Good question Ding-Batty. I have no idea what "social groups" are for.
I have no idea why I should have to be a member of one in order view any content there.
I can perhaps imagine some subset of Parallax heads have a common problem. Like for example they live in Australia. In which case a place to discuss local meetings/happenings might be appropriate. Even that seems a bit contrived to me.
Jazzed is sort of co-opting social group as a project area.
Anyway, first thing is FORUM SEARCH DOES NOT FIND STUFF IN GROUPS. Try finding some terms in Jazzed's group description like "RAM-kind" or "external RAM design".
Contrived? As in planned? Lol. I was playing in what I thought was a deep section of the user control panel. Only later discovered its directly accessible from the "Community" drop down menu.
There is a definite need sometimes for individuals to form a small project team, and I'm always interested in the best way to collaborate. Google Wave showed some promise but they pulled the plug. I wonder whether the social groups here can be tailored to suit our project needs, because having these associated/adjacent to the forum is beneficial. Yes, I agree they should ideally be public and googlable, perhaps there is a way to flag the bots to trawl them (or perhaps vBulletin admin options need to change?).
Sorry Steve for venturing so far OT. Fwiw I like your DRAM concept and would like a board/kit/whatever if you have one spare. I also think its good to play with these things before Prop2, so we can hit the ground running (and truly appreciate Chip's efforts!)
Good question Ding-Batty. I have no idea what "social groups" are for.
I have no idea why I should have to be a member of one in order view any content there.
I can perhaps imagine some subset of Parallax heads have a common problem. Like for example they live in Australia. In which case a place to discuss local meetings/happenings might be appropriate. Even that seems a bit contrived to me.
Jazzed is sort of co-opting social group as a project area.
Anyway, first thing is FORUM SEARCH DOES NOT FIND STUFF IN GROUPS. Try finding some terms in Jazzed's group description like "RAM-kind" or "external RAM design".
Well I thought "social groups" was a feature we should try to use. Maybe someone will find a good fit for that later. If there is no value to having such a group, it's probably best to unsubscribe so I can delete it (edit/delete doesn't seem to work with subscribers).
Yes, a project area made sense to me ... kind of like a committee who's agenda doesn't necessarily make sense for everyone. I'm not really interested in Scribblers for example, but others really enjoy them.
On the image resolution ... I tried uploading the .png several times since I know .png is practically lossless. What ended up being posted appears to have been converted to .jpg since there's lots of "eddying" in the image. Slightly disturbing. I used a different color scheme since others mentioned that the one I like was hard to read.
Here's a link to the new schematic; I can't shrink it without losing detail. The VGA connector and TV/headphone connectors overlap in the FAB to provide a stuffing option. A board with no connectors can also be made.
There is a unique experimental option on the schematic that I don't think anyone has tried and this schematic spills the idea out there for anyone to use/try. Since it's there to see, it would be nice to know if anyone thinks there's a chance in heck that it can be made to work. I'm concerned about the voltage configuration ... any thoughts?
I'll be making 10 or so boards next on a 2 week turn and will reserve one for you. I can not guarantee lead-free parts on boards I assemble.If these boards are produced in volume for Gadget Gangster for example, lead-free parts will be not be an issue.
Look at the schematic and let me know what options you want stuffed. I'm not sure if I'll have any I2C keyboard/mouse support ready by the time the boards ship but I'll try.
Thanks Steve. No problem with lead or without. As for options I'm really just interested in the SDRAM. I can borrow an atmel programmer later and get the kb/ms over i2c running if needed.
cheers
tubular
With ZOG comfortably running on 32MB SDRAM, it's time to update the hardware.
The 32MB Gadget Gangster module should provide a 2 board Propeller Computer solution on any Propeller Platform board.
After chatting with Nick about the Gadget Gangster platform boards, I've decided to go with a *long-board* so that I can fit a micro-SD card and two separate mini-dins for keyboard/mouse on the design. The TV option will be used for the Propeller Computer, the VGA is a stuffing option for hi-resolution textured video experiments.
