• Please review our updated Terms and Rules here

Raspberry Pi update

MCM Electronics (http://www.mcmelectronics.com) is now offering some cases and kits for the Raspberry Pi.

Strange case. It's like a lid from something. It seems like it would waste space and make it difficult to attach switches, lights, or any of the things which normally install onto a flat surface. Perhaps I'm missing something.

patscc said:
Heatsinking always helps. That's physics for you.

Umm, that would be physics if there was a physical definition for the word "helps" or if it was specified what it helps. :)
 
Shouldn't the thing fit quite comfortably into an Altoids tin?

200px-Altoid_and_tin.JPG
 
Heatsinking always helps. That's physics for you.

Not in this case. The SoC used on the Raspberry Pi is specified to work at up to 85 degrees Celsius. In practice what sets the clock speed limit is the maximum switching speed of the transistors in the SoC. They will reach this limit well before reaching an operating temperature of 85 degrees Celsius in virtually any real world scenario. Heatsinking won't help them switch faster. Overvolting will though, but doing that voids the warranty unless the dynamic frequency scaling option is used.
 
Last edited:
I didn't mean heatsinking will help it run faster. A heat sink will dissipate heat from the part quicker than if there's no heatsink, which helps increase reliability and stability.
patscc
 
Right, my heatsink comment was regarding the comment that if overclocked it becomes unstable either instantly or after an hour. Sounds like typical overheating from the overclocking so I was just thinking putting a heatsink on that chip might help alleviate the heat issues and get you a stable overclocked chip. Just a thought though and I realize the average heatsink is about the size of the entire unit.
 
I didn't mean heatsinking will help it run faster. A heat sink will dissipate heat from the part quicker than if there's no heatsink, which helps increase reliability and stability.
patscc

I don't think it will. My lockups and crashes while testing overclocking on my Raspberry Pi don't seem to be related to heat.
Doing a continuous cat /sys/class/thermal/thermal_zone0/temp shows an unremarkable 58 degrees Celsius at the time of my last crash after approximately 5 hours of running a Quake 3 demo loop. I've run it at 67 degrees Celsius without problems.
 
After much testing I seem to have settled on my maximum overclock: 840 MHz CPU, 275 MHz GPU and 485 MHz RAM. I can achieve this without overvolting. Here are my nbench results:

Code:
BYTEmark* Native Mode Benchmark ver. 2 (10/95)
Index-split by Andrew D. Balsa (11/97)
Linux/Unix* port by Uwe F. Mayer (12/96,11/97)

TEST                : Iterations/sec.  : Old Index   : New Index
                    :                  : Pentium 90* : AMD K6/233*
--------------------:------------------:-------------:------------
NUMERIC SORT        :          268.96  :       6.90  :       2.27
STRING SORT         :          39.577  :      17.68  :       2.74
BITFIELD            :      8.6927e+07  :      14.91  :       3.11
FP EMULATION        :          55.076  :      26.43  :       6.10
FOURIER             :          2828.1  :       3.22  :       1.81
ASSIGNMENT          :          3.1746  :      12.08  :       3.13
IDEA                :           827.1  :      12.65  :       3.76
HUFFMAN             :          507.81  :      14.08  :       4.50
NEURAL NET          :           3.973  :       6.38  :       2.68
LU DECOMPOSITION    :           92.28  :       4.78  :       3.45
==========================ORIGINAL BYTEMARK RESULTS==========================
INTEGER INDEX       : 13.962
FLOATING-POINT INDEX: 4.613
Baseline (MSDOS*)   : Pentium* 90, 256 KB L2-cache, Watcom* compiler 10.0
==============================LINUX DATA BELOW===============================
CPU                 : 
L2 Cache            : 
OS                  : Linux 3.2.27+
C compiler          : gcc-4.7
libc                : libc-2.13.so
MEMORY INDEX        : 2.989
INTEGER INDEX       : 3.908
FLOATING-POINT INDEX: 2.558
Baseline (LINUX)    : AMD K6/233*, 512 KB L2-cache, gcc 2.7.2.3, libc-5.4.38
* Trademarks are property of their respective holder.

BTW, nbench was compiled with CFLAGS = -s -static -Wall -O3 -fomit-frame-pointer -funroll-loops -mcpu=arm1176jzf-s -mtune=arm1176jzf-s -mfpu=vfp -mfloat-abi=hard
 
If they port Linux to it, maybe, although $99 is a lot more than $35 that includes video, ethernet, usb.
They have a C compiler in their SDK for it, but I don't know if it's free.
patscc
 
They have several competitors but I'm not sure how available the competition is. Even Intel made a similar pricerange (under $100) SBC. However what's interesting to me is that in our Robot group not really anyone jumped onboard with the Raspberry. Effectively it costs a lot more than an arduino or other chip like the Teensy. Sorta the more you know the less you need sort of product. Most of the folks who have worked with making their own PIC, stamp, etc systems don't seem to have an interest in this since it's more of a computer and more expensive than what they'd want for a project or controller.
 
The only thing that is special about the RPi is the price tag. You can get better performance, better expand-ability, or both for a bit more money, And the gap is dropping by the day.
 
Now it comes with 512MB of memory, not 256MB.

It's only been six months since the first Raspberry Pi's were shipped. Even Apple wait longer between updates to their hardware! Judging from their forums many Raspberry Pi owners have been irritated by this change. Many also appear to be worried that the userbase will be fragmented. I'm sure there are many more who would express their annoyance there but are afraid that they'll be banned if they do.

Personally I can't wait until the Raspberry Pi Foundation have some direct competition. I'm just so sick of their hubris and their persistent downplaying and glossing over of the Raspberry Pi's issues.
 
Back
Top