know that many low price ebay parts are not suitable, due to high resistance cables causing a voltage drop under load. They get away with it, because cell phones can still charge with them (just more slowly) but the RasPi will just quit if the voltage drops too far under load.


I was looking on Farnell for a good power supply, although I could only find 2 and they were rated at 500ma, know any other reputable sources as I dont really trust amazon with these specific requirements.

selsinork
Posts: 151
Joined: Mon Apr 16, 2012 8:31 am

Re: Raspberry Pi Power Requirements

Sat Apr 21, 2012 7:40 pm

ThePunisher1024 said:

I was looking on Farnell for a good power supply, although I could only find 2 and they were rated at 500ma, know any other reputable sources as I dont really trust amazon with these specific requirements.

How about  http://uk.farnell.com/_/st-099.....dp/2081769

Can't say if it's good or not though.. "Country of Origin: CN China" so it may be exactly the same as the ones on ebay for half the price...

bredman
Posts: 1415
Joined: Tue Jan 17, 2012 2:38 pm

Re: Raspberry Pi Power Requirements

Sat Apr 21, 2012 8:43 pm

If you look at the troubleshooting forum, about half of the problems are caused by inadequate power supplies.

The recommendation of 700mA does not seems to be enough. It looks like 1000mA (1A) is necessary to be guaranteed problem-free.

Also, a good power supply can't help you if you use a rubbish power cable for the RPi.

hippy
Posts: 7360
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: Raspberry Pi Power Requirements

Sat Apr 21, 2012 9:40 pm

ThePunisher1024 said:


So we can assume 700ma basic for the Pi running full whack with two basic USB devices (keyboard and mouse)


No. From the Wiki …


Board A: 5V, 500 mA (2.5W) without any devices connected
Board B: 5V, 700 mA (3.5W) without any devices connected

http://elinux.org/RPi_Hardware

So, for a Model B, you need a charger capable of supplying 700mA plus whatever any connected devices use.

If you have two USB devices connected, each drawing 100mA, you need a power supply capable of providing 900mA.

This could explain why some Kindle power supplies seem to be giving problems for some people, being only rated for 850mA.

ThePunisher1024
Posts: 11
Joined: Sat Apr 14, 2012 11:00 am

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 12:30 am

So the 700ma requirement is slightly inadequate, given any USB device will likely draw more power than could be supplied? I know we are supposed to use powered hubs for power hungry devices, but surely using a 700ma supply will leave the Pi out of pocket power wise regardless of what USB device is plugged in?

User avatar
n3tw0rk5
Posts: 216
Joined: Fri Jan 06, 2012 6:04 pm

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 1:35 am

Mine seems to have bombed out for the 2nd time in a day, I'm using a 700ma charger but it's a cheap cable.

Also I forgot to unplug the keyboard before I left for work which wont have helped.

I'm swapping both out and will see how it goes.

ThePunisher1024
Posts: 11
Joined: Sat Apr 14, 2012 11:00 am

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 11:42 am

Think I have found a reliable power supply that supplys 5v 1000ma:

Technical specs:

http://www.expansys.com/htc-de.....le-203734/

Place to purchase from:

http://www.amazon.co.uk/HTC-TC.....038;sr=1-6

Michael
Posts: 340
Joined: Sat Jul 30, 2011 6:05 pm

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 12:09 pm

hippy said:


No. From the Wiki …


Board A: 5V, 500 mA (2.5W) without any devices connected
Board B: 5V, 700 mA (3.5W) without any devices connected

http://elinux.org/RPi_Hardware

So, for a Model B, you need a charger capable of supplying 700mA plus whatever any connected devices use.

If you have two USB devices connected, each drawing 100mA, you need a power supply capable of providing 900mA.

This could explain why some Kindle power supplies seem to be giving problems for some people, being only rated for 850mA.


That is correct – 700 mA is a minimum requirement, not maximum.

Recently, Gert pasted this in a blog comment, which I think sums up things well:


AMPS are not like volts. AMPS says how much you CAN draw from a supply. It mean you can safely connect a 100A supply to a 0.01A device. (Compare it to liters/pints. If you need 1 liter/pint of water you don’t care that you have a lake available. As long as there is enough).
If a supply [doesn't have] enough amps your Raspberry-Pi will [not] work [reliably]. We have seen that the first thing that starts failing is the Ethernet, next is the keyboard.


hippy
Posts: 7360
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 12:50 pm

ThePunisher1024 said:


So the 700ma requirement is slightly inadequate, given any USB device will likely draw more power than could be supplied? I know we are supposed to use powered hubs for power hungry devices, but surely using a 700ma supply will leave the Pi out of pocket power wise regardless of what USB device is plugged in?


