USB programmable Open Source Hardware
in [FPGA]
|
From: nobody on 30 Sep 2009 12:02 The idea of obsolete has come up a couple of times in this discussion and I would like to take a minute to address it. Being a single engineer in an office somewhere I do not have much time to redesign, therefore one design to fit a multitude of applications. My applications are growing with the times and speed is key, a single bit stream ripping through copper at an increasing bandwidth, even though a GT may be pushing several gigs it requires 10s of gigs to support the rising edge of the digital signal. It has been said that the chip is outdated the board design was not meant specifically for the chip but for upcoming chips and the gigabit tranciever resources that are coming at a decreasing price. The chip may be obsolete but my four layer board design is not and will support a bga chip and GT resources. There is a reason for xilinx application notes such as xapp623 and xapp489 get it done once and be get it done right. I am not saying that anybodies project is wrong but there is a reason for four and more layers, it may not be for hobbyist but again I am not a hobbyist. respectfully Cy Drollinger
From: Nico Coesel on 30 Sep 2009 12:18 emeb <ebrombaugh(a)gmail.com> wrote: >On Sep 29, 2:00=A0pm, n...(a)puntnl.niks (Nico Coesel) wrote: >> I recenty did an FPGA design on a 2 layer board. Used >> the bottom layer as a ground plane and put 2 rings and a plane >> underneath the FPGA on the top layer. And of course a size 0402 >> decoupling capacitor on each power supply pin. > >That's similar to what I did: on the backside I had ground plane plus >a 'snail-shell' of 3 concentric supply traces with vias through to the >top where the VQ100 package sits. I used 0603 caps on every supply pin >- some on the back, some on the front. Didn't have any supply >problems. > >FWIW - I'm impressed that you used 0402 parts. I haven't tried to >handle parts that small yet, but I'm getting pretty good at the 0603s. We had the board (18" x 8") assembled. But it is possible to mount 0402 parts by hand. It just takes some getting used to like handling 0603 parts once did. You'll need a magnifier for inspection though. -- Failure does not prove something is impossible, failure simply indicates you are not using the right tools... "If it doesn't fit, use a bigger hammer!" --------------------------------------------------------------
From: Antti.Lukats on 30 Sep 2009 12:44 On Sep 30, 7:02 pm, nobody <cydrollin...(a)gmail.com> wrote: > The idea of obsolete has come up a couple of times in this discussion > and I would like to take a minute to address it. Being a single > engineer in an office somewhere I do not have much time to redesign, > therefore one design to fit a multitude of applications. My > applications are growing with the times and speed is key, a single bit > stream ripping through copper at an increasing bandwidth, even though > a GT may be pushing several gigs it requires 10s of gigs to support > the rising edge of the digital signal. It has been said that the chip > is outdated the board design was not meant specifically for the chip > but for upcoming chips and the gigabit tranciever resources that are > coming at a decreasing price. The chip may be obsolete but my four > layer board design is not and will support a bga chip and GT > resources. There is a reason for xilinx application notes such as > xapp623 and xapp489 get it done once and be get it done right. I am > not saying that anybodies project is wrong but there is a reason for > four and more layers, it may not be for hobbyist but again I am not a > hobbyist. > > respectfully > > Cy Drollinger Cy.. do whatever you like.. :) 4 layers is too little actually.. 6 to 10 is more likely layer count for "real boards" just i got the impression you hoped that lots of hobby open-source people would build your 4-layer board using your gerber files. I do not think that will be case.. that the point i was trying to make. Antti
From: nobody on 30 Sep 2009 13:19 Antti, You seem to need to slam my stuff as if it is incorrect, that's a difference between you and I. My project as it stands is what it is and suits my purpose, however others may find use of such a project and am trying to make it available. Other boards and projects mentioned on this forum are what they are and suit their purpose. But when an individual makes remarks about obsolete, not useful, not necessary and the like explanation is necessary. As engineers we are to take the body of knowledge we have been allotted and use it, break it, expand it and make a contribution. If younger engineers read our comments for their enhancement they should be unbiased and factual as possible. If one were to read over these post they might think that there is no need for a four layer board and as you commented that is just not true. I take offense with your comment about a "real board" needing 6 - 10 layers, well, taken in jest, nope, not funny. Let us try and keep our communication at as high technical level as possible. My points about this post's origination: 1. I have completed a project that is affordable and can be built by a hobbyist. 2. Is there any need for an open source project with these mentioned capabilities? I appreciate all the comments on this post and has been very educational. Cy Drollinger
From: Andy Peters on 30 Sep 2009 13:32
On Sep 30, 9:02 am, nobody <cydrollin...(a)gmail.com> wrote: > The chip may be obsolete but my four > layer board design is not and will support a bga chip and GT > resources. You do realize that most chip families are not pinout-compatible with one another? IOW, a Spartan6 in 256BGA is not going to work on a board designed for a Spartan3AN? And as I tried to get you to understand on the Xilinx PicoBlaze forum -- there's no such thing as a "standardized FPGA platform" (as much as Xilinx would have you believe otherwise). An audio-processing design has very little in common with a video-processing system. You choose an FPGA and design a board to meet the product requirements. You don't try to make your product requirements fit any random board you might already have built. -a |