I myself (redundant no?) have little experience with ARMs and do not find them exiting at all (too many kinds on the block ?) but a MIPS processor ? you mean a real MIPS a R3xxx or R4xxx ? or a PIC32 ? that sounds like fun. real fun and real challenge.
I can send you the eagle files for both projects, no problem. Did you already check the wiki in my sig ?
Eagle is hard to learn. It took me almost a year to use it comfortably and not to think wtf ?
I have gobs of MIPS chips around here as I deal in SGI hardware. Unfortunately, many of them are bga.
There is also a 900Mhz R7000 that would be great to get working.
I checked your wiki and it was good to see that you are doing surface mount chips. Why not surface mount props?
I did not see the links to eagle files on your wiki.
Thanks,
Doug
The eagle files are not there. As the prototype has some bugs I decided to redo it correcting them and adding some extra functionality but it is a work in progress. I can send you them if you want. Send me a PM with an email address.
Edit: why not SMT props ? well, I have none. I only have 6 DIP propeller that I use in several projects and the user could put its own props and thus they would not be in the price. As the other ICs are cheaper... If I were to make a run of these boards I'd use SMTs.
I have a working video driver that can produce a nice 512x256 4 bit picture (see page at wikispaces in my sig) with a 5 MHz crystal or with a 12 MHz crystal (PLL8) that should work well with a 6 MHz and PLL16.
I found that when NVIDEOW is asserted due to the address and DS, the propeller needs around 110 ns to negate DTACK and that puts it well beyond the transition of S4 to S5, so it is not recognized then it misses the next write cycle. (If you do not know what I'm talking about, get the disassembly of the JS ROM and the circuit posted in the wiki). At 96 MHz, the prop should make it to assert DTACK on time for S4->S5, I'll test it tonight. I did not have a soldering iron nor a faster crystal at home yesterday :-(.
Another workaround would be to make the MC68008 slower... but I'd rather not do that. Maybe 7.3 or 7.4 MHz is enough.
This problem was not present in the 020 version because the minimum cycle is 3 clocks and DS is valid a bit earlier.
With all the modifications that the board underwent I may well do a new one with just CPLDs, a 68EC020 and 512K RAM if it fits in 10x8 cm...
The machine starts in 8 color 256x256 mode (but I only have 512x256 4 colors but the line length is the same) and asks for a key F1 or F2 for Monitor (512x256 4 colors the mode that this screenshot shows) and TV 256x256 8 colors.
The red and green dots are missed, by the propeller, writes to video RAM, even at 96 MHz it is not enough to negate DTACK in time for the S4->S5 transition. I think an automatic wait state generator for video writes may be needed... the prop is a bit slow. I can try with a 6.55 or 7.x MHz crystal... I'm ignoring DS for writes anyways.
Comments
I myself (redundant no?) have little experience with ARMs and do not find them exiting at all (too many kinds on the block ?) but a MIPS processor ? you mean a real MIPS a R3xxx or R4xxx ? or a PIC32 ? that sounds like fun. real fun and real challenge.
I can send you the eagle files for both projects, no problem. Did you already check the wiki in my sig ?
Eagle is hard to learn. It took me almost a year to use it comfortably and not to think wtf ?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Visit some of my articles at Propeller Wiki:
MATH on the propeller propeller.wikispaces.com/MATH
pPropQL: propeller.wikispaces.com/pPropQL
pPropQL020: propeller.wikispaces.com/pPropQL020
OMU for the pPropQL/020 propeller.wikispaces.com/OMU
There is also a 900Mhz R7000 that would be great to get working.
I checked your wiki and it was good to see that you are doing surface mount chips. Why not surface mount props?
I did not see the links to eagle files on your wiki.
Thanks,
Doug
The eagle files are not there. As the prototype has some bugs I decided to redo it correcting them and adding some extra functionality but it is a work in progress. I can send you them if you want. Send me a PM with an email address.
Edit: why not SMT props ? well, I have none. I only have 6 DIP propeller that I use in several projects and the user could put its own props and thus they would not be in the price. As the other ICs are cheaper... If I were to make a run of these boards I'd use SMTs.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Visit some of my articles at Propeller Wiki:
MATH on the propeller propeller.wikispaces.com/MATH
pPropQL: propeller.wikispaces.com/pPropQL
pPropQL020: propeller.wikispaces.com/pPropQL020
OMU for the pPropQL/020 propeller.wikispaces.com/OMU
Post Edited (Ale) : 8/11/2009 7:57:15 PM GMT
I have a working video driver that can produce a nice 512x256 4 bit picture (see page at wikispaces in my sig) with a 5 MHz crystal or with a 12 MHz crystal (PLL8) that should work well with a 6 MHz and PLL16.
I found that when NVIDEOW is asserted due to the address and DS, the propeller needs around 110 ns to negate DTACK and that puts it well beyond the transition of S4 to S5, so it is not recognized then it misses the next write cycle. (If you do not know what I'm talking about, get the disassembly of the JS ROM and the circuit posted in the wiki). At 96 MHz, the prop should make it to assert DTACK on time for S4->S5, I'll test it tonight. I did not have a soldering iron nor a faster crystal at home yesterday :-(.
Another workaround would be to make the MC68008 slower... but I'd rather not do that. Maybe 7.3 or 7.4 MHz is enough.
This problem was not present in the 020 version because the minimum cycle is 3 clocks and DS is valid a bit earlier.
With all the modifications that the board underwent I may well do a new one with just CPLDs, a 68EC020 and 512K RAM if it fits in 10x8 cm...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Visit some of my articles at Propeller Wiki:
MATH on the propeller propeller.wikispaces.com/MATH
pPropQL: propeller.wikispaces.com/pPropQL
pPropQL020: propeller.wikispaces.com/pPropQL020
OMU for the pPropQL/020 propeller.wikispaces.com/OMU
Here are a couple of screenshots.
The machine starts in 8 color 256x256 mode (but I only have 512x256 4 colors but the line length is the same) and asks for a key F1 or F2 for Monitor (512x256 4 colors the mode that this screenshot shows) and TV 256x256 8 colors.
The red and green dots are missed, by the propeller, writes to video RAM, even at 96 MHz it is not enough to negate DTACK in time for the S4->S5 transition. I think an automatic wait state generator for video writes may be needed... the prop is a bit slow. I can try with a 6.55 or 7.x MHz crystal... I'm ignoring DS for writes anyways.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Visit some of my articles at Propeller Wiki:
MATH on the propeller propeller.wikispaces.com/MATH
pPropQL: propeller.wikispaces.com/pPropQL
pPropQL020: propeller.wikispaces.com/pPropQL020
OMU for the pPropQL/020 propeller.wikispaces.com/OMU
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
http://www.propgfx.co.uk/forum/·home of the PropGFX Lite
·
Well, the circuit needs some changes so I'll program a CPLD and send you the extension too.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Visit some of my articles at Propeller Wiki:
MATH on the propeller propeller.wikispaces.com/MATH
pPropQL: propeller.wikispaces.com/pPropQL
pPropQL020: propeller.wikispaces.com/pPropQL020
OMU for the pPropQL/020 propeller.wikispaces.com/OMU
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
http://www.propgfx.co.uk/forum/·home of the PropGFX Lite
·