Here's the feature-set:
Memory: 32MB SDRAM available for ZOG or other compilers.
Language compilers must provide at least 25 bit addressability.
Video: TV & SDCARD (or VGA without SDCARD).
Audio: Optional TV audio headphone connector (0.25" or 0.125" possible).
SDCARD: Push-push type on the right pins - no hardware hacks required.
Ethernet: SDCARD pins allow optional Gadget Gangster E-Net module use.
Ethernet or SneakerNet your choice
Keyboard: Accesible via I2C and atTiny85 AVR.
Mouse: Accesible via I2C and atTiny85 AVR.
AVR: DIP-8 Allows for custom code connecting to Keyboard/Mouse port pins.
Optional applications include I2C interface A/D conversion on one or both ports.
Input or feedback on the design is welcome!
Other board design ideas are also in process.
One that makes sense for me at least is to have a Propeller execution processor module that will free almost all the Propeller Platform motherboard pins for other purposes. Such a design is feasible because of the enormous work RossH did with Catalina on Morpheus and Cluso's Propeller Blades.
Another option is to create a slower SDRAM board that allows more pins for 1 Propeller (add another latch).
Additionally, a 100MHz Propalyzer PC-Propeller Logic Analyzer module is planned and could potentially be used with any PC application including Viewport.
Finally, I plan to produce "instructables" for the adding I2C Keyboard/Mouse AVR design to other hardware solutions such as the and VGA-A/V board.
One that makes sense for me at least is to have a Propeller execution processor module that will free almost all the Propeller Platform motherboard pins for other purposes. Such a design is feasible because of the enormous work RossH did with Catalina on Morpheus and Cluso's Propeller Blades.
Another option is to create a slower SDRAM board that allows more pins for 1 Propeller (add another latch).
Additionally, a 100MHz Propalyzer PC-Propeller Logic Analyzer module is planned and could potentially be used with any PC application including Viewport.
Finally, I plan to produce "instructables" for the adding I2C Keyboard/Mouse AVR design to other hardware solutions such as the and VGA-A/V board.
Its looking very good, jazzed. I thought it was looking short on resistors around the KB and Mouse but thats because of the avr running at 5 volts. One minor addition would be to put a USB A footprint as well as PS2, since you have room. You can get USB female A's that hang over the side of the board (so their shell wouldn't short anything)
One minor addition would be to put a USB A footprint as well as PS2, since you have room. You can get USB female A's that hang over the side of the board (so their shell wouldn't short anything)
Tubular, can you explain how the USB connectors would be used?
the USB A's can be used with the keyboards and mice that support dual PS2/USB mode. Ie it eliminates the clumsy PS2 to USB adapter required with many modern keyboards or mice.
Here's a photo of a board edge USB A female connector
I'm debugging the I2C Keyboard/Mouse interface chip.
Unfortunately, my development progress on the I2C will stop as of this evening for a few weeks while I attend to family business. During this time, though several boards will be built thanks to Nick.
Sorry, maybe I will ask a stupide question, but...
Is SDRAM faster than Micro-SD?
If yes, how much faster?
Also, how many PINs will this board use? I only have 5 PINs remaining on my project...
Last question
If I don't need VGA, Keyboard and Mouse, can I still use those dedicated PINs for my project? Or they are already connected to other pasive components which might interfer with what I will connect?
So look like I can forget it I can free up to 8 pins, but no more. So I will have to look at another option...
JM
That's a fair summary. I'm designing a new keyboard/mouse interface using I2C just so we can have that. SDRAM will be much faster than SDCARD by about 10x.
What's the number of pin required for this? If I don't need the keyboard/mouse, but just the SDRAM, what's the minimum? I'm looking for some kind of memory, faster than SD cards, not necessay big (1Mo is enough for my needs), which does'nt require more than 8 pins. Do you know what I can use?
What's the number of pin required for this? If I don't need the keyboard/mouse, but just the SDRAM, what's the minimum?
An 8 bit data bus interface requires more than 8 pins for control signals like latch enable (1 or more depending on the number of latches), and a write enable for storing data.
