User avatar
jackokring
Posts: 816
Joined: Tue Jul 31, 2012 8:27 am
Location: London, UK
Contact: ICQ

Re: Minor improvements for a model C

Fri Aug 10, 2012 2:31 am

http://www.raspberrypi.org/phpBB3/viewt ... 46&t=14165 memory topology. Not really worth it given the nature of cache and size. Now a split between super fast big memory, and slower small memory but more dense memory. And the SD swap thrash stall is severe.

Yes. A 1GB or less high speed flash, with an SRAM page restoration eviction cache on it. It would prevent early onset of the low memory swap thrash best (especially if only swapfs could use it).

Cheers Jacko
Pi[NFA]=B256R0USB CL4SD8GB Raspbian Stock.
Pi[Work]=A+256 CL4SD8GB Raspbian Stock.
My favourite constant 1.65056745028

User avatar
jackokring
Posts: 816
Joined: Tue Jul 31, 2012 8:27 am
Location: London, UK
Contact: ICQ

Re: Minor improvements for a model C

Fri Aug 10, 2012 3:31 am

Or in software reserve many buffers for swapfs, and use kernal (TLB free) copy. The page table virtual memory TLB mechanism, can work as the cache association table. The fact that only swapfs pages use these buffers, means CPU usage goes up, and IO efficiency goes up. The fact that some memory is used causes more virtual memory swapping of pages (a few %), and makes for a preference for code which does not heavily use the ext based file system, as the CPU does not care for buffer contents and use, just that it needs to be removed to make space, using an LRU style algorithm.

Perhaps a +1 use on read, a +2 use on write, with a >>1 on overflow and flag_not_done and an XOR flag_done there and a two cycle small number inconsideration. Not linear LRU, monotonic enough of a sequence? For some hardware?

http://www.win.tue.nl/~aeb/linux/lk/lk-9.html for a lot about kernel memory as once was, but typing takes time. And it's mostly relevant. A swapfs block slab cache, with added write out to cache even if page read only or unwritten (but not to flash)?

Cheers Jacko
Pi[NFA]=B256R0USB CL4SD8GB Raspbian Stock.
Pi[Work]=A+256 CL4SD8GB Raspbian Stock.
My favourite constant 1.65056745028

Olavxxx
Posts: 7
Joined: Fri Aug 10, 2012 9:05 am

Re: Minor improvements for a model C

Fri Aug 10, 2012 9:08 am

Hi,
I have my thoughts about the hardware, size and placement of ports.
As some requested, I'd also like the ports all on one side.

Also I would like to use smaller ports for HDMI.
As far as the USB-ports, I'd like them to be not dual height.

If I understood correctly, the USB is really an usb hub which also connects the LAN.
My thought is that I would like a REALLY LOW card :-) Then maybe use breakout cables like ultraportables do.

I know it's not the main goal of the PI to be used as a mediacentre, etc. but this would allow one to create very slim devices. of course I can desolder the parts and buy components, but I'd rather pay 5$ more for a "nanoportable" version of the PI.

User avatar
reiuyi
Posts: 165
Joined: Sun Oct 09, 2011 4:59 pm
Contact: Website

Re: Minor improvements for a model C

Sat Aug 11, 2012 12:03 pm

None of the points you mentioned matter to the educative goals. You're allowed to use raspi as whatever (it's yours, you bought it), but don't expect any hardware changes for non-critical aspects.

USB power: Works fine for its intended goal: keyboard and a mouse. Even usb sticks work!

Sound: It's about FM quality, way better than my old soundblasters. Nowadays PWM is almost always used for DACs (often in combination with a low-brand opamp). I don't see the issue again, it works fine for educative goals and there is a digital audio out if you really must. I'll just assume you never envisioned TDA1541-like quality.

Size: Not relevant for educative purposes at all. I can list tens of commercial devices that don't have mounting holes (seriously, open up your mobile phone or your mp3 player and there's never mounting holes in the modern ones). The low number of IO pins has to do with pcb routing, not size of the board. If you want more IOs, the entire thing would need to be redesigned as an 8 or 10 layer PCB and that's just too expensive (currently standard 6-layer). Although "connectors on all sides" is an often criticized point, it's not relevant for its purpose at all.