That is correct. If you are using a 700mA supply, the R-Pi is drawing 700mA, no matter what any USB device draws you are over-stretching the power supply and everything could collapse in a heap at any time.

Reality is likely to be that the R-Pi isn't drawing it's maximum of 700mA so there's some spare capacity for the USB devices so everything seems fine. But if the R-Pi does then draw more power ( you run something that's GPU intensive, or write to SD Card, etc ) and the combined draw rises above 700mA it can collapse into a heap at that point.

This IMO seems to be what's happening with those people who may have borderline power supplies.

User avatar
Montala
Posts: 675
Joined: Mon Mar 05, 2012 11:54 pm
Location: Herefordshire (U.K.)

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 1:06 pm

ThePunisher1024 said:


Think I have found a reliable power supply that supplys 5v 1000ma:

Technical specs:

http://www.expansys.com/htc-de.....le-203734/

Place to purchase from:

http://www.amazon.co.uk/HTC-TC.....038;sr=1-6


There is certainly no need to pay that much for one.

Another cheaper product, also available from Amazon at http://www.amazon.co.uk/gp/pro.....03_s00_i00 also has a 1 amp (1000 ma) output, is a genuine Nokia product, and only (currently) costs £3.38.

User avatar
jojopi
Posts: 3226
Joined: Tue Oct 11, 2011 8:38 pm

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 1:17 pm

Despite the fact that the schematic says 1.1A, F3 on both beta and production photos is a 750mA (hold current) device.  This protects the entire board, including USB outputs.  So it cannot be true that the 700mA figure for the Model B is intended to exclude its peripherals?

ThePunisher1024
Posts: 11
Joined: Sat Apr 14, 2012 11:00 am

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 1:35 pm

Montala said:


ThePunisher1024 said:


Think I have found a reliable power supply that supplys 5v 1000ma:

Technical specs:

http://www.expansys.com/htc-de.....le-203734/

Place to purchase from:

http://www.amazon.co.uk/HTC-TC.....038;sr=1-6


There is certainly no need to pay that much for one.

Another cheaper product, also available from Amazon at http://www.amazon.co.uk/gp/pro.....03_s00_i00 also has a 1 amp (1000 ma) output, is a genuine Nokia product, and only (currently) costs £3.38.


Thats actually quite good, we should get a list of approved power supplies going I think? Looking on the nokia website, it gives out up to 1200ma of current, morethan enough for the Pi (which trips out at 1000ma anyway?).

If we have a 1200ma power supply, the Pi should never be able to requuest more than say 800ma? So wont have to trip its breaker?

hippy
Posts: 7360
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 2:07 pm

jojopi said:


Despite the fact that the schematic says 1.1A, F3 on both beta and production photos is a 750mA (hold current) device.  This protects the entire board, including USB outputs.  So it cannot be true that the 700mA figure for the Model B is intended to exclude its peripherals?


Or maybe the wrong fuse was fitted during manufacturing ?

It would be nice to have a final, definitive and unambiguous statement on what the R-Pi maximum consumption is without USB devices drawing power. And also get the documented power consumption figures of the R-Pi board to reflect those.

It's been asked for often enough and I thought we'd got there but perhaps not.

User avatar
Montala
Posts: 675
Joined: Mon Mar 05, 2012 11:54 pm
Location: Herefordshire (U.K.)

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 2:15 pm

My apologies... I was going by memory (never a good idea at my age!)

I have just checked the rating on mine, which appears to be a genuine Nokia AC-10X and the output is indeed quoted at 1200mA and so makes it an even better buy!

The mains plug itself is not over large either, although I have seen smaller micro USB ones, but at least it is easy to get hold of!

User avatar
rew
Posts: 435
Joined: Fri Aug 26, 2011 3:25 pm

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 4:11 pm

Hi,

My Raspberry Pi uses about 380mA when idle, and about 420mA when the CPU is fully active.

I just remembered I had a keyboard connected. I don't see the difference between keyboard connected and not.

My Raspberry Pi is currently driving an HDMI display at 1920x1080. The difference without the screen is about 20-30 mA. i.e. I get 370mA with the CPU idle and about 400mA with the CPU at full blast.