Other slower designs can be done. If you only need access to small portions of memory at any time, a cache or virtual interface would deliver what you need on SPI RAM for example with serial SPI interfaces.
Since I found a little time to relax from my heavy out of the ordinary personal schedule, I thought I would answer your question about pin usage better.
The SDRAM interface requires 20 total Propeller pins. That leaves 2 for serial download and debug, 2 for I2C, and 8 pins free. The board Nick shows is a variation that allows for a Propeller Computer with any of the Gadget Gangster Propeller Platform boards. The 8 free Propeller pins when using only SDRAM are P16-23 and are available on the connector near the SDCARD footprint.
Hope that helps answer your question a little more.
--Steve
Yep, thanks. It's really to much for my application. I have 10 pins available, and still need to add an FM receiver. So I will try to reduce the number of used pins.
Can you point me to some memory I can use with 10 pins? It need to be faster than the SDcard... I just need 1Mo.
Yep, thanks. It's really to much for my application. I have 10 pins available, and still need to add an FM receiver. So I will try to reduce the number of used pins.
Can you point me to some memory I can use with 10 pins? It need to be faster than the SDcard... I just need 1Mo.
(Sorry Steve - did not mean to hijack your thread!)
Yes, FRAM's are non-volatile.
- 1MB FlexMem kit with four 256Kx8 FRAM's would be around $150+s/h
- 512KB FlexMem kit with four 128Kx8 FRAM's (SOIC-8's) would be around $75+s/h
Up to three FlexMem's can be stacked on P0-P7
I just checked, the SOIC-8 256Kx8 parts are not available yet, only TDFN-8 in the 256Kx8 (2Mbit) configuration. My cost will be around $20-$25/chip when available in SOIC-8. I'd have to check to see if I can get away with puttting TDFN on FlexMem.
I'd have to check what the lead time is on the chips, or charge more and get them from Mouser.
Jazzed anyway you can enlarge the schematic for some reason I can not
Sure. Is this better?
In Google Chrome I use Ctrl++ [press and hold Ctrl and + at the same time] and the web page content gets bigger. Ctrl+- makes page content smaller. I looked around and it seems Firefox and Internet Explorer also support Ctrl++ (and for those who use Opera, I'm sure there's a way to do it there too). I don't know how clicking the attachment would work ... it just seems to show a box with an X-it control.
In Google Chrome I use Ctrl++ [press and hold Ctrl and + at the same time] and the web page content gets bigger. Ctrl+- makes page content smaller. I looked around and it seems Firefox and Internet Explorer also support Ctrl++ (and for those who use Opera, I'm sure there's a way to do it there too). I don't know how clicking the attachment would work ... it just seems to show a box with an X-it control.
The schematic on page 1 is out of date. I'm not sure if I'll be posting another one any time soon. My priority is to get the latest hardware tested and into production.
The schematic on page 1 is out of date. I'm not sure if I'll be posting another one any time soon. My priority is to get the latest hardware tested and into production.
Hi Doug (@hinv). Here are answers and a matrix of possible configurations.
How is your board coming along?
The SDRAM design works fine. I'm still developing the I2C mouse/keyboard section - sometimes I wish this was a separate module, but I think it's important for users who want a Propeller-PC with gobs of memory. The mouse works great, but I don't have the keyboard fully functional yet.
What is the top speed?
The highest burst speed is 10MB/s (note 1).
A 64 byte page access speed is 5.3MB/s (note 2).
All performance measured on an 80MHz clock system.
Memory Performance Notes:
1. The top speed is 10MB/s. That is the rate at which bytes are read/written in a single burst after an address phase. That's "putting the best foot forward" as some people say
2. The most useful and realistic measure is page write or read speed. A 64 byte page write or read clocks in at 5.3MB/s or 64 bytes every 12us. The longer the page, the faster the access. Page access performance is useful primarily because of caching as I've implemented for ZOG. I have a VMCOG version also, but it is slower because there's no room for fast page brusts; VMCOG also can not support 32MB.
What is the turn around time for a byte, word, long?