Model ID: This is just purely off-topic; an unreasonable demand. Raspi has gone all the way to supply some decoding codecs without there being any need, and now people are demanding even more? What on earth are you going to do with the h264 encoding license? There is no video-input on this thing and the camera connector is the only video-input and it requires specialized GPU drivers which you will never be able to produce yourself.

alexchamberlain
Posts: 121
Joined: Thu Jun 14, 2012 11:20 am
Location: Leamington Spa, UK
Contact: Website

Re: Minor improvements for a model C

Sat Aug 11, 2012 12:36 pm

reiuyi wrote:None of the points you mentioned matter to the educative goals. You're allowed to use raspi as whatever (it's yours, you bought it), but don't expect any hardware changes for non-critical aspects.
I agree with your sentiment, but not everything you have said...
reiuyi wrote:USB power: Works fine for its intended goal: keyboard and a mouse. Even usb sticks work!
This is causing A LOT of issues with beginners. Just check the forums or Raspberry Pi.SE. I don't think the fuses are quite performing as expected, and sorting this will increase stability a lot. Stability has to be at the top of the list of goals.
Developer of piimg, a utility for working with RPi images.

techpaul
Posts: 1512
Joined: Sat Jul 14, 2012 6:40 pm
Location: Reading, UK
Contact: Website

Re: Minor improvements for a model C

Sat Aug 11, 2012 8:25 pm

reiuyi wrote:None of the points you mentioned matter to the educative goals. You're allowed to use raspi as whatever (it's yours, you bought it), but don't expect any hardware changes for non-critical aspects.

USB power: Works fine for its intended goal: keyboard and a mouse. Even usb sticks work!
Yes no maybe, in general yes, but limiting the power for USB BELOW the spec, means there will be problems when even more users or teachers or children go its a USB socket I can add my USB wifi/cam or other gadget, but it wont work. Waiting for a 2.5in USB hard drive to be added.

This is like standard UK 13A mains sockets wired up with 3A cable, because its intended uise was a clock.
However sometime later it gets cold someone comes along a 3kW heater, finds the convenient 13A socket and melts the cable and starts a fire.

If you put a standard connector and standard interface on a board for general use, then it should follow the standard, for general users.
Just another techie on the net - For GPIO boards see http:///www.facebook.com/pcservicesreading
or http://www.pcserviceselectronics.co.uk/pi/

alexchamberlain
Posts: 121
Joined: Thu Jun 14, 2012 11:20 am
Location: Leamington Spa, UK
Contact: Website

Re: Minor improvements for a model C

Sat Aug 11, 2012 9:27 pm

techpaul wrote:limiting the power for USB BELOW the spec
The spec says a unit load is 100mA and a device gets one unit load until it asks for more... 140mA is ABOVE spec?
Developer of piimg, a utility for working with RPi images.

User avatar
jackokring
Posts: 816
Joined: Tue Jul 31, 2012 8:27 am
Location: London, UK
Contact: ICQ

Re: Minor improvements for a model C

Sat Aug 11, 2012 9:42 pm

alexchamberlain wrote:
techpaul wrote:limiting the power for USB BELOW the spec
The spec says a unit load is 100mA and a device gets one unit load until it asks for more... 140mA is ABOVE spec?
I think the spec is more like, a device gets one unit load and it may ask for more, but like all negotiations, more may not be available. That's the difference between unit load, and maximum allowed ratings. Because something you add should be a unit, as in single thing. The 500mA thing came from a reasonable estimate of cable and possible socket performance, but as usual PCs gave 500mA without needing to be asked, as the circuit was simpler, and some equipment even went so far as to draw any power they could get, and only relax when the voltage dropped. In short, the good old days of sucking out loads of power with a maximal power boost regulator are what some fans still want. :P Look at my USB cooker, it has two rings too....

