My Raspberry Pi is a Lemon

[jamesh]

There are more than likely undiscovered flaws as well, but since they are undiscovered after close on a million sold, they cannot be that bad...

One cannot usually tell how bad flaws are until they are discovered or, more correctly, start affecting things.

A million sold doesn't mean there isn't an undiscovered flaw which could be severely damaging when it does reveal itself in the future.

All non-trivial products present such concerns; so I'm not criticising the R-Pi, just the notion that quantity sold and/or current lack of evidence of flaws means there isn't a flaw or any flaw will not be serious. It doesn't.

I disagree. How can a bad flaw suddenly appear, and affect the whole million? If there were a bad flaw, then surely it would have appeared by now? The only possible thing I could think of was some aging problem with a component, but they have all been thoroughly tested. I can see a bad flaw appearing in a very minor subsystem, used by very few people (or again, it would have appeared by now)

That said, if it happens, I'll buy you a pint.

[jamesh]

You get what you pay for.
The Pi needs an extremely accurate 5V supply.
When the TP1 TP2 voltage drops below 4.85V, nothing is guaranteed anymore.
It has some usb issues. Streaming usb devices like webcams and audio adapters usually work if you don't push them to hard.
If you try a high resolution webcam or a 24 bit surround audio adapter, you will likely run into problems.
Some usb2serial adapters with older ftdi chipsets don't work very well either.
Usb keyboards (specially wireless ones) attempt to lose keyup or keydown events.
The usb port supply pins are directly connected to the 5V supply.
Hotplugging usb devices can cause a reboot caused by the inrush current and the temporary voltage drop due to that.
The analog audio out is not fully 16 bit and has a high noise level due to the fact it uses the same ground plane as all the digital surroundings. It produces an annoying plop when the audio driver is turned on and off.
The os normaly boots from an sd card. This requires to always properly shutdown the system, as the sd card background wear levelling algoritms can corrupt the sd card when power is removed during a write operation to the card. Making the root fs read only might prevent such writes. Most oses also use a swap file. Due to the limited write cycles of sd card flash technology, this might shorten the sd card life.
The system comes without real time clock and requires synchronisation with a time server to acquire the correct time.
The cpu core is an arm V6 architecture. Some linux distro's don't support this architecture anymore.
It's having a capable GPU, but the linux graphical X interface isn't currently using it.
Most of the issues are being worked upon, but due to the fact that the foundation is a charity with only one payed employee, no one can guarantee when and if they will ever be fixed.
They already sold approx 1000000 pieces, so it looks like for most people those minor issues are not a problem.
Let's not forget: the Pi is a low cost platform for educational purposes.

Have you taken that job at Cubieboard?

Dear reader, the above list of issues may seem a disaster, but it isn't. Most people see few if any of the issues above, and many of them are actively being worked on to improve the situation. Saying just one employee does not really cover the work being done.

obcd, please give it a rest. If this were product forum for almost anything else that post would have been immediately deleted. Next one like it will be.

[ShiftPlusOne]

I don't quite like the "you get what you pay for" approach. You should get what you were promised, and that is the case for most buyers. It is not ok that some people are getting defective units and they should have them replaced for ones that work properly through the distributor/manufacturer. I do think the list obcd gives is legitimate though. Those are all problems I ran into, but only the USB ones bothered me (at launch). All of the ones which affected me have been fixed and I am sure the rest will be too.

Anyway, the point is, the faulty units are not the norm and you have every right to have them replaced.

I don't quite like the "you get what you pay for" approach. You should get what you were promised, and that is the case for most buyers. It is not ok that some people are getting defective units and they should have them replaced for ones that work properly through the distributor/manufacturer. I do think the list obcd gives is legitimate though. Those are all problems I ran into, but only the USB ones bothered me (at launch). All of the ones which affected me have been fixed and I am sure the rest will be too.

Anyway, the point is, the faulty units are not the norm and you have every right to have them replaced.