Byte at a time access speed is 0.5MB/s.
Dreadful of course When forced to read or write one byte, word, or long at a time, access would be about 0.5MB/s for 1 byte, near 0.9MB/s for a word, and near 1.6MB/s for a long (any of these accesses will happen in about 2us). Considering the setup time is the prevailing factor, the time it takes for any single access is about the same. The more you access at one time the higher the throughput.
What are some possibilities?
Nick and I are building more boards this week. The common assembly will probably be a board that supports 32MB SDRAM. Different flavor boards can be built according to instruct-able documents and may include one of the following configurations:
Configuration Notes:
1) 8MB to 64MB SDRAM is possible, but the only shipping configuration will be 32MB (Development/testing of other sizes is negotiable).
2) VGA can not be used with SDCARD, Ethernet, TV+SDCARD, or TV+Ethernet.
3) An I2C attached up to 3 single end channels 5V range ADC is possible instead of Keyboard/Mouse - can have 1 differential channel ADC.
4) Theoretically possible with a new LCD module board to be designed.
8 Propeller pins are free for any use when only SDRAM is attached.
Note: At this time the I2C Keybd and I2C Mouse are not "officially supported" on the SDRAM board.
Keyboard and Mouse working simultaneously requires the RESET pin to be set permanently to I/O.
A forthcoming boot-loader will make reprogramming possible regardless of the RESET pin feature.
Once the boot-loader is available, the keyboard and mouse will be officially supported.
Comments
I have no idea why I should have to be a member of one in order view any content there.
I can perhaps imagine some subset of Parallax heads have a common problem. Like for example they live in Australia. In which case a place to discuss local meetings/happenings might be appropriate. Even that seems a bit contrived to me.
Jazzed is sort of co-opting social group as a project area.
Anyway, first thing is FORUM SEARCH DOES NOT FIND STUFF IN GROUPS. Try finding some terms in Jazzed's group description like "RAM-kind" or "external RAM design".
There is a definite need sometimes for individuals to form a small project team, and I'm always interested in the best way to collaborate. Google Wave showed some promise but they pulled the plug. I wonder whether the social groups here can be tailored to suit our project needs, because having these associated/adjacent to the forum is beneficial. Yes, I agree they should ideally be public and googlable, perhaps there is a way to flag the bots to trawl them (or perhaps vBulletin admin options need to change?).
Sorry Steve for venturing so far OT. Fwiw I like your DRAM concept and would like a board/kit/whatever if you have one spare. I also think its good to play with these things before Prop2, so we can hit the ground running (and truly appreciate Chip's efforts!)
regards
tubular
Yes, a project area made sense to me ... kind of like a committee who's agenda doesn't necessarily make sense for everyone. I'm not really interested in Scribblers for example, but others really enjoy them.
On the image resolution ... I tried uploading the .png several times since I know .png is practically lossless. What ended up being posted appears to have been converted to .jpg since there's lots of "eddying" in the image. Slightly disturbing. I used a different color scheme since others mentioned that the one I like was hard to read.
Here's a link to the new schematic; I can't shrink it without losing detail. The VGA connector and TV/headphone connectors overlap in the FAB to provide a stuffing option. A board with no connectors can also be made.
http://www.brouhaha.com/~sdenson/propeller/sdram8232_3.png
There is a unique experimental option on the schematic that I don't think anyone has tried and this schematic spills the idea out there for anyone to use/try. Since it's there to see, it would be nice to know if anyone thinks there's a chance in heck that it can be made to work. I'm concerned about the voltage configuration ... any thoughts?
BTW, does anyone not remember Bob Hope?
TIA,
--Steve
I'll be making 10 or so boards next on a 2 week turn and will reserve one for you. I can not guarantee lead-free parts on boards I assemble.If these boards are produced in volume for Gadget Gangster for example, lead-free parts will be not be an issue.
Look at the schematic and let me know what options you want stuffed. I'm not sure if I'll have any I2C keyboard/mouse support ready by the time the boards ship but I'll try.