Cheers Jacko
Pi[NFA]=B256R0USB CL4SD8GB Raspbian Stock.
Pi[Work]=A+256 CL4SD8GB Raspbian Stock.
My favourite constant 1.65056745028

techpaul
Posts: 1512
Joined: Sat Jul 14, 2012 6:40 pm
Location: Reading, UK
Contact: Website

Re: Minor improvements for a model C

Sat Aug 11, 2012 9:57 pm

alexchamberlain wrote:
techpaul wrote:limiting the power for USB BELOW the spec
The spec says a unit load is 100mA and a device gets one unit load until it asks for more... 140mA is ABOVE spec?
Jez, the unit load is for devices that do not need to enumerate for higher power loads up to 500mA, during enumeration the device requests it max current level if more than 100mA to the host. The host or even unpowered hub must try to honour what requests it can, if unable to fail to enumerate the device (due to power combinations on its ports) and report an error to the host OS including port wants too much power.

Powered Hosts to follow USB 2.0 spec (and the USB 2.0 subset of USB 3.0) should be able to supply up to 500 mA.

Wondering how many others on here like me have deigned, manufactured and sold lots of devices with USB
as devices and fully powered fully compliant USB 2.0 Hubs, with upto 500mA per port loads.
Just another techie on the net - For GPIO boards see http:///www.facebook.com/pcservicesreading
or http://www.pcserviceselectronics.co.uk/pi/

techpaul
Posts: 1512
Joined: Sat Jul 14, 2012 6:40 pm
Location: Reading, UK
Contact: Website

Re: Minor improvements for a model C

Sat Aug 11, 2012 10:06 pm

techpaul wrote:
alexchamberlain wrote:
techpaul wrote:limiting the power for USB BELOW the spec
The spec says a unit load is 100mA and a device gets one unit load until it asks for more... 140mA is ABOVE spec?
Jez, the unit load is for devices that do not need to enumerate for higher power loads up to 500mA, during enumeration the device requests it max current level if more than 100mA to the host. The host or even unpowered hub must try to honour what requests it can, if unable to fail to enumerate the device (due to power combinations on its ports) and report an error to the host OS including port wants too much power.

Powered Hosts to follow USB 2.0 spec (and the USB 2.0 subset of USB 3.0) should be able to supply up to 500 mA.

Wondering how many others on here like me have deigned, manufactured and sold lots of devices with USB
as devices and fully powered fully compliant USB 2.0 Hubs, with upto 500mA per port loads.
By powered I mean in USB terms self-powered as in has its own power supply as compared to bus-powered which referes to devices and unpowred hubs that take their power from the USB lead and host port.

By 500mA I mean the max is 500mA per port to be spec compliant, if a host wishes to allow more then that is its problem.
Just another techie on the net - For GPIO boards see http:///www.facebook.com/pcservicesreading
or http://www.pcserviceselectronics.co.uk/pi/

User avatar
jackokring
Posts: 816
Joined: Tue Jul 31, 2012 8:27 am
Location: London, UK
Contact: ICQ

Re: Minor improvements for a model C

Sat Aug 11, 2012 10:36 pm

The spec says a unit load is 100mA and a device gets one unit load until it asks for more... 140mA is ABOVE spec?
Jez, the unit load is for devices that do not need to enumerate for higher power loads up to 500mA, during enumeration the device requests it max current level if more than 100mA to the host. The host or even unpowered hub must try to honour what requests it can, if unable to fail to enumerate the device (due to power combinations on its ports) and report an error to the host OS including port wants too much power.

Powered Hosts to follow USB 2.0 spec (and the USB 2.0 subset of USB 3.0) should be able to supply up to 500 mA.

By powered I mean in USB terms self-powered as in has its own power supply as compared to bus-powered which referes to devices and unpowred hubs that take their power from the USB lead and host port.
So if the power supply came with the Pi, it would have it's own supply. So you are only compliant when using USB loopback from a hub powering, as the self powering is not done, and no negotiation is made to get 500mA from a power adapter or hub. :roll:
By 500mA I mean the max is 500mA per port to be spec compliant, if a host wishes to allow more then that is its problem.
So if the max is 500mA, what is the absolute minimum?
Pi[NFA]=B256R0USB CL4SD8GB Raspbian Stock.
Pi[Work]=A+256 CL4SD8GB Raspbian Stock.
My favourite constant 1.65056745028

techpaul
Posts: 1512
Joined: Sat Jul 14, 2012 6:40 pm
Location: Reading, UK
Contact: Website