You do get what you pay for.

This device is still at pre-release

The full educational roll out is yet to come and that is its stated aimed target use.

Ok there are fast approaching 1,000,000 units out there not quite that number of users because quite a few of us will have >1 RPi its not the foundations fault that there are a lot of us who want to play.

So far we have seen NO reports of refused replacement of defective units(lot's of bitching about delayed delivery and third party profiteering) none of it by the foundation.

I would therefore be interested to learn what you think was promised becuase mine has done everything that was expected within the parameters set when I bought it.

[ShiftPlusOne]

Sorry, I guess I was a bit ambiguous there. I got exactly what I was promised and then some. It seems like some people tend to dismiss the defective units with "hey, it's $35, what do you expect?" and I am saying that the price doesn't matter, if you've got a faulty unit, you should get it replaced, that's all. I myself have ordered 3 raspis and they all work fine and are free of manufacturing defects, so I certainly wasn't trying to be negative, quite the opposite.

Have you taken that job at Cubieboard?

Dear reader, the above list of issues may seem a disaster, but it isn't. Most people see few if any of the issues above, and many of them are actively being worked on to improve the situation. Saying just one employee does not really cover the work being done.

obcd, please give it a rest. If this were product forum for almost anything else that post would have been immediately deleted. Next one like it will be.

Come on James, the appraisal by obcd was pretty accurate and fair . Does your job at Broadcom depend on wringing every last sale out of the Pi ?

You've been on my case for perhaps overdwelling on some of the Pi's issues (USB). I'm still here because I give out some useful advice (I have it on good authority) ,does obcd fall into that category as well ?

The Pi has faults, but it didn't stop me buying 3 of 'em and I'll be first in line for an 'A' and a camera module when they become available.

If your job is on the line, then you have a case and I'm out of line, but I think you should declare it.

I was only clarifying in case someone read you the wrong way ShiftPlusOne That probably included me

[jamesh]

Have you taken that job at Cubieboard?

Dear reader, the above list of issues may seem a disaster, but it isn't. Most people see few if any of the issues above, and many of them are actively being worked on to improve the situation. Saying just one employee does not really cover the work being done.

obcd, please give it a rest. If this were product forum for almost anything else that post would have been immediately deleted. Next one like it will be.

Come on James, the appraisal by obcd was pretty accurate and fair . Does your job at Broadcom depend on wringing every last sale out of the Pi ?

You've been on my case for perhaps overdwelling on some of the Pi's issues (USB). I'm still here because I give out some useful advice (I have it on good authority) ,does obcd fall into that category as well ?

The Pi has faults, but it didn't stop me buying 3 of 'em and I'll be first in line for an 'A' and a camera module when they become available.

If your job is on the line, then you have a case and I'm out of line, but I think you should declare it.

It wasn't an appraisal, it was an assassination job. He made no effort to put any good points in his list, and some of his points are wrong or out of date anyway.

So I'll do it since its 11.20pm, and I'm still in the office waiting for customers to phone.


