MaxK1
Posts: 1043
Joined: Sun Aug 26, 2012 11:34 pm

Re: USB - the Elephant in our Room

Tue Nov 20, 2012 6:29 pm

tufty wrote:
MaxK1 wrote:And judging from some of the exchanges, there are more than a few people out there who are declaring that they *know* their cheap, POS "charger" is good because it says 5V, 1.+A right on the printed label. Don't need no stinkin' meter to check that....
And there are others who have stabilised lab supplies on their desks, who /do/ know which end of a multimeter is which, and who are /still/ having problems. With keyboards.

Last time I checked, as I said, voltages were well within usable range, the polyfuses were fine, and I was having problems. I'll bridge the fscking polyfuses if it will satisfy you, or direct feed the power via a hacked up usb cable, but I can almost guarante it won't change anything for the better (I tried the latter before, and it didn't fix the problem then, I can't see why it would do so now).

Like I said before, I frankly couldn't care less about linux on the pi, but there's 3 things here that are liable to turn people less savvy than myself off, /instantly/ :

1 - keyboard that works elsewhere doesn't work on the pi => "the device is broken"
2 - arcane shit to try and make it work => "linux is a sack of wank"
3 - attitudes like "it's a power problem" or " buy another keyboard" => "the community is stuffed with arseholes"
I'm not saying there aren't problems with some devices. I am saying that we need to separate the true problem devices from the power supply issues. Sounds like you have been unlucky enough to get hold of a bunch of keyboards that fall into the former category and you are not willing to go any further with it. I'm OK with that. Maybe someone else who is reading this can benefit from it. There have been several (no idea how many) who _have_ found that switching to a different supply solved the problems they were having. Shrug.
You are in a maze of twisty little passages, all alike.
When General Failure and Major Disaster get together, Private Parts usually suffers.

ski522
Posts: 394
Joined: Sun Sep 30, 2012 2:22 pm

Re: USB - the Elephant in our Room

Tue Nov 20, 2012 6:56 pm

tufty wrote: I'm not saying there aren't problems with some devices. I am saying that we need to separate the true problem devices from the power supply issues. Sounds like you have been unlucky enough to get hold of a bunch of keyboards that fall into the former category and you are not willing to go any further with it. I'm OK with that. Maybe someone else who is reading this can benefit from it. There have been several (no idea how many) who _have_ found that switching to a different supply solved the problems they were having. Shrug.
After reading through this thread I'm under the general impression that software work was need for USB. An update here would be nice, not sure where the dev twitter's feed is?

As for PS's not sure why we need to keep bringing that up. As an EE myself, I'm having USB problems with brand name components and I know it's not PS related.

tufty
Posts: 1456
Joined: Sun Sep 11, 2011 2:32 pm

Re: USB - the Elephant in our Room

Tue Nov 20, 2012 7:49 pm

MaxK1, not tufty wrote: I'm not saying there aren't problems with some devices. I am saying that we need to separate the true problem devices from the power supply issues. Sounds like you have been unlucky enough to get hold of a bunch of keyboards that fall into the former category and you are not willing to go any further with it. I'm OK with that. Maybe someone else who is reading this can benefit from it. There have been several (no idea how many) who _have_ found that switching to a different supply solved the problems they were having. Shrug.
And there you have it. "Shrug". Way to be helpful.

Although I don't care about Linux on the Pi personally, I do care about the success of the Pi generally, and that is intrinsically tied to Linux. It's hard enough to get a linux box of any kind into educational establishments, where Windows is king and the only thing the admins know is Windows; if the device itself is "problematic" with known good peripherals (*every* keyboard I own works perfectly with every other computing device I have, they simply are not at fault) and the solution suggested is "replace the keyboard" or "replace the PSU" as a first guess, it ain't gonna go *anywhere*. For me to replace a PSU (not that I need to, my PSUs are absolutely adequate) will maybe set me back 20 euros. Yeah, "first world problem". However, if you're looking at a classroom full of 30+ machines when budgets are tight, that's closer to 500 euros even with discounts and still no guarantee that it's all gonna work. Oh, changing the PSUs didn't work? Swap out the keyboards. Another 5 - 15 euros per machine, and *still* no guarantee. Plus we now have 30-odd redundant PSUs and keyboards on our hands. Okay, so your technician now has to spend hours piss-anting about configurations and rpi-update and so on, - "hell, let's just send the ******* lot back and get some PCs running Windows - at least we understand them and they work out of the box".