Re: Minor improvements for a model C

Sat Aug 11, 2012 10:43 pm

jackokring wrote:
The spec says a unit load is 100mA and a device gets one unit load until it asks for more... 140mA is ABOVE spec?
Jez, the unit load is for devices that do not need to enumerate for higher power loads up to 500mA, during enumeration the device requests it max current level if more than 100mA to the host. The host or even unpowered hub must try to honour what requests it can, if unable to fail to enumerate the device (due to power combinations on its ports) and report an error to the host OS including port wants too much power.

Powered Hosts to follow USB 2.0 spec (and the USB 2.0 subset of USB 3.0) should be able to supply up to 500 mA.

By powered I mean in USB terms self-powered as in has its own power supply as compared to bus-powered which referes to devices and unpowred hubs that take their power from the USB lead and host port.
So if the power supply came with the Pi, it would have it's own supply. So you are only compliant when using USB loopback from a hub powering, as the self powering is not done, and no negotiation is made to get 500mA from a power adapter or hub. :roll:
By 500mA I mean the max is 500mA per port to be spec compliant, if a host wishes to allow more then that is its problem.
So if the max is 500mA, what is the absolute minimum?
If no protection or current limit set sensibly on host, whatever will cause a failure, or fire or other cataastrophe in the host port/hub, cheap usb cable, device.

Basically whichever parts is first to chicken out and choose to send smoke signals :)
Just another techie on the net - For GPIO boards see http:///www.facebook.com/pcservicesreading
or http://www.pcserviceselectronics.co.uk/pi/

techpaul
Posts: 1512
Joined: Sat Jul 14, 2012 6:40 pm
Location: Reading, UK
Contact: Website

Re: Minor improvements for a model C

Sat Aug 11, 2012 10:45 pm

In reality whether the Pi comes with or without a PSU makes no difference its hub ports are self-powered host ports, as nothing works without a PSU.

Unless of course you have created a Telepathy Interface and Telpathy Power Transfer for you Pi :lol:
Just another techie on the net - For GPIO boards see http:///www.facebook.com/pcservicesreading
or http://www.pcserviceselectronics.co.uk/pi/

User avatar
mahjongg
Forum Moderator
Forum Moderator
Posts: 12123
Joined: Sun Mar 11, 2012 12:19 am
Location: South Holland, The Netherlands

Re: Minor improvements for a model C

Sat Aug 11, 2012 11:16 pm

jackokring wrote:Umm, difficult to say.

Some analog inputs on the GPIO.
A DC filter on the sound out.
A VGA set of holes on the board.

Cheers Jacko
Some analog inputs on the GPIO isn't possible, the SoC doesn't have any analog inputs, so the PI will never have them either! What would be possible though, is to have is an I2S (codec) interface, so you could add an audio codec if you need one. Actually all the signals are present on the current board, you just have to solder the board now to get to them.

The current PWM analog audio output does have a "DC filter" (AKA "capacitor") on the two audio channels. Namely C48 & C34.

I understand holes, but what are "VGA holes", do you mean the 4 screw holes that are available on the back of a monitor, so you can mount the monitor on the wall? Do you want to use these to mount a PI to the monitor? There are cases that will allow that now!

User avatar
jackokring
Posts: 816
Joined: Tue Jul 31, 2012 8:27 am
Location: London, UK
Contact: ICQ

Re: Minor improvements for a model C

Sat Aug 11, 2012 11:18 pm

techpaul wrote:Unless of course you have created a Telepathy Interface and Telpathy Power Transfer for you Pi :lol:
dE/dL = L.I.dI/dL + (1/2).I^2 ... L > 0, I > 0.. unfortunately the regulator can get sparky. :D
Pi[NFA]=B256R0USB CL4SD8GB Raspbian Stock.
Pi[Work]=A+256 CL4SD8GB Raspbian Stock.
My favourite constant 1.65056745028

User avatar
reiuyi
Posts: 165
Joined: Sun Oct 09, 2011 4:59 pm
Contact: Website

Re: Minor improvements for a model C

Sat Aug 11, 2012 11:38 pm

techpaul wrote:Yes no maybe, in general yes, but limiting the power for USB BELOW the spec, means there will be problems when even more users or teachers or children go its a USB socket I can add my USB wifi/cam or other gadget, but it wont work. Waiting for a 2.5in USB hard drive to be added.