The ethernet chip is a power-hog. During the boot process the whole system drops to 90mA, which increases to 250mA without the ethernet, and then increases to 380 when the ethernet becomes active.
Check out our raspberry pi addons: https://www.bitwizard.nl/shop/

User avatar
jbeale
Posts: 3641
Joined: Tue Nov 22, 2011 11:51 pm
Contact: Website

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 5:10 pm

In my experience all ethernet chips are in the 100+ mA range, having to do with the need to drive a relatively low impedance. For example, on the RasPi schematics p.3 you see some resistors, each 50 ohms, pulling up  the LAN9512 TXP, TXN outputs to +3.3 V. On the far side of the cable, there is another such resistor.  It's connected through a transformer, but it's a high-frequency signal so I'll ignore that. 3.3V / 25 ohms = 132 mA.

selsinork
Posts: 151
Joined: Mon Apr 16, 2012 8:31 am

Re: Raspberry Pi Power Requirements

Sun Apr 22, 2012 5:41 pm

jbeale said:


In my experience all ethernet chips are in the 100+ mA range, having to do with the need to drive a relatively low impedance. For example, on the RasPi schematics p.3 you see some resistors, each 50 ohms, pulling up  the LAN9512 TXP, TXN outputs to +3.3 V. On the far side of the cable, there is another such resistor.  It's connected through a transformer, but it's a high-frequency signal so I'll ignore that. 3.3V / 25 ohms = 132 mA.



The LAN9512 datasheet, p42, says 231mA when running at 100M, 188mA at 10M. presumably this is while passing traffic.

Interestingly it also lists power consumptions for the magnetics:

100BASE-TX: ~42mA
10BASE-T: ~104mA

So that gives you the reason for the A @ 500mA and B @ 700mA.

Checking my cheap & nasty 1m micro usb cable, I was dropping ~0.5v across it at around 400mA and had ocasionally seen the repeating key problems.

I've now upgraded to some 16/0.2 conductor cable and over a similar length have around 0.01v drop. As far as I can tell that's stopped the problems under load.

hippy
Posts: 7360
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: Raspberry Pi Power Requirements

Wed Apr 25, 2012 7:56 pm

hippy said:


jojopi said:


Despite the fact that the schematic says 1.1A, F3 on both beta and production photos is a 750mA (hold current) device.  This protects the entire board, including USB outputs.  So it cannot be true that the 700mA figure for the Model B is intended to exclude its peripherals?


Or maybe the wrong fuse was fitted during manufacturing ?


I found this from Pete Lomas who designed the hardware ...


The [ USB protection ] fuses kick in hard around 280mA and fold back and limit to 140mA. If you remove them then all you have for protection is the 700mA inbound fuse.

http://www.raspberrypi.org/arc.....ves/tag/qa


So it seems that the circuit diagram is in error, as are the power consumption and current draw specifications on the Wiki. If it's a 750mA fuse, and each USB port is allowed to draw 100mA, then the rest of the board cannot draw more than 550mA. If the board draws over 550mA you can't draw 100mA out of each USB socket at the same time.

The good news is that rew's measured figures - the board draws 420mA with CPU maxed-out - fits acceptably within this scheme, the board itself actually consuming 2.1W

If it's a 750mA fuse then that's not the minimum that can be consumed, it's a maximum, and no one needs to look for a 1A supply because that's 250mA more than can ever be used. Which begs the question why it's being suggesting there's any need for a 1A or higher capacity supply?

It would be really good to get this issue cleared up once and for all considering there are people who are having issues which appear to be related to power supplies.

User avatar
jbeale
Posts: 3641
Joined: Tue Nov 22, 2011 11:51 pm
Contact: Website

Re: Raspberry Pi Power Requirements

Wed Apr 25, 2012 8:26 pm

I do not know about the RPi specifically, but here are some observations:

1) current drawn by a computer is often "bursty", not a smooth steady flow. There are often brief spikes of current, during a write to network, write to CF card etc.

2) a "750 mA" polyfuse will not blow instantly at 751 mA. Consider for example a Littlefuse 1206L075TH unit. The data sheet says Ihold = 0.75 A and Itrip = 1.50 A.

Ihold = Hold Current: maximum current the device will pass without tripping in 20 deg.C still air.

Itrip = Trip current: minimum current at which the device will trip in 20 deg.C still air

http://www.littelfuse.com/data....._1206L.pdf

In between 0.75A and 1.5 A that particular part may, or may not trip, it is not guaranteed to either hold or trip. (Note, my example is not necessarily the same as the actual part used on RasPi.)