That's a pretty reasonable scenario of how the Pi /isn't/ going to get a foothold, and it's before student 1 starts getting blown away by how linux sorta kinda works a bit sometimes and "looks a bit wierd and it's slow".
ski522 wrote:As for PS's not sure why we need to keep bringing that up. As an EE myself, I'm having USB problems with brand name components and I know it's not PS related.
Because it's the standard, simple solution to everything. If it doesn't work, it's because your PSU isn't up to scratch. If it still doesn't work after that, it's because you haven't got the right type of USB peripheral. Yes, the "U" in "USB" does stand for "Universal", but not when we're talking about the Pi.

None of this, of course, is intended to denigrate what Gordon and the others are actually doing, which is actual valuable and useful work. If I was near Cambridge, or even in the UK, I'd pop round and deliver my Pi, keyboard, cabling and PSU to Gordon in person as an example of a setup that refuses to work. But if another "guru" suggests that my PSU, which can source enough current to vapourise most of the Pi's components, is somehow "not powerful enough" to power a keyboard that draws well under 25mA - well - "keyboard enema" :)

I should mention that the apple white keyboard, at least, *has* worked in the past on my Pi. But it doesn't any more. All that's changed is the software.

MaxK1
Posts: 1043
Joined: Sun Aug 26, 2012 11:34 pm

Re: USB - the Elephant in our Room

Tue Nov 20, 2012 9:03 pm

"I should mention that the apple white keyboard, at least, *has* worked in the past on my Pi. But it doesn't any more. All that's changed is the software."

Hmm - it would have been useful to have known that - I would *not* have suggested the PSU/cabling
was possibly at fault, but maybe suggested something else. When you say the software changed, I assume you mean the firmware files rather than an older kernel (3.1.?) or are you referring to a change in the USB driver and it suddenly broke with one of Gordons changes?
You are in a maze of twisty little passages, all alike.
When General Failure and Major Disaster get together, Private Parts usually suffers.

davidmam
Posts: 101
Joined: Tue Dec 06, 2011 4:13 pm

Re: USB - the Elephant in our Room

Tue Nov 20, 2012 10:58 pm

Is there any update on the issue with libusb where repeated acquisitions of a device fail? Ie you can send one request to a camera with gphoto2 but the second will hang the whole USB system, and the same I see with my data logger. Acquire once, fine, acquire twice (consecutively, not simultaneously) and the USB disappears. It can be prevented by doing a USB reset between acquisitions

It really is entirely reproducible.
A reproducible show stopper for me.

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

Re: USB - the Elephant in our Room

Wed Nov 21, 2012 4:55 am

I'd just like to jump in real quick and point out that Tufty is *EXTREMELY* intelligent on these things and if you take a look at his post count you can probably realize that one can forgo the usual "have you tried swapping the PSU yet" ;-)

While power issues are a very common (though thankfully decreasingly so) root cause for USB issues, there's no need to keep harping on it ignoring other more likely candidates :-)

I have seen several threads from the launch till now indicating that there does seem to be reoccurring issues with Apple keyboards not working with the Pi. While I don't find this overly surprising as Apple has a nasty habit of subtly changing standards to make their devices subtle proprietary (I'm looking at you USB charging), since Apple is so annoying prolific (no offense Simon :-P) the pi should be more readily compatible with apple keyboards as those are likely to be on hand.
Dear forum: Play nice ;-)

tufty
Posts: 1456
Joined: Sun Sep 11, 2011 2:32 pm

Re: USB - the Elephant in our Room

Wed Nov 21, 2012 11:34 am

abishur wrote:I'd just like to jump in real quick and point out that Tufty is *EXTREMELY* intelligent on these things and if you take a look at his post count you can probably realize that one can forgo the usual "have you tried swapping the PSU yet" ;-)
Awww, shucks <blush>.