If you put a standard connector and standard interface on a board for general use, then it should follow the standard, for general users.
I don't agree. Even a 500mA limit won't boot up most external harddrives, nor will it heat up your cup of tea. There's a reason why external hdds often have a Y-shaped cable. The polyfuse was a silly design choice, though PSU requirements would be insane without them (try finding a decent good quality 5v 2A with a micro usb plug, you won't find one!). The PCB lanes are also too thin to carry such loads; adding additional costs to pcb manufacturing.

All in all; having full-sized full-power usb ports is an unreasonable demand that's not necessary for its educative goals. Kb and mouse work fine, usb stick works fine, some wifi adapters work fine.. The list of compatible USB devices is miles long!

techpaul
Posts: 1512
Joined: Sat Jul 14, 2012 6:40 pm
Location: Reading, UK
Contact: Website

Re: Minor improvements for a model C

Sun Aug 12, 2012 12:05 am

reiuyi wrote:
techpaul wrote:Yes no maybe, in general yes, but limiting the power for USB BELOW the spec, means there will be problems when even more users or teachers or children go its a USB socket I can add my USB wifi/cam or other gadget, but it wont work. Waiting for a 2.5in USB hard drive to be added.

If you put a standard connector and standard interface on a board for general use, then it should follow the standard, for general users.
I don't agree. Even a 500mA limit won't boot up most external harddrives, nor will it heat up your cup of tea. There's a reason why external hdds often have a Y-shaped cable. The polyfuse was a silly design choice, though PSU requirements would be insane without them (try finding a decent good quality 5v 2A with a micro usb plug, you won't find one!). The PCB lanes are also too thin to carry such loads; adding additional costs to pcb manufacturing.

All in all; having full-sized full-power usb ports is an unreasonable demand that's not necessary for its educative goals. Kb and mouse work fine, usb stick works fine, some wifi adapters work fine.. The list of compatible USB devices is miles long!
I note you did not quote my bit about 13A (thirteen Amp) sockets wired with 3A (three Amp) cable very teling.

Live in the real world a general computer with USB ports will have users plugging in any USB device that was the point of the standard.

The micro USB connector for power was a silly idea from day one as many micro USB cables are price manufactured down for SPECIFIC very low power devices and users will pick up any USB cable thinking they are all the same and many will put excessive volt drops and current limits.

As to power tracking I bet most of the 5V is actually in inner planes used only for power, so in effect is very wide.

Anyway a track capable of carrying 1A to the USB (two ports) is not as wide as you may think, it is around 0.4 to 0.5 mm wide for outer layers which it probably is at the moment or wider. It would have to be less than 1oz Cu to need wider than 1mm track on outer layers. Yes I lay out PCBs for my designs as well.

FYI last lot of USB 2.5 inch hard drives I used draw from USB power at ratring of 500mA last measured one was 400mA. The 500mA limit works with all self-powered USB 2.5 hard drives that I have used current job for one customer requires TEN of them they all work fine. 2.5 inch drives are laptop drives meant for low power usage and 5V only. Oh I have used them oin many systems of various vintages as well.

So far you are the only person who supports the current scheme I see no one else supporting in ANY thread anywhere. Having the attitude of it is fine is someone who does not understand the amount of things that can go wrong or has experience of real world users in their day to day life with electronic equipment.

I am someone who has for years had to design things to ensure users cannot do something beyond the spec with trying really hard, not doing what they are supposed to do.

I just hope you don't do any electrical wiring.
Just another techie on the net - For GPIO boards see http:///www.facebook.com/pcservicesreading
or http://www.pcserviceselectronics.co.uk/pi/

gritz
Posts: 449
Joined: Sat Jan 28, 2012 2:33 am

Re: Minor improvements for a model C

Sun Aug 12, 2012 12:32 am

I have to concur with techpaul to a degree. The most disappointing thing about this whole situation is that the USB current limit was deliberate but not documented and was left to users to find out for themselves, rather than made obvious in the spec so that users could make their own decisions.