Thanks,
--Steve
cheers
tubular
With ZOG comfortably running on 32MB SDRAM, it's time to update the hardware.
The 32MB Gadget Gangster module should provide a 2 board Propeller Computer solution on any Propeller Platform board.
After chatting with Nick about the Gadget Gangster platform boards, I've decided to go with a *long-board* so that I can fit a micro-SD card and two separate mini-dins for keyboard/mouse on the design. The TV option will be used for the Propeller Computer, the VGA is a stuffing option for hi-resolution textured video experiments.
Here's the feature-set:
Memory: 32MB SDRAM available for ZOG or other compilers. Language compilers must provide at least 25 bit addressability.
Video: TV & SDCARD (or VGA without SDCARD).
Audio: Optional TV audio headphone connector (0.25" or 0.125" possible).
SDCARD: Push-push type on the right pins - no hardware hacks required.
Ethernet: SDCARD pins allow optional Gadget Gangster E-Net module use. Ethernet or SneakerNet your choice
Keyboard: Accesible via I2C and atTiny85 AVR.
Mouse: Accesible via I2C and atTiny85 AVR.
AVR: DIP-8 Allows for custom code connecting to Keyboard/Mouse port pins. Optional applications include I2C interface A/D conversion on one or both ports.
Input or feedback on the design is welcome!
Other board design ideas are also in process.
One that makes sense for me at least is to have a Propeller execution processor module that will free almost all the Propeller Platform motherboard pins for other purposes. Such a design is feasible because of the enormous work RossH did with Catalina on Morpheus and Cluso's Propeller Blades.
Another option is to create a slower SDRAM board that allows more pins for 1 Propeller (add another latch).
Additionally, a 100MHz Propalyzer PC-Propeller Logic Analyzer module is planned and could potentially be used with any PC application including Viewport.
Finally, I plan to produce "instructables" for the adding I2C Keyboard/Mouse AVR design to other hardware solutions such as the and VGA-A/V board.
Cheers.
--Steve
Thanks.
--Steve
Here's a photo of a board edge USB A female connector
The 32MB SDRAM, SDCARD, and TV DAC work fine.
I'm debugging the I2C Keyboard/Mouse interface chip.
Unfortunately, my development progress on the I2C will stop as of this evening for a few weeks while I attend to family business. During this time, though several boards will be built thanks to Nick.
--Steve
Full Size
Sorry, maybe I will ask a stupide question, but...
Is SDRAM faster than Micro-SD?
If yes, how much faster?
Also, how many PINs will this board use? I only have 5 PINs remaining on my project...
Last question
If I don't need VGA, Keyboard and Mouse, can I still use those dedicated PINs for my project? Or they are already connected to other pasive components which might interfer with what I will connect?
Thanks,
JM
SDRAM is faster than microSD. I don't know how much, though.
SDRAM alone will require more than 5 pins. This module with everything connected uses almost all of the Propeller pins.
Thanks for your reply.
So look like I can forget it I can free up to 8 pins, but no more. So I will have to look at another option...
JM
That's a fair summary. I'm designing a new keyboard/mouse interface using I2C just so we can have that. SDRAM will be much faster than SDCARD by about 10x.
--Steve
What's the number of pin required for this? If I don't need the keyboard/mouse, but just the SDRAM, what's the minimum? I'm looking for some kind of memory, faster than SD cards, not necessay big (1Mo is enough for my needs), which does'nt require more than 8 pins. Do you know what I can use?
Thanks,
JM
Other slower designs can be done. If you only need access to small portions of memory at any time, a cache or virtual interface would deliver what you need on SPI RAM for example with serial SPI interfaces.
Since I found a little time to relax from my heavy out of the ordinary personal schedule, I thought I would answer your question about pin usage better.
The SDRAM interface requires 20 total Propeller pins. That leaves 2 for serial download and debug, 2 for I2C, and 8 pins free. The board Nick shows is a variation that allows for a Propeller Computer with any of the Gadget Gangster Propeller Platform boards. The 8 free Propeller pins when using only SDRAM are P16-23 and are available on the connector near the SDCARD footprint.