Post counts don't necessarily signify anything though :)

abishur wrote:I have seen several threads from the launch till now indicating that there does seem to be reoccurring issues with Apple keyboards not working with the Pi. While I don't find this overly surprising as Apple has a nasty habit of subtly changing standards to make their devices subtle proprietary (I'm looking at you USB charging), since Apple is so annoying prolific (no offense Simon :-P) the pi should be more readily compatible with apple keyboards as those are likely to be on hand.
It's an interesting thought, but I really don't think that's the problem, no matter what you might think of Apple. For starters, I'm seeing this on more than just Apple keyboards, particularly a Matias tactile pro, and various usb firmwares running on "teensy" microcontrollers. I could accept that the teensy firmwares might play "fast and loose" with the USB HID spec, but the Matias certainly doesn't (it's an early board, later ones do some weird stuff to get full NKRO).

Secondly, if it were a problem with the way that Apple boards interact with the Linux (or other) USB stack, I'd probably be seeing the same problems on other machines (like, for example, the Dell running kubuntu 12.04 that I'm typing this post on, using the same Apple white keyboard that spectacularly fails on the Pi).

It's all very odd. Some keyboards, and almost always those with onboard hubs, seem to trigger interrupt storms on the Pi, when combined with certain combinations of firmware and / or kernel. My guess is that it's kernel related (there were reports that Apple boards worked with debian but failed miserably under Arch - or vice versa, I forget).

I really think we should put the "swap your PSU" thing to sleep, though. The white Apple boards have been measured as drawing under 25mA when in use. I've measured my Matias at 15mA.

gsh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 1436
Joined: Sat Sep 10, 2011 11:43 am

Re: USB - the Elephant in our Room

Wed Nov 21, 2012 12:05 pm

I assume that if you're powering it from a bench supply then you have removed the fuses? Just measuring between the two test points only tells you what the quiescent voltage drop is across the fuse, we don't know if there is a short term spike in power consumption leading to a short term power drop.

The point here is that there seems to be a difference between your hardware and mine and although software looks to be the problem, there is significant evidence to believe it's not the keyboard or the software but maybe the hardware that's the problem, we should try to fix down where the problem is coming from.

I've not noticed a key drop on my system for about two months (since I enabled the fiq and we removed the big latencies in the SD driver). Although I did put a bluetooth device into my RaspBMC build and found I dropped a key the other day. (too difficult to reproduce to be useful though)

Not been able to look at isochronous at all because I don't have a suitable device I've only seen USB1.0 devices and that's just going to fail for all the split transaction reasons.

EDIT:

OK, having bothered to go back and actually read some posts then I think I'll have a look at an apple keyboard if I can get my hands on one...

I've now removed the fuses from my board though and don't know whether that's effecting stuff...

Gordon
--
Gordon Hollingworth PhD
Raspberry Pi - Director of Software Engineering

User avatar
RaTTuS
Posts: 10458
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK

Re: USB - the Elephant in our Room

Wed Nov 21, 2012 12:13 pm

tufty wrote:... Some keyboards, and almost always those with onboard hubs, seem to trigger interrupt storms on the Pi, when combined with certain combinations of firmware and / or kernel. ...
I've seen this with Dell keyboads with USB hubs in them.
How To ask Questions :- http://www.catb.org/esr/faqs/smart-questions.html
WARNING - some parts of this post may be erroneous YMMV

1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX
Covfefe

thradtke
Posts: 492
Joined: Wed May 16, 2012 5:16 am
Location: Germany / EL

Re: USB - the Elephant in our Room

Wed Nov 21, 2012 12:21 pm

gsh wrote:[...] I've only seen USB1.0 devices and that's just going to fail for all the split transaction reasons.
What are those "split transaction reasons"? Do we have a chance to eliminate those problems, too?
Rocket Scientist.

