The Pandaboard has to be heavily subsidized by TI. There's no way it can cost $175, since it's an 8-layer board, manufactured/tested partly in the US, and has 1GBytes of RAM using package-on-package technology. I don't think it's fair to compare.
Since FPGA's are good at processing massively parallel data, high bandwidth is necessary to do any work. I suppose CPU to FPGA bandwidth isn't important if they can both access RAM.
I did look into the LM3404 but I couldn't figure out how to set its current with a wide input voltage range.
The HV9918 I picked is ridiculously cheap at only $1 a chip. For my project the LED driver chip was the main BOM cost. For you it probably wouldn't reduce costs by much. Unfortunately my board has some high EMI emissions despite its extremely compact layout which I haven't been able to figure out. Then again, one of my other switching power supplies also had mysterious EMI so it could just be my layout skills.
I feel bad that the slashdotters have been pretty critical of your project. The cost is understandable as selling small quantities of these projects at near parts cost definitely does not make up for development time. Which is why I've decided to just sell them at cost in terms of labor which of course isn't feasible in the long run. But designing a product is something I enjoy doing so I don't mind.
The board is in Arduino shield form factor and it can drive RGB 700mA LED's of any voltage up to 30V. It has an onboard micro that communicates with I2C that allows you to dim the channels without having to do the PWM on the host micro. This is optional, and there will be cheaper versions that just takes in PWM input signals directly.
It seems like all these high power LED stuff is way overpriced. I designed this after seeing Sparkfun selling a similar board for $50! And now this, which costs at least $290 for a bare board.
I haven't gotten around to finishing it yet but I intend to sell these boards for only $25 for the basic feature version.
TCP connection timeouts on some routers default to 3600 seconds or one hour. So, when you use some Bittorrent or such, opening lots of connections, your router keeps these connections (even after disconnection) in its memory for up to an hour. It fills up and your router grinds to a halt, opening connections very slowly.
There's other timeouts too, but I'm not sure exactly what they do. Firmware like HyperWRT lets you change these timeouts to something much shorter, like 90 seconds, which typically prevents lock-ups like that.
(I'm actually not 100% sure that this is the sole cause for router lock-ups)
Sorry to reply to myself, but I forgot to add that Dell is not the only one that distributes drivers that disable stereo mix. Lenovo has these problems too:
How often is stereo mix actually used? People are overreacting; they think recording is disabled. That's ridiculous: tons of people use microphones, for gaming, voice chat, they couldn't afford to turn recording off. What is disabled is some boolean in the driver, not in the hardware that allows stereo mix to work.
How many people even use stereo mix? You record what's playing through your speakers. Any decent sound editing program is going to have some kind of mixer that will combine your recording with whatever you were playing along with.
The only use I see is for a program like FRAPS, which records your screen and sound.
One contract that I was at had an incompetent DBA. (The first time I began to question her abilities was when she asked if changing the desktop bitmap would impact system performance).
It does. At least on very old computers with little RAM, if you had a large desktop image, it'd have to swap it from disk to display it, so the desktop would take very long to display. I still remember the days...
The submitter used stronger language to describe the comment than the comment itself. Connes just said "The 'proof' is that of Theorem 7.3 page 29 in Li's paper, but I stopped reading it when I saw that he is extending the test function h from ideles to adeles by 0 outside ideles and then using Fourier transform (see page 31). This cannot work... "
Slashdot Mirror
The Pandaboard has to be heavily subsidized by TI. There's no way it can cost $175, since it's an 8-layer board, manufactured/tested partly in the US, and has 1GBytes of RAM using package-on-package technology. I don't think it's fair to compare.
I have to do this on every single website I visit anyway.
Remember the days when ctrl-+ only changed the font sizes and didn't scale the page layout? Wasn't too long ago...
That's really not of concern to anyone building electronics though.
Is it really unlimited? At what point will they start complaining?
Really? I wouldn't think that'd be a good idea for power consumption reasons.
Or if your meter knew the rate, it just needs to report a cost to the power company.
Damn how'd you know?
Since FPGA's are good at processing massively parallel data, high bandwidth is necessary to do any work. I suppose CPU to FPGA bandwidth isn't important if they can both access RAM.