User avatar
mahjongg
Forum Moderator
Forum Moderator
Posts: 12123
Joined: Sun Mar 11, 2012 12:19 am
Location: South Holland, The Netherlands

Re: Minor improvements for a model C

Sun Aug 12, 2012 12:40 am

jackokring wrote:
techpaul wrote:Unless of course you have created a Telepathy Interface and Telpathy Power Transfer for you Pi :lol:
dE/dL = L.I.dI/dL + (1/2).I^2 ... L > 0, I > 0.. unfortunately the regulator can get sparky. :D
hmmm, I think Tesla did something like this. Conspiracy theorists say he was murdered in his hotelroom because of it. :roll:

User avatar
Lob0426
Posts: 2198
Joined: Fri Aug 05, 2011 4:30 pm
Location: Susanville CA.
Contact: Website

Re: Minor improvements for a model C

Sun Aug 12, 2012 1:13 am

I have replaced the 140ma fuses with 500ma. Some of my USB drives still will not run directly from the RasPi. So it is not just the fuses that are limiting the current. It was the same with a 700ma poly fuse (the original F3 placed into F2 on my board). I was using a 2A iPad charger for these tests. The test voltage at TP1/TP2 do not reflect the voltage at F1/F2. The fuses size is only half of the problem. The Resistance of these smaller fuses is much higher. 140ma poly fuses run 2 to 6 ohms by spec (mine were 5.9 and 7.2) the the 500 ma replacements run at .014 to .025. A big difference in those figures.

The Micro connector was chosen as it is a "standard" power connector in the UK/EU. Nothing is really standard in the U.S.. I did not agree with the change, but there were a lot of people that asked for "cell phone" chargers, so I did make my opinion of the change known. Most of these people really did not know what they were asking for. The Foundation latched onto this idea because phone chargers are readily available and they felt it would lower the overall cost of operation. I would be interested to see if it is still considered as a great idea now, that the quality of phone chargers has really been tested on the RasPi. It did however allow the Foundation to meet their goals. Their goals were mainly price based as they were trying to get these into the lower income brackets and third world countries childrens hands. there is always a trade off in any such endeavor. The Alpha model was really efficient, voltage tolerant and overall a tougher customer when placed into neophyte hands. Technically it was vastly superior to the production model in all aspects. But it was harder to meet the cost goals that had been set. Also note that with the change to 5v versus 6v to 20v, many components were able to be eliminated. This also let the foundation design at a low enough price to bring in "for profit" partners that would not have been possible with the Alphas. In other words a profit margin was built into the production versions. You may not like this but there are a whole lot more Raspberry Pi's out in the world this way than there would have been if the Alphas had been produced by the foundation alone.
If you really want to get a better RasPi then ask that the Alpha model go into production with a 512MB memory module in place of the 256MB. The Alpha design was funded by Broadcom. So the board design really does not belong to the foundation. But if someone asked nicely maybe Broadcom would allow it to be made!

Alpha specs included:
6v to 20v operation through a 5.5mm power connector.
full spec (500ma) USB ports
Switching rather than linear regulators. Idle current was like under a watt with 2.25 watt max load (I believe that included a draw from keyboard and mouse)
More GPIO! all 26 pins were connected. Included I2s and I2c, 2 PWM's (if memory serves me)
There were a couple of other pin strips on the board as well.
512MB version 2.0 as WordPress Server
Motorola Lapdock with Pi2B
Modded Rev 1.0 with pin headers at USB

http://rich1.dyndns.tv/
(RS)Allied ships old stock to reward its Customers for long wait!

payturr
Posts: 64
Joined: Mon Jul 09, 2012 9:46 pm

Re: Minor improvements for a model C

Sun Aug 12, 2012 3:18 am

I think if there was to be a Model C, it would be a high end model - leaving the A to be basic for schools and hackers, B for house use and recreational purposes, and the C for people who need more performance (people like me who try to use the Pi as a primary desktop) - maybe a ARMv7 CPU, 1-1.2GHz CPU, double the RAM of the A/B, and perhaps a built in wifi card. That is, if all of this could still cost underneath a 75$ price point.