Triode
Posts: 4
Joined: Mon Nov 19, 2012 11:18 pm

Re: USB - the Elephant in our Room

Wed Nov 21, 2012 7:33 pm

gsh wrote: Not been able to look at isochronous at all because I don't have a suitable device I've only seen USB1.0 devices and that's just going to fail for all the split transaction reasons.
Hi Gordon - can you comment on this further or point me to more detailed discussion - I'd be interested in understanding more and providing more test cases if I can.

Are split transactions not supported at all, or is it just any embedded transaction translator doesn't work correctly. I've found problems on other SoC for my use case which are resolved by using an external high speed hub as this enables the hub's transaction translator to be used (in this case its the ARC usb core on other SoCs which doesn't seem to support in and outbound SiTDs in the same 1ms frame..). Is it possible to disable any internal TT and force use of an external one? Does the normal linux ehci-sched code get used?

My use case is a 1.0 device hifi dac which works in asynchronous mode. This creates 2 isochronous streams - the output stream which needs to reserve space in every 1ms frame, and a second inbound one which also needs space in every 32 frames. I think you are saying this is not expected to work at present. Is there any help we can give to help resolve?

GT-Force
Posts: 4
Joined: Wed Nov 21, 2012 6:20 am

Re: USB - the Elephant in our Room

Wed Nov 21, 2012 11:21 pm

I am one of the displeased RaspberryPi customers/owners. I thought this was going to be the little magic media-center for $35, but it doesn't even work for more than 10 mins due to all USB devices failing. Utter failure! Didn't ANYONE test this thing before it was pushed into production? I am very disappointed, and telling people that I am so. :(

Frodo1B
Posts: 18
Joined: Sat Nov 03, 2012 12:09 am

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 12:19 am

I am strongly suspect that instances like this are due to either bad Raspberry Pi boards, or bad SDHC cards. It seems that once one finds correctly functioning SDHC cards along with a good power supply things work correctly. The problem is that I have seen batches of 3 new SDHC cards that did not function correctly. Once I found good cards that were compatible with my RPis everything smoothed out. This SDHC card issues makes it hard to recommend the RPi to a non-technical user though as it is impossible to know that they have good compatible SDHC cards and I have seen cases where cards on the compatibility list only work in less than 50% of the cases.

GT-Force
Posts: 4
Joined: Wed Nov 21, 2012 6:20 am

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 12:38 am

According to these guys, it's neither SDHC cards, nor the power supply but only RapberryPi itself to blame:

http://forum.stmlabs.com/showthread.php?tid=4271&page=3

I measured mine after finding this thread, and to no surprise, I also get very low voltages.

I have neither time, nor will to exchange boards until I hit a good one. From my perspective, This is a huge opportunity wasted with a bad design. I wish it was as advertised, but looks like it just isn't.

User avatar
rurwin
Forum Moderator
Forum Moderator
Posts: 4258
Joined: Mon Jan 09, 2012 3:16 pm
Contact: Website

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 9:07 am

It's a development release.

Bridge that fuse and put one in the power cable. Report your findings.

gsh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 1436
Joined: Sat Sep 10, 2011 11:43 am

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 9:20 am

Split transactions are supported, (otherwise keyboards wouldn't work at all on model B!)

If you absolutely have to have USB1.1 working then the best way is to force the speed to 1.1 using dwc_otg.speed=1 this means that everything works at that speed and therefore the interrupt overheads are significantly reduced meaning that the transaction timing is more acurate.

There have been some reports of things not working at all in this mode (i.e. the whole thing crashing!) but I've not seen this...

