|
|
|
|
 |
|
|
I'm Still Alive. . .
Yes, for all who actually might read my lowly blog and have been wondering why I haven't updated in so long, I can say for sure I'm still alive :-)
These past few months have been quite busy, and I have mainly lost an interest in blogging, which explains why I haven't updated in a long while. I would like to start blogging again so I can explain a new project I have been working on. It actually relates to my last post about a bus protocol. I'm essentially making a parallel computational system out of 8051 microcontrollers for a research course I am taking. The idea is to build a system that I can later build software around to further research in parallel computing. I chose 8051s for the main CPU nodes just because they're cheap and easy to use. It makes for building the project much easier than if I used a more powerful CPU, like an ARM. My goal is to eventually go for an ARM CPU in a later prototype.
Anyway, I just want to say I'm still alive and kicking (mainly on IRC these days) and more info on this project will be coming soon.
Bona Fide Work • 1 Comments/Trackbacks • Posted by: J Silverman |
|
|
|
|
|
|
|
|
|
 |
 |
February 3, 2006, 5:55 pm |
|
|
|
 |
|
|
|
|
Bus interface
8 data lines (Bus)
1 clock line
1 handshaking line
1 disable bus interrupts (DBI) line
Each node in the bus will be mapped to a specific address. When that address is sent on the bus, it will allow for the requested node to be interrupted and respond to the requester. The DBI line will disable this ability so that no extraneous interrupts occur during data transmission. The handshaking line will prevent other nodes in the bus from initiating communications during a transmission. The clock line signifies that proper data is on the bus and that the requested node should acknowledge that.
To prevent a fatal race condition from happening where two or more nodes on the bus pull the handshaking line down at the same time, each node making the request must disable bus interrupts and put each bit of the reqeusted node's address, starting with the LSB, one at a time with a slight delay on the bus and check to see if the bit placed on the bus was what it outputted. If the node expected to see a one or high on a line, yet it is zero or low, it will back off from continuing the communication and wait until the handshaking and DBI lines are high again. By backing off, it will immediately toggle the clock line, to signify to any others in the arbitration to pause, and set its connection to the bus, DBI, and handshaking lines to input only. The one node to win the first phase of the bus arbitration will raise the DBI line to allow for the requested node to respond. This same process must also happen when the address of the requester is transmitted, thereby preventing two requesters requesting the same one node. Each bit will be placed on the bus one at a time, and the one requester expecting to see a one on the line where it is zero will back off and toggle the clock line and set its bus, DBI, and handshaking lines to input only.
Master Requests Slave for Communication
master checks if handshaking or DBI lines are low
if low, communication is in progress
master pulls handshaking line low
master disables bus interrupts
master outputs slave address bit by bit on data bus
master performs bus arbitration as described above
if master wins arbitration, it enables bus interrupts
slave responds by disabling bus interrupts
master outputs its address bit by bit on the bus
master performs bus arbitration as described above
if master wins arbitration, it toggles the handshaking line
master sets data byte
master toggles clock
repeat data/clock until all data is sent
master releases handshaking line
slave releases DBI line
Slave Requests Master for Communication
slave checks if handshaking or DBI lines are low
if low, communication is in progress
slave pulls handshaking line low
slave disables bus interrupts
slave outputs master's address bit by bit on data bus
slave performs bus arbitration as described above
if slave wins arbitration, it enables bus interrupts
master responds by disabling bus interrupts
slave outputs its address bit by bit on the bus
slave performs bus arbitration as described above
if slave wins arbitration, it toggles the handshaking line
slave sets data byte
slave toggles clock
repeat data/clock until all data is sent
slave releases handshaking line
master releases DBI line
Slave0 Requests Slave1 for Communication
slave0 checks if handshaking or DBI lines are low
if low, communication is in progess
slave0 pulls handshaking line low
slave0 disables bus interrupts
slave0 outputs slave1's address bit by bit on the data bus
slave0 performs bus arbitration as described above
if slave0 wins arbitration, it enables bus interrupts
slave1 responds by disabling bus interrupts
slave0 outputs its address bit by bit on the bus
slave0 performs bus arbitration described above
if slave0 wins arbitration, it toggles the handshaking line
slave0 sets data byte
slave0 toggles clock
repeat data/clock until all data is sent
slave0 releases handshaking line
slave1 releases DBI line
Bona Fide Work • 0 Comments/Trackbacks • Posted by: J Silverman |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
serialShift on the way!
Well, after a few issues, I finally got my serialShift project being manufactured by Olimex. It took a couple of international faxes (wonder how much the phone bill this month is going to be :>) and a good couple of emails, but so far I'm pleased with Olimex. I'll be getting 12 copies of my board so I'll have quite a few boards to sell. If you'd like one, just email me or wait and see for a post saying I'm selling them on eBay.
I also finally got my Motorola 68000 processor books and have been trying to read them in between helping my parents with re-setting up a new inventory system for their store. The m68k is a really interesting processor, compared to the stuff I've seen before. The addressing modes alone are quite a bit different. I'm looking forward to building the SBC and being able to test out the m68k.
G1Technology • 0 Comments/Trackbacks • Posted by: J Silverman |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
First Post of the Year
Happy 2006 everyone! This new year should be quite interesting. I, of course, have more college to do, but I'm hoping to really get some stuff for sale at G1Technology.com. With the serialShift, serialOUT (new name for the powerSOD), and a m68k SBC, I'll have some fairly good things to sell. I think I'll use eBay first to judge the demand before ordering large quantities. More info on this to come. . .