Hope that helps answer your question a little more.
--Steve
Can you point me to some memory I can use with 10 pins? It need to be faster than the SDcard... I just need 1Mo.
Any idea?
Thanks,
JM
You might also take a look at Cluso99's RamBlade.
Yes, FRAM's are non-volatile.
- 1MB FlexMem kit with four 256Kx8 FRAM's would be around $150+s/h
- 512KB FlexMem kit with four 128Kx8 FRAM's (SOIC-8's) would be around $75+s/h
Up to three FlexMem's can be stacked on P0-P7
I just checked, the SOIC-8 256Kx8 parts are not available yet, only TDFN-8 in the 256Kx8 (2Mbit) configuration. My cost will be around $20-$25/chip when available in SOIC-8. I'd have to check to see if I can get away with puttting TDFN on FlexMem.
I'd have to check what the lead time is on the chips, or charge more and get them from Mouser.
How is your board coming along?
What is the top speed?
What is the turn around time for a byte, word, long?
Thanks,
Doug
Sure. Is this better?
In Google Chrome I use Ctrl++ [press and hold Ctrl and + at the same time] and the web page content gets bigger. Ctrl+- makes page content smaller. I looked around and it seems Firefox and Internet Explorer also support Ctrl++ (and for those who use Opera, I'm sure there's a way to do it there too). I don't know how clicking the attachment would work ... it just seems to show a box with an X-it control.
I think He mean SCH on page 1
That is OK for me.
I not need it -- If needed can made my one.
How is your board coming along?
The SDRAM design works fine. I'm still developing the I2C mouse/keyboard section - sometimes I wish this was a separate module, but I think it's important for users who want a Propeller-PC with gobs of memory. The mouse works great, but I don't have the keyboard fully functional yet.
What is the top speed?
The highest burst speed is 10MB/s (note 1).
A 64 byte page access speed is 5.3MB/s (note 2).
All performance measured on an 80MHz clock system.
Memory Performance Notes:
1. The top speed is 10MB/s. That is the rate at which bytes are read/written in a single burst after an address phase. That's "putting the best foot forward" as some people say
2. The most useful and realistic measure is page write or read speed. A 64 byte page write or read clocks in at 5.3MB/s or 64 bytes every 12us. The longer the page, the faster the access. Page access performance is useful primarily because of caching as I've implemented for ZOG. I have a VMCOG version also, but it is slower because there's no room for fast page brusts; VMCOG also can not support 32MB.
What is the turn around time for a byte, word, long?
Byte at a time access speed is 0.5MB/s.
Dreadful of course When forced to read or write one byte, word, or long at a time, access would be about 0.5MB/s for 1 byte, near 0.9MB/s for a word, and near 1.6MB/s for a long (any of these accesses will happen in about 2us). Considering the setup time is the prevailing factor, the time it takes for any single access is about the same. The more you access at one time the higher the throughput.
What are some possibilities?
Nick and I are building more boards this week. The common assembly will probably be a board that supports 32MB SDRAM. Different flavor boards can be built according to instruct-able documents and may include one of the following configurations:
Configuration Notes:
1) 8MB to 64MB SDRAM is possible, but the only shipping configuration will be 32MB (Development/testing of other sizes is negotiable).
2) VGA can not be used with SDCARD, Ethernet, TV+SDCARD, or TV+Ethernet.
3) An I2C attached up to 3 single end channels 5V range ADC is possible instead of Keyboard/Mouse - can have 1 differential channel ADC.
4) Theoretically possible with a new LCD module board to be designed.
8 Propeller pins are free for any use when only SDRAM is attached.
Note: At this time the I2C Keybd and I2C Mouse are not "officially supported" on the SDRAM board.
Keyboard and Mouse working simultaneously requires the RESET pin to be set permanently to I/O.
A forthcoming boot-loader will make reprogramming possible regardless of the RESET pin feature.
Once the boot-loader is available, the keyboard and mouse will be officially supported.
Hope that helps. Cheers.
--Steve