Only PCI-E 1x interconnect between the CPU and FPGA? Kinda disappointing.
In related news, and also very interesting: http://www.eetimes.com/electronics-news/4210263/Intel-to-fab-FPGAs-for-startup-Achronix
Heh. No other way to navigate around their huge stock. There's just too many choices for every common component like caps.
I did look into the LM3404 but I couldn't figure out how to set its current with a wide input voltage range.
The HV9918 I picked is ridiculously cheap at only $1 a chip. For my project the LED driver chip was the main BOM cost. For you it probably wouldn't reduce costs by much. Unfortunately my board has some high EMI emissions despite its extremely compact layout which I haven't been able to figure out. Then again, one of my other switching power supplies also had mysterious EMI so it could just be my layout skills.
I feel bad that the slashdotters have been pretty critical of your project. The cost is understandable as selling small quantities of these projects at near parts cost definitely does not make up for development time. Which is why I've decided to just sell them at cost in terms of labor which of course isn't feasible in the long run. But designing a product is something I enjoy doing so I don't mind.
No it's (untracked) USPS first class, which is usually around two days for me. Overnight costs a lot more of course.
Use Digikey in the future. No minimum order and $3 shipping.
How often does one need a JTAG debugger anyway? You can do plenty without one.
Shamelessly linking my own blog here:
http://electronoblog.allanw.org/2010/10/triple-high-brightness-led-driver-arduino-shield/
The board is in Arduino shield form factor and it can drive RGB 700mA LED's of any voltage up to 30V. It has an onboard micro that communicates with I2C that allows you to dim the channels without having to do the PWM on the host micro. This is optional, and there will be cheaper versions that just takes in PWM input signals directly.
It seems like all these high power LED stuff is way overpriced. I designed this after seeing Sparkfun selling a similar board for $50! And now this, which costs at least $290 for a bare board.
I haven't gotten around to finishing it yet but I intend to sell these boards for only $25 for the basic feature version.
Who else read the title and was excited thinking there was an astonishing speed-up in SSD systems?
Did you know TI's educational division only accounted for 4% of their 2007 revenue?
TCP connection timeouts on some routers default to 3600 seconds or one hour. So, when you use some Bittorrent or such, opening lots of connections, your router keeps these connections (even after disconnection) in its memory for up to an hour. It fills up and your router grinds to a halt, opening connections very slowly.
There's other timeouts too, but I'm not sure exactly what they do. Firmware like HyperWRT lets you change these timeouts to something much shorter, like 90 seconds, which typically prevents lock-ups like that.
(I'm actually not 100% sure that this is the sole cause for router lock-ups)
Sorry to reply to myself, but I forgot to add that Dell is not the only one that distributes drivers that disable stereo mix. Lenovo has these problems too:
http://forums.lenovo.com/lnv/board/message?board.id=Special_Interest_General&thread.id=316&view=by_date_ascending&page=1
It's quite a popular thread, and Lenovo reps have posted too.
How often is stereo mix actually used? People are overreacting; they think recording is disabled. That's ridiculous: tons of people use microphones, for gaming, voice chat, they couldn't afford to turn recording off. What is disabled is some boolean in the driver, not in the hardware that allows stereo mix to work.
How many people even use stereo mix? You record what's playing through your speakers. Any decent sound editing program is going to have some kind of mixer that will combine your recording with whatever you were playing along with.
The only use I see is for a program like FRAPS, which records your screen and sound.
One contract that I was at had an incompetent DBA. (The first time I began to question her abilities was when she asked if changing the desktop bitmap would impact system performance).
It does. At least on very old computers with little RAM, if you had a large desktop image, it'd have to swap it from disk to display it, so the desktop would take very long to display. I still remember the days...
See the first result on google for "phonon.dll":
http://raviratlami1.blogspot.com/2008/07/how-to-download-and-install-phonondll.html
The submitter used stronger language to describe the comment than the comment itself. Connes just said "The 'proof' is that of Theorem 7.3 page 29 in Li's paper, but I stopped reading it when I saw that he is extending the test function h from ideles to adeles by 0 outside ideles and then using Fourier transform (see page 31). This cannot work... "
How about Ubuntu: Toaster edition?