The main problem with split transactions is that the timing is critical, if you send a packet one uframe too early the hub will drop it, if you send it late (past the end of the frame) then it'll drop it. Unfortunately when you couple this with a hardware implementation that uses interrupts to decide when to send this and couple it with a poor interrupt latency operating system like linux then you end up with us dropping packets. This isn't a problem with bulk transactions of course because they'll get retransmitted but interrupt and isochronous will actually lose packets.

The solution to this problem is basically re-writing the driver (no it is not an ehci driver) I've had an initial stab at creating an EHCI compatible stub, and I've got it to the point where I can do simple device enumeration but it requires that people help me now and that means I need to get the documentation released and am in the process of organising that now... Keep your fingers crossed, the company who owns this documentation is not famous for giving it out!

Gordon
--
Gordon Hollingworth PhD
Raspberry Pi - Director of Software Engineering

thradtke
Posts: 492
Joined: Wed May 16, 2012 5:16 am
Location: Germany / EL

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 9:26 am

Again thanks, Gordon, for your explanations. I wish I could help, but it's all way over my head :-(. Good luck and thank your for caring.
Rocket Scientist.

tufty
Posts: 1456
Joined: Sun Sep 11, 2011 2:32 pm

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 9:34 am

GT-Force wrote:According to these guys, it's neither SDHC cards, nor the power supply but only RapberryPi itself to blame:

http://forum.stmlabs.com/showthread.php?tid=4271&page=3

I measured mine after finding this thread, and to no surprise, I also get very low voltages.
What they are talking about there is voltage drop across the F3 polyfuse. This is a self-resetting fuse that runs between the PSU and the rest of the circuit. A polyfuse is a bit like a "classic" fuse, except that when it blows, it resets itself afterwards. Sounds great, right? It is, except that polyfuses have a non-negligible resistance to start with (so you will see a certain amount of voltage drop across them even in a normal situation), and once they have blown once, they have a tendency to increase in resistance (and thus voltage drop) even in the normal case. They also have a non-negligible "reset time", if you blow the fuse once, and then retry straight away, you're liable to make things worse. It is possible (and appears to have been the case in the thread you linked to) that excessive polyfuse resistance can cause voltages seen by USB devices to be well below the spec, and cause problems at the USB device end (anything less than 4.75 volts at the device, like the 4.5 volts measured in the linked thread, is out of range).

There's a few issues that can cause "out of range" voltages on the Pi.

The first is the PSU. If it's too weedy, then you will see voltage drops as the processor or any pi-powered peripheral "wakes up" and pulls more juice. Obviously, the problem here is the PSU, and the solution is obvious. As a hint, you should generally look at your PSU, take the rated supply current, and divide it by the number of outlets, because manufacturers are twats; if you have a PSU that claims to supply 2A, but it has 4 USB outlets, chances are it will supply 500mA maximum per port, and that's not going to be enough for a Pi with SD card, keyboard and mouse. The only really safe option is to take a brand name PSU that's intended for something like an iPad, and not some $0.99 no-name chinese wall wart from eBay.

Secondly, there's the F3 polyfuse, which might be out of range (either manufacturing flaw, or due to having been tripped heavily in the past). This is measurable using the method described in the thread you linked to - if it's OOR then you have a decent chance at getting your Pi replaced - it's defective.

Thirdly, on "first batch" Pis, there were also F1 and F2 polyfuses on the USB ports themselves (other side of the board to F3, next to the USB ports). These have been removed / replaced with 0Ohm resistors on later batches. If you have them (they should be green SMDs like F3, probably marked "14", the replacement resistors will probably be black and will definitely be marked "0"), the measurement process is exactly the same as that for F3. Again, OOR -> defective Pi.