It doesn't need an extremely accurate voltage supply - it needs an adequate one.
When the voltage between tp1/2 drops below 4.85v the device stops working. Yes, because you are not giving it enough power - why is that an issue? It's statement, a bit like saying "My car bloody well stopped. Just because I haven't got any petrol in it. It's an outrage I tell you"
Yes, webcams are problematic - I agree with that one. USB fixes on the way we hope.
FTSI chipsets - see the thread, but getting much better.
Keyboards do not attempt to lose characters. With the right settings they mostly work fine,
The usb port supply pins are directly connected to the 5V supply. - and what, pray, is the problem with that - again a simple statement of fact made to look like its a real problem.
Hotpluggin can cause problems -agreed
The Foundation has ALWAYS, from before release, stated that the analog audio out is a bit pants. It was never meant to be high quality. Sorry.
"The os normaly boots from an sd card. This requires to always properly shutdown the system, as the sd card background wear levelling algoritms can corrupt the sd card when power is removed during a write operation to the card. Making the root fs read only might prevent such writes. Most oses also use a swap file. Due to the limited write cycles of sd card flash technology, this might shorten the sd card life." No empirical evidence that this is the case at all. I've never had problems with SD cards, asnfd I've never shut the thing down properly. Most people will never see a problem with wearing them out.
"The system comes without real time clock and requires synchronisation with a time server to acquire the correct time. " Well, duh! That's just a statement of fact designed to make the product look bad
We have our own distro. The fact that other distro's don't support Armv6 is completely irrelevant. There is one for the Raspi, and it works fine.
See teh_orphs thread about GPU accleration for X - currently in testing.
Foundation having one employee is again irrelevent - people ARE working on fixing stuff, and ARE improving things all the time - X and USB are cases in point.
"They already sold approx 1000000 pieces, so it looks like for most people those minor issues are not a problem." TRUE!!


Hmm. That wasted 10 minutes. Still no customer call. Think I might bunk back to the hotel.

[jamesh]

I don't quite like the "you get what you pay for" approach. You should get what you were promised, and that is the case for most buyers. It is not ok that some people are getting defective units and they should have them replaced for ones that work properly through the distributor/manufacturer

......clip.....

Anyway, the point is, the faulty units are not the norm and you have every right to have them replaced.

Agreed entirely - and this is I believe what is happening - not sure where the idea that defective boards are not being replaced has come from.

[ShiftPlusOne]

I don't quite like the "you get what you pay for" approach. You should get what you were promised, and that is the case for most buyers. It is not ok that some people are getting defective units and they should have them replaced for ones that work properly through the distributor/manufacturer

......clip.....

Anyway, the point is, the faulty units are not the norm and you have every right to have them replaced.

Agreed entirely - and this is I believe what is happening - not sure where the idea that defective boards are not being replaced has come from.

I wasn't trying to say that the distributors are not replacing them, but that some customers aren't. In some of the other threads a guy has mentioned that he needs to aim a fan at his pi to prevent it from shutting down. A few have said that their pi just doesn't stay powered for long in general (with power supply issue eliminated). Some people even reflow their pi in the oven with some success. It seems that some people are just not having them replaced when maybe they should.

jamesh = moderator

ShiftPlusOne = moderator

Go sort it out in a PM

Or as they used to say "Get a room"

It wasn't an appraisal, it was an assassination job. He made no effort to put any good points in his list, and some of his points are wrong or out of date anyway.

You seemed determined to read it as an assassination job, I read it as largely a statement of fact. Most of it was just a list of the Pis 'features' in light of his opening statement 'You get what you pay for'. Its a cheap computer, and this is what you get. A guy who has been around for 6 months and made 500+ largely helpful posts obviously doesn't think the Pi is a lost cause. I read the final 2 sentences as quite positive. You shouldn't be so defensive......

Some of the issues you list are software related and have nothing to do with the Pi hardware, like GPU acceleration and X not supporting that. No one is forcing Linux upon you

As, you said though, you get what you pay for. But, to me $40 is an awesome price for a device that is so versatile. You can buy cheap and nasty DVD players for that price, but what can you do with them? Only play DVD's, thats it....

What can you do with a Pi?

There are more than likely undiscovered flaws as well, but since they are undiscovered after close on a million sold, they cannot be that bad...

One cannot usually tell how bad flaws are until they are discovered or, more correctly, start affecting things.

A million sold doesn't mean there isn't an undiscovered flaw which could be severely damaging when it does reveal itself in the future.

All non-trivial products present such concerns; so I'm not criticising the R-Pi, just the notion that quantity sold and/or current lack of evidence of flaws means there isn't a flaw or any flaw will not be serious. It doesn't.

I disagree. How can a bad flaw suddenly appear, and affect the whole million?

