SheevaPlug USB Power Supply problems. Can it Drive a USB Ext Disk?

[superpat]

In the Hardware thread, "is my shaevaplug dead", http://plugcomputer.org/plugforum/index.php?topic=1318.0, there are two or maybe three users who appear to have a broken sheevaplug  internal  mains to 5V power supply unit.

The common theme in these faults appears to be connecting  and running an external USB 2.5in disk drive that is drawing it's 5V supply via the sheevaplug's USB connecter.

Looking at the pictures of one of the blow ups, it appears that C1, a 6.8ufd 400v electrolytic capacitor dumped its electrolyte into the supply. This capacitor is connected across the o/p of the full wave bridge rectifier (4 diodes) that rectifies the incoming ac mains voltage.

Now a digression on electrolytic capacitors.  (Wikipedia has a full explanation for enquiring minds)

In the theoretical  world, a capacitor consists of two plates and an 100% insulating dialectric. In the real world,  the dialectric has a dc leakage resistance and an ac leakage impedance. This impedance is composed of a reactive and resistive component. The in phase resistive component is called the "Equivalent Series Resistance" (ESR).

 The ac component in the current flowing in the capacitor, is caused by the ripple on the dc current. If the capacitor is drawing a CONSTANT current, there is only the ripple current is the charging part of the cycle when the  mains bridge conducts and charges the capacitor, and the constant load discharges it to the mean value. However if the output of the capacitor is connected to a rapidly changing load, then the ripple can assume a large percentage of the total supplied current. 

The ripple current is converted to heat (Irip^2 x ESR = watts) and this causes the electrolyte to heat.  When the electrolyte heats the ESR increases and the capacitor can go into thermal runaway. To stop the capacitor exploding violently the top of the capacitor is scribed in a cross, as a deliberate weak point.  Capacitors that are on the way to self distruction sometimes bulge the  flat end out into a dome instead of a flat before they vent.

To prevent this, capacitors that are going to be used in a power supply or high ripple environment ( audio amps etc), are normally carefully chosen for a low ESR and a high nominal operating temperature.

There are also stories that many manufacturers used a "magic" formula from another manufacturer and got it wrong, and made the ESR problem worse. PC motherboards have suffered from electrolytic cap failure for several years, and if you Google for motherboard cpacitors you will find dozens of sites talking about it and the replacement caps procedure.

Returning to the Sheevaplug power supply.

I was worried about these failures so I stripped the power supply out of my  sheeva plug and had a look at the manufacturer and type of capacitor used.   I found the two main electrolytics were Chinese in origin, Yalecon RX 400V, 105 deg C devices.   I searched vainly for a proper specification for these capacitors. All I found was a one line buying spec from one of the Chinese sales factors. The capacitors are manufactured by  Gagaku Yiyang Electronics Co,and have the following characteristics (Example RX only):-

Brand     Domestic      Model      RX
 Dielectric    Aluminum    Applications  Filter
Profile    Cylindrical    Power characteristics     Power
Frequency    HF     Adjusting Fix
Lead Type    In the same direction pinout    Tolerance    ± 20 (%)
Withstand voltage    500 (V) Nominal Capacity    68UF
Loss    5.0

The USB V 2.0 spec defines that the host shall supply up to 0.5A at 5V per port.

A typical 2.5in 500GB hard disk drive spec, my Samsung HM500L1 states the power consumption is:-
      
Power Requirements:
- Voltage: +5V+/-5%
- Spin-up Current (Max.): 1000 mA
- Seek (typical): 2.6 W
- Read/Write (typical): 2.3 W
- Low Power Idle(typical): 0.6 W
- Standby(typical): 0.25 W
- Sleep(typical): 0.2 W

So on idle its using 0.6W = 0.12A,  but when seeking it uses 2.6W or 0.52 A

Depending on HOW MUCH FILTERING is in the external enclosure, the drive is going to create some or a lot of ac ripple in the Sheevaplug 5V supply.

This is ignoring the ripple of the sheevaplug SOC's  own  current consumption variations

The other problem is that the capacitors are "domestic" rated at 105 deg C max.   There are higher quality devices on the market rated at 130 deg C.

I have NO knowledge what the internal operating temperature of the Sheevaplug power supply is when it is drawing a heavy load.

Also I have NO idea what the design output current for the Sheevaplug power supply is.

Therefore I would recommend that you power your USB hard drive enclosures from a separate +5v wall wart or brick

WARNING....  MAINS SWITCHING POWER SUPPLIES, ESPECIALLY THE FILTER CAPACITORS CAN HAVE SEVERAL HUNDRED VOLTS ACROSS THEM EVEN AFTER THE SHEEVAPLUG IS DISCONNECTED FROM THE MAINS. THEY CAN  KILL OR GIVE YOU A NASTY FLASH BURN.  OBSERVE ALL NORMAL HIGH VOLTAGE SAFETY PRECAUTIONS WHEN WORKING ON MAINS SWITCHERS. 

If you don't know the normal HV safety rules then you should not be into a mains switcher!

cheers

Patrick

 





 

[restamp]

An interesting and informative post, superpat.  Thanks.