How do polyfuses blow? By putting devices on the USB port that try to draw too much current. Anything more than 100mA is too much. A good example would be a USB-powered hard drive. Putting one of these on your Pi will probably fry the polyfuses straight away. The Pi has /nothing/ to stop a device trying to pull track-melting currents from USB apart from the polyfuses. Many, if not all, USB WiFi dongles will pull more than 100mA...
GT-Force wrote:I have neither time, nor will to exchange boards until I hit a good one. From my perspective, This is a huge opportunity wasted with a bad design. I wish it was as advertised, but looks like it just isn't.
I can sympathise with this, but it should probably be pointed out that the Pi was never advertised as a magical $35 media centre.

FWIW, Gordon, I just pulled out the multimeter and re-measured my Pi. Powered off the USB charger, 0.02V drop over F3 and 0.04 / 0.03V over F1 and F2 respectively. Worst case, I'm seeing 4.92V at the device (My Matias is currently gutted, and I have access to the incoming USB lines...)

Simon

User avatar
mahjongg
Forum Moderator
Forum Moderator
Posts: 12214
Joined: Sun Mar 11, 2012 12:19 am
Location: South Holland, The Netherlands

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 9:43 am

Yes, It seems some people have a polyfuse that has "blown" and has not completely recovered.

By the way, normal fuses have a non negligible resistance also, otherwise they couldn't blow either. (yes, D'oh! :lol: )

Why the fuse has been blown is the big question, it could happen when putting in more than 5.6 Volt into the PI, as it is at that voltage that the PI's overvoltage device D17 trips, and then drains a lot of current through the poyfuse, thereby fully or partly "blowing it".

One reason, (Im theorizing here) could be a charger that when unloaded has an abnormally high output voltage, say six volts, and has a build in large capacitor (a few hundred microfarad) that holds a charge at this six volt, then when you plug in this charger (connect the micro-B USB connector) D17 triggers, and the capacitor in the PSU discharges through the polyfuse, thereby increasing its resistance significantly.

I still feel that F3 is dimensioned too small, off-factory these fuses seem to all have a resistance way above their "average" spec.

RPF people are aware (obviously, I might add) of the issue, (of "bad" polyfuses, whatever the cause behind it is) and are investigating it, and will most probably solve (or at least largely mitigate) it in the near future.

tufty
Posts: 1456
Joined: Sun Sep 11, 2011 2:32 pm

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 10:06 am

mahjongg wrote:By the way, normal fuses have a non negligible resistance also, otherwise they couldn't blow either.
Of course. D'oh.
mahjongg wrote:One reason, (Im theorizing here) could be a charger that when unloaded has an abnormally high output voltage, say six volts, and has a build in large capacitor (a few hundred microfarad) that holds a charge at this six volt, then when you plug in this charger (connect the micro-B USB connector) D17 triggers, and the capacitor in the PSU discharges through the polyfuse, thereby increasing its resistance significantly.
Sounds quite possible to me. Having seen the guts of a few cheapo chargers, it wouldn't surprise me to see 5V +- 240V being dumped into the Pi :)

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 23636
Joined: Sat Jul 30, 2011 7:41 pm

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 10:21 am

GT-Force wrote:I am one of the displeased RaspberryPi customers/owners. I thought this was going to be the little magic media-center for $35, but it doesn't even work for more than 10 mins due to all USB devices failing. Utter failure! Didn't ANYONE test this thing before it was pushed into production? I am very disappointed, and telling people that I am so. :(
Well done. Thanks for 'dissing' this educational product that doesn't do exactly what you want (yet), even though exactly what you want isn't what it's being sold as. Please be aware that slagging of a charity led product designed for education is rather an unpleasant and callous thing to do. For something that cost you $35. Many people out there have no problems with USB at all, so for a start your facts are wrong (did you really do any research on the topic?). How you can say utter failure when it obviously isn't for the vast majority of people is rather self centered don't you think? Of course it was tested, and the software has been improving all the time. So, all in all, troll fail.

