Request urgent help , newbie Raspberry Pi ModelB issues

The correct fuse should allow for the amount drawn by the Pi added to the maximum current draw allowed by the USB port. The USB standard allows only a limited current to be drawn, which can be altered by negotiation between the hub and what's plugged into it.

--
Tciao for Now! 

John.
Reply to
John Williamson
Loading thread data ...

just plug a powered hub in both ends. that way accessory power doesn't flow through the rPi and the usb power cable can be short

--
?? 100% natural
Reply to
Jasen Betts

I haven't looked at a schematic but the likely reason for the input fuse is to work with a reverse biased diode that will crowbar the supply in the event it is connected with reversed polarity or to a LV AC supply.

Looking at the underside of the board I would put my money on D17 being that component.

--
Graham. 

%Profound_observation%
Reply to
Graham.

I was wondering if one could feed in power via the GPIO pins.Pin 1 is 3.3 V, 2 is 5.0 V and pin 6 is I think GND. That would make the power supply and RPI a self-contained unit without having to get external powered USB hubs and so forth.

Reply to
dakupoto

This is documented somewhere. The recommendation is that you take steps to ensure your alternative power to these 'raw' pins has polarity and fuse protection of it's own.

I dont see the point of bypassing the supplied protection, just power the extra devices direct from the power supply or easier still, use a powered hub. On my setup I run two wires from the 5V input on the powered hub circuit board to the protected input side of the fuse and the card mount earth. If one USB port is driving a keyboard/mouse and the other goes to the hub usb input, the standard fuse system is adequate and very suitable for protecting the pi. I can recall cheap electronic goods that could be blown if the polarity was wrong, and I cursed them when one cheap diode could have saved the whole device. I think the pi's protection is something to be pleased about. Even if the special fuse develops a fault it is large enough to replace easily and indeed it is where I attached my positive in wire also.

Ron M.

Reply to
Ron

You can feed regulated 5v in. Don't feed 3.3 as the Pi generates it on-board. Some sort of protection might be advisable - it something on-board accidentally shorts the 5v (e.g. an external USB peripheral) there is concern about the PCB tracks ability to carry more than the

700mA that the regular polyfuse allows. Some people have soldered extra wires from the microUSB to the main USBs to totally bypass the fuses too.

Gordon

Reply to
Gordon Henderson

I'm only lurking in this NG because I'm merely, at the moment, contemplating the use of a Rpi for for XBMC as a custimisable and much cheaper alternative to the crappy commercial alternatives (I once 'borrowed' one from Aldi for a 50 quid deposit a year or two back, returning it because it didn't work 'as advertised').

Having said that, the various topics of discussion that come up in this NG impell me to contribute from time to time and this is one of those times.

The idea of using an anti-reverse polarity protection diode which blows the fuse rather than simply by blocking the reverse voltage is a very old one, probably predating its use in CB radios that became available in the UK back in the early 80s.

The problem with using a series wired polarity protection diode on equipment designed for 12v car battery power is the significant volt drop it introduces to anything that uses more than a trivial amount of power, such as CB radios and the like.

Since the reverse polarity issue should only arise infrequently due to initial installation mistakes, putting the diode in parallel so that it sacrifices a "Ten Cent Fuse"[1] neatly avoids the volt drop issue[2]. Indeed, this alternative approach to the problem of incorrect polarity connection was exactly the one I took when I designed my hybrid "Top Band/Marine RT" 10 watt output transciever back in 67 / 68.

The transciever was all solid state apart from a 6BW6 pentode for the Tx final stage. The 'Key Down' power consumption must have been in the region of 30 watts or more suggesting the need for at least a 3A fuse.

I chose a 5A fuse and wired a 50 or 60A rated diode across the fused supply (I took the view that in situations like this, it really, really pays to err on the side of using a 'massively over-engineered' solution rather than rely on a merely 'adequately engineered'[3] solution - it neatly eliminates the "Fingers Crossed Syndrome" when connecting up to the battery).

Provided you don't bypass the 700mA polyfuse to permit each of the two USB ports to supply half an amp each, the diode is more than adequate for the job. If you are going to strap out the polyfuse for such purpose, it's best done when the Rpi is mounted in a box large enough to allow you to fit an additional higher current rated reverse polarity protection diode with a 3A polyfuse[5] on the incoming supply rail.

[1] For once (and at long last) a case that was the opposite to the more usual one described by the definition for a transistor as being "A ten dollar device designed to protect a ten cent fuse" (provided you over-engineered the diode rating sufficiently to guarantee this would remain true for most classes of "Idiot" - no _absolute_ guarantees against the more "Inventive Idiot" can be assumed of course

- you'd be an idiot yourself to make such an assumption, there _are_ limits, after all!).

[2] A series diode will avoid relying upon a sacrificial 'ten cent fuse' of the correct rating but introduces a significant volt drop when you have only 12 volts to start with. Also, it has to have a continuous current rating that at least matches the peak consumption, preferably higher to withstand the peak fault current during the time it takes for the (correctly rated) fuse to blow.