G1Technology • 0 Comments/Trackbacks • Posted by: J Silverman |
|
|
|
|
|
|
|
|
|
|
|
December 31, 2005, 5:09 pm |
|
|
|
|
|
|
|
|
Happy New Year!
Happy New Year everyone! It's been one crazy year and maybe next year won't be too bad. Doubt it though.
Anyway, I've been planning another (yes, another) project which I think is really neat. I'm going to build a computer, from the component level and up. I've already ordered four 68000 processors which is the same processor found in the original Mac. I also ordered a couple of books about building SBCs using a m68k processor. It should be a fun project especially if I can get a form of Linux running on it. Btw, here's one already made:
http://chaokhun.kmitl.ac.th/~kswichit/68k/68k.html
Also, I found these couple of sites yesterday, which has got me quite excited:
http://www.applefritter.com/replica
http://www.brielcomputers.com/index.htm
A remade replica Apple 1! Thats just awesome. I'd love to build one of those. Now I just need to find someone to pay for the kit ;-)
G1Technology • 0 Comments/Trackbacks • Posted by: J Silverman |
|
|
|
|
|
|
|
|
|
|
|
December 30, 2005, 4:56 pm |
|
|
|
|
|
|
|
|
|
|
|
December 15, 2005, 12:05 am |
|
|
|
|
|
|
|
|
Another semester almost over. . .
Well, its finals time again, and in a few days I'll finally finish this very busy semester. I just finished my first final of the week and have three more till Saturday (yea, they're making us have Saturday finals). I've also made some progress on projects for G1Technology, which I'll explain later.
And I'm now also the President of our local ACM chapter, yay! :-)
More News · Bona Fide Work • 0 Comments/Trackbacks • Posted by: J Silverman |
|
|
|
|
|
|
|
|
|
|
|
December 3, 2005, 7:41 pm |
|
|
|
|
|
|
|
|
The powerSOD is coming soon
The powerSOD, power Serial Output Device, (thanks aquatix for the name ;-) will be finished soon and become G1Technology's first product for sale. It will be essentially a serial lcd device but also include a bank of 8 LEDs, and an extra 8 pin port ready for port expansion. The first version will feature an enhanced 8051 microcontroller from Dallas Semiconductor running at 11.057 mhz that will execute one command per clock cycle, giving around 11 MIPS. All necessary pins for in-circuit and in-application programming will be brought out to the main headers which can be plugged into a breadboard for easy testing. In addition to that, the pins that control the 16x2 character LCD will be brought out to the headers to give as much control of the LCD to the user as possible.
The powerSOD's firmware is still in development, but at this point, I can say it will give full control of the LCD, LEDs, and the extra port to the user with an easy to use API. The firmware source code and the schematic will be made available when I complete the project. The final feature set might differ some from this description as I finish the design. If you would like to join the development, please email me at sales@g1technology.com.
G1Technology • 0 Comments/Trackbacks • Posted by: J Silverman |
|
|
|
|
|
|
|
|
|
|
|
November 24, 2005, 4:06 am |
|
|
|
|
|
|
|
|
Happy Thanksgiving!
For all in the US, have a happy Thanksgiving weekend! For all those outside the US, have a great Morphest 2005 weekend!
Anyway, since it's a national holiday in this country, I get a nice four day weekend to really work on my electronics projects. Earlier this week, I had gotten an at89c51 chip, lcd screen, a few ZIF sockets, and a 20 to 40 pin adapter for my programmer, and I've been playing with the full 40 pin at89c51 chip ever since I got it. I'm working on creating my own serial lcd module circuit (with a few additions, like a bank of LEDs). This is so I don't have to purchase expensive serial lcds, and I'll get a bit more function out of mine as well. My module should cost roughly $12 including all parts for a 16x2 display. Maybe I should sell some after I finish the design ;-)
More News · Bona Fide Work • 0 Comments/Trackbacks • Posted by: J Silverman |
|
|
|
|
|
|
|
|
|
|
|
November 15, 2005, 11:15 pm |
|
|
|
|
|
|
|
|
Just when I thought I could relax. . .
A line of thunderstorms laden with tornadic activity is approaching my house. And I thought maybe I could have a relaxful evening, bah!
Anyway, life's been crazy for me as usual, although college work has slowed for now. I only have a month to go, so it all be over soon. Maybe I can finally do some tinkering with my electronics projects once again. I'm hoping to play around coding my 8051 chips with assembly + C this time so I can get more efficient programs.
In other news, I've been following this Sony DRM CD rootkit story and must say this might finally put the greedy companies in line. There's been so much backlash against Sony for their rootkits that maybe some other companies might think twice before continuing DRM technology. We'll see what happens there.
Oh yea, just finished up teaching those accounting students excel for this semester. It was quite fun doing the quiz that's administered at the end of the course, as we didn't have enough machines for everyone. So luckily with some help from a friend of mine, we split the group into two labs. It still was a bit interesting to conduct.
One last thing (warning: this is mainly an inside joke running through our CS department) I must say here for anyone at my college reading this:
The Essence of Computer Science
1. Efficiency
2. Speed
3. Abstraction
4. Queue-like
5. Nibble
6. Accumulator
Aimless Stuff • 0 Comments/Trackbacks • Posted by: J Silverman |
|
|
|
|
|
|
|
|
|
|
Navigation
Search the Blog
Calendar
Friends who blog
Sections