W. H. Heydt
Posts: 10762
Joined: Fri Mar 09, 2012 7:36 pm
Location: Vallejo, CA (US)

Re: Minor improvements for a model C

Sun Aug 12, 2012 5:56 am

payturr wrote:I think if there was to be a Model C, it would be a high end model - leaving the A to be basic for schools and hackers, B for house use and recreational purposes, and the C for people who need more performance (people like me who try to use the Pi as a primary desktop) - maybe a ARMv7 CPU, 1-1.2GHz CPU, double the RAM of the A/B, and perhaps a built in wifi card. That is, if all of this could still cost underneath a 75$ price point.
The problem with that is, that anyone other than a charity (aka "non-profit") is going to expect to be paid for time and resources to design the beast AND make a decent profit.

Now, except for the price point, the Odriod-x comes pretty close to what you're asking for...1.4GHz quad core processor, 1GB RAM, 6 USB ports and some other stuff. IT costs $130. which is still pretty cheap. The capabilities there are why I refer to it as the server for the Pi (and I'm thinking of doing just that with a couple of them next year...provided one can run a full, open version of Linux on it instead of Android).

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 23358
Joined: Sat Jul 30, 2011 7:41 pm

Re: Minor improvements for a model C

Sun Aug 12, 2012 9:41 am

techpaul wrote:
reiuyi wrote:
techpaul wrote:Yes no maybe, in general yes, but limiting the power for USB BELOW the spec, means there will be problems when even more users or teachers or children go its a USB socket I can add my USB wifi/cam or other gadget, but it wont work. Waiting for a 2.5in USB hard drive to be added.

If you put a standard connector and standard interface on a board for general use, then it should follow the standard, for general users.
I don't agree. Even a 500mA limit won't boot up most external harddrives, nor will it heat up your cup of tea. There's a reason why external hdds often have a Y-shaped cable. The polyfuse was a silly design choice, though PSU requirements would be insane without them (try finding a decent good quality 5v 2A with a micro usb plug, you won't find one!). The PCB lanes are also too thin to carry such loads; adding additional costs to pcb manufacturing.

All in all; having full-sized full-power usb ports is an unreasonable demand that's not necessary for its educative goals. Kb and mouse work fine, usb stick works fine, some wifi adapters work fine.. The list of compatible USB devices is miles long!
I note you did not quote my bit about 13A (thirteen Amp) sockets wired with 3A (three Amp) cable very teling.

Live in the real world a general computer with USB ports will have users plugging in any USB device that was the point of the standard.

The micro USB connector for power was a silly idea from day one as many micro USB cables are price manufactured down for SPECIFIC very low power devices and users will pick up any USB cable thinking they are all the same and many will put excessive volt drops and current limits.

As to power tracking I bet most of the 5V is actually in inner planes used only for power, so in effect is very wide.

Anyway a track capable of carrying 1A to the USB (two ports) is not as wide as you may think, it is around 0.4 to 0.5 mm wide for outer layers which it probably is at the moment or wider. It would have to be less than 1oz Cu to need wider than 1mm track on outer layers. Yes I lay out PCBs for my designs as well.

FYI last lot of USB 2.5 inch hard drives I used draw from USB power at ratring of 500mA last measured one was 400mA. The 500mA limit works with all self-powered USB 2.5 hard drives that I have used current job for one customer requires TEN of them they all work fine. 2.5 inch drives are laptop drives meant for low power usage and 5V only. Oh I have used them oin many systems of various vintages as well.

So far you are the only person who supports the current scheme I see no one else supporting in ANY thread anywhere. Having the attitude of it is fine is someone who does not understand the amount of things that can go wrong or has experience of real world users in their day to day life with electronic equipment.

I am someone who has for years had to design things to ensure users cannot do something beyond the spec with trying really hard, not doing what they are supposed to do.

