element14 (AU) power supply for Raspberry Pi - poor voltage

What voltage do you measure when it is not connected to the Pi?

Colin

Wish I could answer this for you, but I have no method of being able to place the multimeter leads on the MicroUSB plug's pins, since they are tiny and recessed within the plug. I tried using some breadboard wires, but didn't succeed in measuring any voltage. I have just purchased some microusb line sockets that will make it much easier, but you know what ebay purchases can be like .. might take a while to get to me..

Damn

I've just ordered one of these.

I have been able to find another, but these are specifically for the Pi and at a good price.

I hope it doesn't turn out to be a wasted exercise.

Mark

Mark, I'm curious to know what voltage you get with yours. Hopefully just mine are duds.

Actually, 4.77V is within the specified RasPi range for TP1-TP2 which is the standard USB Vbus range of 4.75V to 5.25V.  A proper USB power supply is usually 5.1V which helps compensate for the voltage drops across the micro USB cable and polyfuse F3.

You might try checking the voltage drop across polyfuse F3.  It may be that the power supply is fine, but your RasPis have iffy polyfuse F3 which have higher voltage drops than normal.  However, since you've checked two RasPis I would suspect the power supply and its cable.

You can also check TP1-TP2 with no cables attached and the SD card removed.  This should have minimum voltage drop so you can get close to what the power supply would be with no load.

Some micro USB cables have very thin conductors which result in a substantial voltage drop, and replacing them fixes the problem.  However, according to the photo your power supply has the cable built in so there's no way to run this test.  I suspect that the Chinese manufacturer has skimped on the conductors in the cable and that's where the voltage drop is.  I just checked the spec sheet on the supply and it lists the cable length as 1.8m.  That's a lot of resistance with thin conductors -- IMO you're much better off with 1m or less.  If the problem is the cable, element14 needs to know that its RasPi-targetted power supplies "have been measured and found wanting".

Lots of power supply troubleshooting info at the RasPi Troubleshooting wiki: http://elinux.org/R-Pi_Troubleshooting#Troubleshooting_power_problems

Nev, there's a long history of official "Pi accessories" not necessarily being Pi-compatible --- take a quick peep at my old thread about Incompatible self-powered hubs .

I suspect that  the Pi's  low cost and the tiny profit margins mean that only a very cursory attempt to verify Pi compatibility is being made, because to invest any significant amount of engineering time would wipe out the business margins entirely.

It's not an ideal situation, but it's hard to see a sustainable solution for a commercial enterprise like Farnell.  Maybe they should avoid claiming Pi compatibility entirely and leave that matter for the community to judge.  Alternatively, perhaps RPF could give certain items its stamp of approval, since it is best positioned to have good knowledge of its board design and hence of accessory types or brands that could cause Pi a problem..

John Beetem wrote:

 

You might try checking the voltage drop across polyfuse F3.  It may be that the power supply is fine, but your RasPis have iffy polyfuse F3 which have higher voltage drops than normal.  However, since you've checked two RasPis I would suspect the power supply and its cable.

hi John, I think you may have misread .. I checked only one RPi at the time, but I had purchased 2 of these exact same adaptors (one adaptor showing 4.77 volts and the other showing 4.69 volts when fully booted with XBMC, HDMI & LAN connected, voltage was checked over TP1/TP2).

I have just performed some more tests as per your suggestions

My original RPi adaptor is an old Canon Printer AC adaptor, made in Japan (about 15 yrs ago), output is 13V 1.8A, and I've attached a down-step converter bringing it down to 5.2 volts (it has a variable resistor to set the voltage, so I should bring it down a bit more I know). Power is sent through the GPIO pins.

Canon Adaptor, RPi with everything removed, only power connected:

Input: 5.20V (tested before connecting to RPi)

TP1/TP2: 5.18V

TP2/F3 (outside side): 5.18V

TP2/F3 (inside side): 5.18V

TP1/TP2 Raspbmc booted (SDCard, & HDMI only): 5.11V

Element14 AC adaptor, RPi with everything removed, only power connected (I tested with the worst performing out of the two Element14 adaptors I purchased):

TP1/TP2: 4.84V

TP2/F3 (outside side): 4.86V

TP2/F3 (inside side): 4.84V

TP1/TP2 Raspbmc booted (SDCard, & HDMI only): 4.72V

The above voltage tests were done on a 512Mb RPi. I also tried the TP1/TP2 test on the above Element14 AC adaptor using an old 256Mb RPi. Result was 4.83V.

Nev

John Beetem wrote:

 

Some micro USB cables have very thin conductors which result in a substantial voltage drop, and replacing them fixes the problem.  However, according to the photo your power supply has the cable built in so there's no way to run this test.  I suspect that the Chinese manufacturer has skimped on the conductors in the cable and that's where the voltage drop is.  I just checked the spec sheet on the supply and it lists the cable length as 1.8m.  That's a lot of resistance with thin conductors

That's correct, it's a long, fixed cable. I am tempted to crack open the adaptor to measure the voltage at the start of the output wire.

These actually make a decent, cheap, phone charger. Nice long cable, and has the Australian certifications.

Just read that thread of yours Morgaine, interesting stuff.

As for tiny profit margins, I understand, but at the same time how much would it add to the cost of each adaptor to get it right .. better parts/design, and thicker cable ... $3 ? (just very much a guess, I would have no idea!) Even at $AU9.00 it would be decently priced.

In fact, here in Australia any element14 order for at least $45 gets free shipping. The RPi is $36. If they were able to make a decent AC adaptor at a $9 selling price, then I'm sure they would sell loads of them as they would tip the order into the free shipping zone.

 

Morgaine Dinova wrote:

 

Nev, there's a long history of official "Pi accessories" not necessarily being Pi-compatible --- take a quick peep at my old thread about Incompatible self-powered hubs .

 

I suspect that  the Pi's  low cost and the tiny profit margins mean that only a very cursory attempt to verify Pi compatibility is being made, because to invest any significant amount of engineering time would wipe out the business margins entirely.

 

It's not an ideal situation, but it's hard to see a sustainable solution for a commercial enterprise like Farnell.  Maybe they should avoid claiming Pi compatibility entirely and leave that matter for the community to judge.  Alternatively, perhaps RPF could give certain items its stamp of approval, since it is best positioned to have good knowledge of its board design and hence of accessory types or brands that could cause Pi a problem..

Nev

I will do when I get it....

In the meantime I thought I would check the one I have been using a Brightlink 1A travel charger from The Warehouse

Its out of spec.

No card/cables = 5.37

Login prompt   =  5.26

Graphic screen = 5.25

I also have  Galaxy Y charger supplied with the phone that is 0.7A ...

No card/cables = 5.07

Login prompt   =  5.02

Graphic screen = 5.04

So hurray for Mr Samsung.

Its also got a 1.5m tiny cable, but obviously better internals.

Mark