To those others who actually think about the problem and help find issues, and for testing fixes etc, without feeling the need to tell all and sundry how terrible your Pi life is - thank you.
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Contrary to popular belief, humorous signatures are allowed. Here's an example...
"My grief counseller just died, luckily, he was so good, I didn't care."

gsh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 1436
Joined: Sat Sep 10, 2011 11:43 am

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 10:52 am

Used Apple keyboard on RaspBMC and got zero key drops...

version: XBMC 12.0-ALPHA Git:bdff099 (Compiled: Nov 10 2012)
Keyboard model number: A1048

Also mouse plugged into keyboard worked fine too...

Can you try this build and make sure that you still see the same problem with the Apple keyboard (my keyboard is quite old so may not be the same version as yours). What version is the keyboard?

Gordon
--
Gordon Hollingworth PhD
Raspberry Pi - Director of Software Engineering

ski522
Posts: 394
Joined: Sun Sep 30, 2012 2:22 pm

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 12:47 pm

jamesh wrote:Well done. Thanks for 'dissing' this educational product that doesn't do exactly what you want (yet), even though exactly what you want isn't what it's being sold as.
Then they need to change the FAQ
What’s a Raspberry Pi?

The Raspberry Pi is a credit-card sized computer that plugs into your TV and a keyboard. It’s a capable little PC which can be used for many of the things that your desktop PC does, like spreadsheets, word-processing and games...
This one paragraph is also used on many vendors websites selling the PI. This leaves many with the impression that the PI can do everything a computer can do including using all kinds of USB devices which is going to happen. The problem with USB is the first word "Universal" which (at the moment) appears not to be the case...that's fine, but it needs to be clear on the PI's website (and maybe at vendors websites too) that the USB port on the PI is quite as "Universal" (yet) that many would like it to be.

rspitz
Posts: 58
Joined: Tue Jul 31, 2012 7:25 pm

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 1:38 pm

Jamesh,

no need to take such criticism personal. And trying to excuse the Pi's braindead USB implementation with its status as "charity led product designed for education" is completely beside the point, too.

Fact is: The Pi is sold as a commercial product via commercial channels, and as such deserves to be treated as such.

As for the "charity" part: When a price was originally calculated for a total sale of 10.000 units and now the number of units sold is climbing towards one million, there must be a margin that somebody is making a nice profit of. I'm pretty sure it is not the Foundation, but whoever is making that profit should take responsibility for improving the known deficiencies, if that is at all possible with the current hardware.

Apparently, this is not being done, so the problem is still only being worked on by dedicated volunteers such as Gordon, who has my greatest admiration for trying to tackle this Sisyphos task.

However, having followed this discussion from the beginning, I'm starting to believe that the USB case is hopeless. The designers of the Pi were mislead by the alleged "USB host" capability of the SoC, which is just badly designed and has limitations that were apparently never really tested or challenged before the SoC was used in the Pi and went into widespread use. Well, shit happens and the Pi might still be mostly sufficient when used just with a keyboard and mouse attached for "educational programming". But the official Pi homepage is still full of "success stories", even for _commercial_ projects, without any prominent warning that there are USB problems. So it is no surprise that there are buyers who are disappointed (to say the least) with the product when they find out the hard way that the USB ports don't work as expected.

Regards, Richard

MaxK1
Posts: 1043
Joined: Sun Aug 26, 2012 11:34 pm

Re: USB - the Elephant in our Room

Thu Nov 22, 2012 2:12 pm

Not sure how you got from this:

"It’s a capable little PC which can be used for many of the things that your desktop PC does, like spreadsheets, word-processing and games..."

to this:

"This leaves many with the impression that the PI can do everything a computer can do including using all kinds of USB devices which is going to happen"


I would infer from the from the 1st blurb that it could do many things, and it was up to creative minds to find out just how far it could be pushed.
You are in a maze of twisty little passages, all alike.
When General Failure and Major Disaster get together, Private Parts usually suffers.

Return to “Troubleshooting”