Since it may be carrying a significant load current, it may also require 'heat sinking' measures to avoid overheating in service. Fitting a diode across the supply on the load side of the fuse eliminates these issues altogether provided the amperage rating is suitably 'over-engineered'[4].

[3] The protective diodes I used to see being used in CB radios were, IMHO, barely adequate for the job, typically 2 or 3 amp diodes with a 1 or 2 amp fuse which could so easily be replaced with a 3 or 5 amp one (or even a blown fuse wrapped in foil!).

Whilst you can almost guarantee that a diode in a high voltage rectifier bridge will fail short cct due either to transient voltage peaks exceeding the diode's PIV rating or the effects of high temperature reducing the effective PIV rating, this isn't necessarily going to be true for excessive forward current overloads.

Relying on the diode to be the second volunteer sacrificial lamb to protect against reverse polarity is rather chancy to say the least. Even if this the case, statisticaly speaking, in 9 out of 10 events, it's still an expensive game of "Russian Roulette" being played out by the manufacturer's bean counter dept.

I'm sure that any 'Real Engineer' (not thwarted by bean counter measures) would elect to over-engeer the diode so that the chances of a really inventive Idiot's actions earning them a Darwin Award are significantly improved.

That's my attitude. The kit shall not suffer the abuse of reverse polarity come Hell or High Water, even if it means consequential damage or injury external to my precious box of tricks. Such consequential damage and / or injury are _NOT_ my concern, those are the idiot's responsibility alone.

[4] I think these diodes are often under-rated on the false premise that they can withstand transient overloads of 50 to 100 times the continuous rating for micro-seconds, neatly ignoring the tens of milliseconds it takes the fuse to 'Do The Right and Honourable Thing". IOW, they're optimists of the worst kind (the 'fingers crossed, hope all goes well' kind). [5] A cheap source for 3A polyfuses are scrapped MoBos where they're used to protect the 5v feeds to PS/2 and USB sockets (typically a common 5v bus to all of the above - no negotiated half amp per USB socket bullshit here!).

If your only source of scrapped MoBos are PC Chips models, don't waste your time looking for any polyfuses. PC Chips went for the cheaper option of employing SMD inductors or the thinnes of circuit traces to emulate the classic one shot fuse effect (PC Chips were bad at _so_ many levels, it was a wonder that they stayed in business for as long as they did).

--
Regards, J B Good
Reply to
Johny B Good

There is an even older protection method. A forward diode operates a relay, which connects the incoming power to the unit to be protected. If it is reverse connected, the relay doesn't switch in. No voltage drop to the load, no sacrificial components.

Of course, it's difficult to make it small enough for an rPi board.

Back to the original topic - I still can't see the point in bypassing the fuse in order to power other usb devices - better surely to use the 5v supply to power both the rpi and a usb hub. There is a significant risk that the tracks on the rpi board won't handle the unanticipated extra current, and either fail, or drop so much voltage on the way to rpi components that the rpi will misbehave.

--
Alan Adams, from Northamptonshire 
alan@adamshome.org.uk 
 Click to see the full signature
Reply to
Alan Adams

It's even older than that. Crowbar diodes were used in the very first synchronised switching PSUs in transistorised colour TVs of the early 1970s. They were a bugger to fix!

--
W J G
Reply to
Folderol

Yes, but the complication is when someone wants to use the pi without a hub. Gordon's mention of people running a wire from the input power (before the Pi fuse) to the USB socket 5V power pins sounds good at first, as the Pi can't be powered from the USB out ports. The remaining question though, is the pi protected from a reversed polarity at the USB out socket 5V? Ron M.

Reply to
Ron

there's a third way to do polarity protection.

fit a MOSFET so that it's biased on by the supply and the body diode faces in the right direction.

--
?? 100% natural
Reply to
Jasen Betts

I'm suspicious of that route. In normal operation the 5v in at the microUSB will be fed through to the 5v out of the main USB. There is unlikely to be anything other than a fuse between them, as that would produce an unacceptable voltage drop, so I strongly suspect the power from the main USB would feed back into the pi. I'm not about to try tracing this on mine though - today's busy enough already.

The only way I can see to create a setup where the 5v out can't also act as 5v in would be to feed the normal 5v in to a switch-mode pus, which creates a higher voltage, that is then supplied to the main USB socket via a blocking diode. That seems unlikely to be what the pi has.

--
Alan Adams, from Northamptonshire 
alan@adamshome.org.uk 
 Click to see the full signature
Reply to
Alan Adams

Choose you MOSFET very wisely to ensure you can create the correct Vgs to make sure it is on as hard as it will go on. Then your Rds will be low enough to start getting close to what a relay contact impedance is.

Seen too many MOSFET installations where folks forgot that Rds is dependant on Vgs. Hence ended up voltage drops ranging from 100mV to

700mV.
--
Paul Carpenter          | paul@pcserviceselectronics.co.uk 
    PC Services 
 Click to see the full signature
Reply to
Paul

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.