3) It is also important to note that these fuses have a time constant, they are sensitive to *average* current over some period of time, perhaps seconds. If you have a short 0.1 second burst of slightly high current, the fuse will probably ignore it (unless it is an extreme overcurrent), but if your power supply droops or cuts out at that load, your computer will halt. In the data sheet, look at the "Average Time Current Curves" plot. The 0.75A fuse I mentioned can actually stand a 2 A current for over 1 second, and 10 A for about 30 msec.

4) Finally, cheap Ebay etc. power supplies do not necessarily achieve the full rating on the label. So I guess there could be more chance of a cheap 1A or 1.2A supply working, than a cheap 0.75A supply working.

User avatar
Jim Manley
Posts: 1600
Joined: Thu Feb 23, 2012 8:41 pm
Location: SillyCon Valley, California, and Powell, Wyoming, USA, plus The Universe
Contact: Website

Re: Raspberry Pi Power Requirements

Thu Apr 26, 2012 6:11 am

As has been mentioned, inadequate power cables are introducing significant voltage drops, and as current flow increases, so does the voltage drop across the cable.  Then, there's the internal resistance of the power supply itself, which is a demonstration of the Heisenberg Uncertainty Principle - in measuring and especially producing the output voltage, it is affected.

The maximum specs quoted for such low-cost/margin devices as cell phone chargers (made by the lowest bidders on contracts for millions of units, each) can be off by as much as 20% of their rated values because of the non-critical nature of the cell phone charging application and minimal attention paid to quality control.  Since such supplies are designed to maintain a constant voltage (a voltage source, rather than a current source), the current can (and must) vary in order to maintain the rated voltage.

However, if a particular supply happens to be on the low side of the performance bell curve for all of the devices manufactured in a lot, it won't be providing enough current, low by as much as ~140 ma for a nominally-rated 700 ma device!  That's not even counting the voltage drop through the power cable, either, so, you can wind up with both low voltage and low current from a marginal cable and power supply operating at its full actual capacity (not even what it's supposed to be providing).  Sometimes, you don't even get that for which you're paying.

The difference between devices operating at the upper and lower ends of the bell curve isn't a matter of design, it's a matter of statistical deviation in production output, which is impossible to predict for any single device.  At the extreme upper limit of this phenomenon are high-power microwave devices, such as mechanical resonance cavities used in microwave production and amplification components (e.g., klystrons and magnetrons) manufactured for radar and medical radiological systems (e.g., for X-rays and CAT scans).  One high-end cavity can cost upwards of a million dollars, not because there is that much invested in material and labor costs in that individual part, but, because, in order to produce one specification-compliant cavity, over 10,000 parts have to be produced, and about 9,999 of them don't meet the spec and the materials are recycled.

It's physically impossible to intentionally manufacture such a part - you have to make all ~10,000 parts and test each of them to find the one that is the Goldilocks part - not too big, not too small, but, juuuust riiiiight.  The same is true of semiconductor components - the fastest and slowest components in a production run of microprocessors, memories, etc., are all manufactured together on the same lines, and it's simply a matter of testing that determines which ones can fetch the highest prices and the rest will earn somewhat less to much less (and complete failures earn nothing).

I know this may be boring to the hardware folks, but, few others are aware of what it takes to produce every great wonder from a single transistor to an entire Pi.
The best things in life aren't things ... but, a Pi comes pretty darned close! :D
"Education is not the filling of a pail, but the lighting of a fire." -- W.B. Yeats
In theory, theory & practice are the same - in practice, they aren't!!!

hippy
Posts: 7360
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: Raspberry Pi Power Requirements

Thu Apr 26, 2012 9:53 am

Jim Manley said:

The difference between devices operating at the upper and lower ends of the bell curve isn't a matter of design, it's a matter of statistical deviation in production output, which is impossible to predict for any single device.  
Absolutely. In assessing the suitability of a particular supply or trying to determine why an R-Pi is not behaving as expected the key question has to be; is it delivering what the R-Pi needs?

If we don't know what the R-Pi actually needs it's impossible to determine, only to speculate.

It is almost impossible to determine which power supplies will work for all simply by plugging them into an R-Pi; one may work for some but not for others, may work under some test conditions but not others.

If there is a problem with a particular supply and R-Pi combination it should be possible to test the supply separate to the R-Pi and determine whether it is suitable or not, determine if the problem rests with the supply or R-Pi. Without knowing what the R-Pi draws, what the supply must deliver it may be impossible to to do that.

That's why it's so fundamentally important to know the R-Pi worst case current draw.

Return to “General discussion”