The 1000ma spin-up current you mention as typical is outside of the limits for USB2.0 supplied power, but I don't doubt many of the USB-powered hard drives want more amps than what is called for in the USB specs.  Mine offers an "optional" (not supplied) USB cord that plugs from the drive into two USB ports, apparently to resolve the case where one port is not be able to provide sufficient power.

I suspect your assessment is probably correct -- that the Plug's power supply is not designed for the load required by an external USB powered drive.  Unfortunately, there is probably no circuit in the Plug to limit the power provided to the USB port.  I use a powered hub with mine, and so far it has worked well.  (I'm actually running two USB powered drives from it as I type this.)

Unfortunately, at the current juncture, it appears that USB-powered hard drives are a catch-22:  I know of no one who has been successful at reliably booting from a HD attached to the Plug through a hub, and if you attach a USB-powered drive directly, the Plug will be providing the power.  That's one of the reasons I've elected to stay with an SDcard for my Plug's root FS.

Perhaps the eSATA port offered on one of the next generation of Plugs, the GuruPlug Server Plus, will offer a solution to these problems, although with 2 or 3 separate USB ports now brought out, if they haven't beefed up their power supply, I suspect it will make such PS failures even more prevalent.

[superpat]

@restamp

Hi

I don't think an occasional spinup 1000 ma is going to be a problem.

I think the problem is constantly varying dc load plus maybe a high overall current requirement and finally perhaps elevated temperature in the unvented metal ps enclosure in the Sheevaplug base.

The ripple current varies as the dc load on the power supply.  During the conducting period of the bridge rectifier, the cap is charged up, and the bridge rect supplies BOTH the charge current and the current for the dc load. During the "off" period of the bridge rectifier, the cap has to provide the dc power for the load, as the charge is used in the cap the voltage across the cap decreases. On the next conducting period of the bridge the cap recharges. Obviously the greater the current drawn, the more "ripple" current is generated.

I will dig out a couple of 10w  0.1 ohm resistors from my parts box, and wire them into the 5v output  and do some 'scope measurements of the current amplitude and waveforms. I will also try to measure the operating temperature inside the power supply case, if I can get my remote temp probe into it.



cheers

Patrick

[restamp]

Hi superpat,
FWIW, I agree with you on the spin-up current; mainly I'm just surprised a drive would exceed the USB2 power specs by that amount.  I think, technically, all USB devices need to come up in a low-power mode and then negotiate with the USB2 controller before attemping to draw more than 100ma.  The controller can always shut a device down if it exceeds 100ma, or 500ma if it has negotiated a high-power profile, but practically I suspect that most, including the Plug, don't do so;  they just attempt to supply the necessary power to the load.  (I do wonder what such a load, even for a relatively short duration, does to the voltage supplied to the processor, though.)

I also agree with you that most capacitor failures are due to either Capacitor Plague or not paying attention to ESR issues in the design.  I've brought several computers and monitors back to life by replacing the obviously bulged caps.  Someone did post a picture of their PS recently with some gunk covering several components.  It was unclear if the gunk was simply excessive glue that was dribbled over the components during the manufacturing process or the remnants of an exploded capacitor.  If the former, it appeared that it had been subjected to a lot of heat.  If the latter, it appeared to be from a component above the PS board, and not one of the PS components itself.

In any event, based on what I've seen here and elsewhere, I've come to consider the Plug's PS a relatively delicate component, and have personally taken some pains not to subject it to excessive current demand.  I look forward to hearing what you find, both thermally and electrically, when you stress-test your Plug's PS.

[cjm]

I'm also very interested in the smoke test results. Thanks for volunteering to do all the dirty work!

I've had a few switching power supply failures due to bulging or venting capacitors myself, the last one being an NSLU2 power supply. It's usually easy to put in a new capacitor with better ESR ratings and the power supply will work again. Just keep in mind superpats warnings about the high voltages in switching power supplies! Given the relatively low price of those caps, it makes me wonder why manufacturers are so cheap about them...

BTW, you can usually connect a power supply directly to the USB disk enclosure so you can apply separate power to the HD without having to use a hub.

BTW2, unless you mercilessly overload the PS to the point of explosively venting capacitors, a typical failure pattern will be that the capacitor doesn't stop working right away; instead, it will slowly lose capacity. This process will accelerate as the electrolyte degrades and by the time you notice something, you probably have only a few days left. One of the things you may notice is that the HD doesn't spin up anymore, trying again and again and generating clicking noises as it does so. But even if you disconnect the HD now, the capacitor is doomed and even a normal plug power load will destroy it in a few hours or days. I had the exact same thing happening on a NSLU2 and ended up powering the NSLU2 by plugging a power supply into the HD enclosure (the NSLU2 can actually be powered by the USB port -- not sure if this is the same with the plug

Thanks,
--Christian

[birdman]

mainly I'm just surprised a drive would exceed the USB2 power specs by that amount.

A lot (most?) bus-powered USB drive comes with 2 USB connectors precisely because 1 can't drive them if it runs exactly to teh max power spec (but many ports run over spec for power).
Certainly, plugging in my external USB drive (which has 2 SUB connectors, but runs OK off one when connected to a desktop or laptop) causes my Plug to become unusable (but the port is driving a Wifi dongle too, so perhaps that's not too surprising).