I never said any flaw would affect the whole million; "One cannot usually tell how bad flaws are until they are discovered or, more correctly, start affecting things".

I stand by that. For the LAN / REG power short in version 1 some people were unaffected, and few, if any, seem to have been seriously affected by it, but that wasn't by design, that was luck, dodged a bullet. Instead of being noticed when it was it could have remained hidden for months and maybe only bitten when a million had been sold. When version 1 went on sale no one could have guessed at that design flaw, nor how bad it would be. Only when the issue was revealed could the likely impact be assessed.

If there were a bad flaw, then surely it would have appeared by now? The only possible thing I could think of was some aging problem with a component, but they have all been thoroughly tested. I can see a bad flaw appearing in a very minor subsystem, used by very few people (or again, it would have appeared by now)

Well we already have the GPIO47 flaw and there may be others that haven't yet been spotted. It's not apparently caused anything bad so far but suppose someone makes a typo in the bootloader, GPU code, Raspbian, some driver or commonly used app that is only triggered some time in the future or through some future change in something else; one day, bang, many Pi's could have their GPIO ports blown and be rendered useless.

Can anyone guarantee that there is not and never will be such an event waiting to bite ? If such an event could happen having sold ten billion won't make it not happen.

Maybe someone could add such a thing maliciously, sit back, wait for the ticking time bomb to go off. Can anyone guarantee that they haven't and won't ?

This of course is exploiting a known flaw, but no one knows if there isn't some other flaw that hasn't caused problems so far but can in the future. Just because we haven't seen it, can't imagine there would be, doesn't mean it isn't there. That's the nasty nature of design flaws; if we could pre-empt them we'd never have any.

Just to reiterate; I'm not suggesting the Pi is riddled with flaws which will become catastrophic. I'm just saying no one can definitively say there are no flaws based on numbers sold. As is the case with every other non-trivial product out there.

[jamesh]

I disagree. How can a bad flaw suddenly appear, and affect the whole million?

I never said any flaw would affect the whole million; "One cannot usually tell how bad flaws are until they are discovered or, more correctly, start affecting things".

I stand by that. For the LAN / REG power short in version 1 some people were unaffected, and few, if any, seem to have been seriously affected by it, but that wasn't by design, that was luck, dodged a bullet. Instead of being noticed when it was it could have remained hidden for months and maybe only bitten when a million had been sold. When version 1 went on sale no one could have guessed at that design flaw, nor how bad it would be. Only when the issue was revealed could the likely impact be assessed.

If there were a bad flaw, then surely it would have appeared by now? The only possible thing I could think of was some aging problem with a component, but they have all been thoroughly tested. I can see a bad flaw appearing in a very minor subsystem, used by very few people (or again, it would have appeared by now)

Well we already have the GPIO47 flaw and there may be others that haven't yet been spotted. It's not apparently caused anything bad so far but suppose someone makes a typo in the bootloader, GPU code, Raspbian, some driver or commonly used app that is only triggered some time in the future or through some future change in something else; one day, bang, many Pi's could have their GPIO ports blown and be rendered useless.

Can anyone guarantee that there is not and never will be such an event waiting to bite ? If such an event could happen having sold ten billion won't make it not happen.

Maybe someone could add such a thing maliciously, sit back, wait for the ticking time bomb to go off. Can anyone guarantee that they haven't and won't ?

This of course is exploiting a known flaw, but no one knows if there isn't some other flaw that hasn't caused problems so far but can in the future. Just because we haven't seen it, can't imagine there would be, doesn't mean it isn't there. That's the nasty nature of design flaws; if we could pre-empt them we'd never have any.

Just to reiterate; I'm not suggesting the Pi is riddled with flaws which will become catastrophic. I'm just saying no one can definitively say there are no flaws based on numbers sold. As is the case with every other non-trivial product out there.