I just hope you don't do any electrical wiring.
I think it unreasonable to expect any USB device plugged in to work at first plug in. After all, that's not the case on PC's either, even Windows might require drivers. Linux may not support the device at all. Now, I know that's at the software level rather than hardware, but the effect to the user is the same- plug in doesn't work. Now, software can be changed (if someone fixes it), as can the board.
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."

techpaul
Posts: 1512
Joined: Sat Jul 14, 2012 6:40 pm
Location: Reading, UK
Contact: Website

Re: Minor improvements for a model C

Sun Aug 12, 2012 10:17 am

Lob0426 wrote:I have replaced the 140ma fuses with 500ma. Some of my USB drives still will not run directly from the RasPi. So it is not just the fuses that are limiting the current. It was the same with a 700ma poly fuse (the original F3 placed into F2 on my board). I was using a 2A iPad charger for these tests. The test voltage at TP1/TP2 do not reflect the voltage at F1/F2. The fuses size is only half of the problem. The Resistance of these smaller fuses is much higher. 140ma poly fuses run 2 to 6 ohms by spec (mine were 5.9 and 7.2) the the 500 ma replacements run at .014 to .025. A big difference in those figures.
None of that surprises me, early in year had customer who wanted similar type of protection on some chips under test, only trouble is with 600 under test at same time, it would have meant 600 trimmers to cope with the variations in voltage drops. Thankfully persuaded them otherwise as they got the spec wrong twice and the mods I had to do was a much smaller number than 600.
The Micro connector was chosen as it is a "standard" power connector in the UK/EU. Nothing is really standard in the U.S.. I did not agree with the change, but there were a lot of people that asked for "cell phone" chargers, so I did make my opinion of the change known. Most of these people really did not know what they were asking for. The Foundation latched onto this idea because phone chargers are readily available and they felt it would lower the overall cost of operation. I would be interested to see if it is still considered as a great idea now, that the quality of phone chargers has really been tested on the RasPi.
Well confirms my suspicions on that subject.
If you really want to get a better RasPi then ask that the Alpha model go into production with a 512MB memory module in place of the 256MB. The Alpha design was funded by Broadcom. So the board design really does not belong to the foundation. But if someone asked nicely maybe Broadcom would allow it to be made!

Alpha specs included:
6v to 20v operation through a 5.5mm power connector.
full spec (500ma) USB ports
Switching rather than linear regulators. Idle current was like under a watt with 2.25 watt max load (I believe that included a draw from keyboard and mouse)
More GPIO! all 26 pins were connected. Included I2s and I2c, 2 PWM's (if memory serves me)
There were a couple of other pin strips on the board as well
Well if the Alpha also made and the pin strip for GPIO actually a shrouded connectror like standard IDC ribbon cable connectors I would agree, as that is the main non-polarised connector on the board. The only other accessible one being the JTAG.

Personally going anywhere near education long term with pin strips and not polarised connectors is asking for trouble. Yes I deal with education market daily, the other half is a teacher for ICT and Computing subjects, so often making things for her or doing talks or helipng on the kids projects as the 'customer'.
Just another techie on the net - For GPIO boards see http:///www.facebook.com/pcservicesreading
or http://www.pcserviceselectronics.co.uk/pi/

techpaul
Posts: 1512
Joined: Sat Jul 14, 2012 6:40 pm
Location: Reading, UK
Contact: Website

Re: Minor improvements for a model C

Sun Aug 12, 2012 10:27 am

jamesh wrote:I think it unreasonable to expect any USB device plugged in to work at first plug in. After all, that's not the case on PC's either, even Windows might require drivers. Linux may not support the device at all. Now, I know that's at the software level rather than hardware, but the effect to the user is the same- plug in doesn't work. Now, software can be changed (if someone fixes it), as can the board.
No you misunderstand how it works there are various general class drivers for supporting functions of devices on many OS (including windows) for such things as

Human Interface - keyboard, mouse, tablet

Storage - pen drives hard drives

Video - generic web cams

Enumerating the current requirements is long before device driver activation to support the functions of the device. This is part of the bus prototcol and root hub driver in the OS to handle this.
Just another techie on the net - For GPIO boards see http:///www.facebook.com/pcservicesreading
or http://www.pcserviceselectronics.co.uk/pi/

Return to “General discussion”