[digitlman]

If your power supply dies, would it be possible to power the plug off a USB wall wart?  It looks like the power supply drops the house 110v down to +5v, .5a, right?

I was thinking you could strip the wires on a usb and the wires from the connect point on the Plug board and solder them together....

Wishful thinking?

[K-Ray]

If your power supply dies, would it be possible to power the plug off a USB wall wart?  It looks like the power supply drops the house 110v down to +5v, .5a, right?
I was thinking you could strip the wires on a usb and the wires from the connect point on the Plug board and solder them together....

I don't think that is going to work properly. The plug takes around 5W in normal use. That means you need at least 5V/1A. That's what most  usb wall warts are.
But at full usage the plug can draw up to 2A and then you are going to be in trouble.
The inside power supply is good for 3 or 4A. If you find a power supply 5V/3A or so you can do it.

[obarthelemy]

I just found out that the WD Scorpio Blue have a Reduced Power Spinup setting that might help: peak power goes from 900mA to 850 (which is not a lot :-p) but the duration that peak power is required goes from 800ms to 125. Spinup time increases from 4.5s to 5.05. My guess is that does help for startup problems, and very little inconvenience. A dual USB cable still sound prudent, though.

Source:
http://www.westerndigital.com/en/library/portable/2579-001116.pdf (old version, diagonal jumper)
http://wdc.custhelp.com/cgi-bin/wdc.cfg/php/enduser/std_adp.php?p_faqid=2670 (new version, straight jumper)

[scaramanga]

Thanks guys for all the info. Especially the last post with the info about WD scorpio blue spin-up options. Very useful!

My setup is like this : the plug, Debian OS on 8gb flash drive + attached 2.5 inch usb hard drive (500gb WDC). I bought it in Nov 2009 and running torrent box (transmissiond) 24h/7.

Here's my story:

-Noticed strange sounds coming from the HD about 2 months ago. Like others suspected the disk is becoming instable. (wrong!)
-This went on for a few weeks, sometimes with no noise. Absolutely no r/w problems with the disk.
-Then, last week disk has read problems. Ok, I thought, the disk is ready for RMA. But to my surprise everything was readable on a Win7 box. (I made a full backup just in case)
-I googled and found this forum with all the folks posting their similar experience. (Power of the web)
-Opened the sheevaplug and found the PSU intact (no cap has blown), but close: One of the caps was bulging!!
-Went down to an electronic hw store and got the 5V 2A Delta Electronics PSU. This is now external and working very well with my setup.

Anyway, what I like to say is, find an external PSU and don't support this crappy company called Globalscale Tech.!

First, the sell you a shitty PSU inside the plug, then they make you pay for another 10usd for a spare PSU (who knows when you will get this and how long it will work with your existing plug).

Customer is supposed to be king! This company doesn't give a damn about you and keeps you guessing whats going on for weeks (when I placed my order).

I got my Delta Electronics (this is a very reputable manufacturer) PSU for less than 5usd!

For me one thing is clear: I'll never deal with those guys (Globalscale Tech) again.

Have fun plugging!
cheers
scara

[CqCn]

The better solution for an external pwr supply is to use small 5v switcher type wall wart.  These days even the usb types for many of the newer phones have rating of about 1 to 2 amp.  Many of these supplies seems to last for ever, compared to the ad hoc designs some companies do specifically for the Plug etc.  You need to buy one designed by a pwr supply specialist.

To handle the peak load fluctuations mentioned in a couple of posts below, one simple way is to put 4 AA NiMh rechargeable cells right across the power supply output.  The newer ones have 2500mAHr capacity,  so they would provide 5+ A for short durations, and also will serve as a good short term power backup for upto 30 minutes plus easily.  BTW, 98+ % of power outages last less than 5 minutes in most places in the USA.  At the typicall <5.5v DC, 4 cells will never be over charged.  You can buy 4 for <$6 these days.

[swaery]

Well, there have been reports that the Sheeva plug has been having constant issues regarding broken internals and I think much of that could be because of connecting and the running of external USB drives which tend to be interfering with the 5v supply! So since the root cause of the error is the USB drive, I don’t think it can drive a USB external disk!

[AmiePoat]

Well, there have been reports that the Sheeva plug has been having constant issues regarding broken internals and I think much of that could be because of connecting and the running of external USB drives which tend to be interfering with the 5v supply! So since the root cause of the error is the USB drive, I don’t think it can drive a USB external disk!

That's true. A USB fix would be the best solution ..

[radael]

Actually, from the reports in this forum, the problem occurred mainly with one or the other of two conditions:

• Using an unpowered USB device
• Running on 240V mains power.

So, it was not only the USB drive that caused the problem.

Most of the reports seem to be associated with the early power supplies.  GlobalScale reported that they got a new supplier for that part (who can say if that is actually true?  I think the "research" is ongoing.)

In another thread, there was an estimate that using a 5V PSU, switchable up to 2.5 to 3.0 Amps, would be enough to handle a disk without over-taxing the power supply.