I didn't saw there are no flaws, I said (or maybe implied) that as the numbers sold increases, the chances of a 'serious' or indeed any flaw decrease. Case study. I'm currently working on something. This something has been working fine in all testing so far. Just started in pre-production, and there is an immediate production stop because of a flaw. It's a really obscure flaw, that simply didn't show up in normal testing, but once hundreds of devices started being thoroughly tested its come to light. BUT, it's only taken a few hundred devices to show this really obscure flaw. Once you get in to the millions, the chances of a flaw NOT being discovered decrease dramatically.

Once you get in to the millions, the chances of a flaw NOT being discovered decrease dramatically.

But a flaw which only manifests itself in time or after a future event will not be discovered until that time no matter how many are sold before that time, one or ten billion. I suspect we are arguing two different things, not against the other, and, as it's an aside to what my original point was, I'll leave it there.

Bottom line. If you think you have a hardware issue, put in a warranty claim.

[mahjongg]

Regardless the issues described here did not warrant the inflammatory subject line of the post which Tails initially used (before moderation).
"Rapsberry PI is a Lemon".
On many other fora such behaviour would have tails an immediate ban for trolling, just for using that subject line!

To be sure -any- product, no matter what can have hidden fault, it simply a given.
If anybody is interested in the subject, it helps to start reading here: http://en.wikipedia.org/wiki/Reliability_engineering

Be fair mahjong re-read the thread title His Raspberry Pi is a Lemon not all RaspberryPi's.

[mahjongg]

Be fair mahjong re-read the thread title His Raspberry Pi is a Lemon not all RaspberryPi's.

Did you read what I said?
He started this thread with the subject "Raspberry PI is a lemon", as in all raspberry PI's are lemons! Even when his message didn't fit the bill! That is why he wasn't discounted as a simple troll.

In any production there is a typical failure rate, even after post production testing, and in the industry a 1% failure rate is about the norm, that means that any device of which half a million is made about 5000 faulty ones typically will find their way to customers. Thats completely normal!
It seems that the Chinese manufacturers are normal producers of small electronic devices, and so its statistically normal for a few thousand units to fail! This is with testing mind you! The failures that are immediately obvious will be eliminated, and won't make it out the door, but any problem that takes 5 to 10 minutes to reveal won't be found. But that is what guarantee is for, simply return it!

Just received my Raspberry Pi from Allied Electronics after months of waiting and it's a lemon.

Just re-read the first post and thread title.

Sorry for being pedantic.

Or did you change them?

[mahjongg]

Nothing in the OP's message was edited (that wouldn't be ethical), but the subject title was changed to be less inflammatory, and a better description of the actual post.
it was changed from "Raspberry Pi is a Lemon" to "My raspberry Pi is a Lemon".

The original title (amongst other things) caused RobHenry to remark that "I've checked mine and it's not a lemon, it's a raspberry.", and is probably the cause of different ideas about the nature of the post by people who responded to the original subject line, versus people who came in later.

What happened on the thread where the OP said his Pi arived with the SoC missing - did he ever post photos?

I searched but could not find the thread. Last time I read it he was asked and said he would post photos more than once but hadn't.

Were any photos ever produced?

Just wondered because I would like to see what the board looks like under the chip..

[ShiftPlusOne]

What happened on the thread where the OP said his Pi arived with the SoC missing - did he ever post photos?

I searched but could not find the thread. Last time I read it he was asked and said he would post photos more than once but hadn't.

Were any photos ever produced?

Just wondered because I would like to see what the board looks like under the chip..

Not sure what thread you're talking about, but you can see unpopulated boards here.
http://www.raspberrypi.org/archives/402

Unless I dreamed it there was a geezer who got into dialogue with various mods about his Pi turning up with the most important bit missing.

Was wondering whether he was a troll or legit.

Thanks for the pic link - it looks cool - I think the foundation should use that for a t-shirt design, or maybe just the silk-screen or photo of the front and back of the board on the front and back of the shirt.