I am pleased to be able to report that the production Morpheus PCB rev.2 has passed all tests with flying colors!
Note: the NTSC video and SVideo build options won't be tested until I build a board configured for that setup - but my 256 color VGA drivers worked just fine.
Mem+ is also almost fully tested, only the SD card interface and the MCP23S17 I/O ports remain to be tested.
I hope to make them available for sale by Monday next week.
This sounds intriguing. Can you please post a description of this for dummies like me?
In particular, are you running external memory for the video buffer? If so, how are you getting the speeds, are you using serial or parallel memory, and are you reading in the data alternate lines at a time?
I did some back-of-an-envelope calculations and I wasn't sure that external video ram was possible. Are you overclocking as well?
Yes, I am using the external memory for the video frame buffer. Catalina also supports it for LARGE programs (thanks Ross!)
You are correct, my drivers do implement two scan-line sized buffers in the hub for the bitmap data.
I designed a parallel bus, using parallel SRAM's that allows up to 16MB of ram.
Here is a list of my currently available Morpheus video drivers - all are fully bitmapped!
- 640x480 with 4 colors per pixel from 256 color palette
- 800x600 with 4 colors per pixel from 256 color palette
- 1024x768 with 4 colors per pixel from 256 color palette
- 256x192 (with 1024x768 timing) with 256 colors per pixel
Coming soon:
- 800x480 with 4 colors per pixel from 256 color palette (for the OEM 7" WVGA/video/svideo module I sell)
- 320x240 with 256 colors per pixel, with sprites (255 colors + transparent) per pixel, the driver supports horizontally splitting the screen in half *with pixel accurate horizontal scrolling* and also pixel accurate independent vertical scrolling
NOTES:
1) All 4 color modes can have a unique palette for each scan line
2) All of my drivers have a common API that provides easy to use API calls for manipulating the palette, point plot, line drawing, rectangles and circles (which may optionally be filled)
Hub-only text driver:
- 128x64/100x50/80x40 RossH's port to Morpheus of Chip's high-rez text driver
My design supports 20MB/sec (MegaBytes per sec) burst access with up to 2.5MB of ram, I need to use faster SOJ ram's to get 20MB/sec with more than 2.5MB due to increased bus capacitance (which explains why I also designed and sell SOJ32->DIP32 PCB adapters)
While I have not tested it yet, I expect that I can get 25MB/sec by using faster memory and my 100MHz crystals - once that is verified, I expect to start selling "Turbo Morpheus" - that is Morpheus with faster memory and 6.25Mhz crystals.
Please find the feature lists below:
[code]
Morpheus Features (rev.2 PCB)
CPU #1:
* “Designed for Largos” which has its root file system on an SPI flash chip and has an SPI ram chip for an edit buffer
* I2C boot EEPROM
* I2C Phillips PCF8563 Real Time Clock
* P0-P7 available for I/O expansion on ProtoBoard compatible 10 pin header
* Compatible with SpinStudio modules from ucontroller.com except for the “Game Stacker”
* P8-P11 normally used for 32KB SRAM or FRAM, required for Largos
* P12-P15 for second 32KB SRAM or FRAM, or WinBond or compatible 4K page erasable FLASH, available up to 8MB (64Mb) in capacity, required for Largos - VMCOG supported
* P8-P15 may optionally be used to provide a second ProtoBoard compatible header if the SPI memory is not populated
* PS/2 Keyboard support (demo board compatible)
* PS/2 Mouse support (demo board compatible)
* Line level stereo audio out (may be used as two DAC’s for other purposes)
* DC input via pads for a DC 2.1mm/5mm jack
* space for 5V voltage regulation based on 7805
* space for 3.3V regulation based on LM1117 or AS2810 3.3V regulators
* Dual filter capacitor support (standard .1uF ceramic, and tantalum/electrolytic capacitor support as per Saphiea’s findings)
* 4 bit bus connecting to CPU #2
* controls RESET line of CPU #2
* can optionally drive XINPUT of CPU #2 from dedicated I/O pin instead of crystal (dial-a-cpu-frequency)
* PropPlug compatible programming header, adds optional fifth pin for 3.3V
CPU #2:
* 4 bit bus to CPU #1 on P28-P31
* normally loaded by CPU #1
* optional Prop plug compatible 5 pin programming header
* optional EEPROM socket for independent boot
* 512KB 32 pin DIP SRAM socket
* 24 bit address / 8 bit data FULL EXPANSION BUS ON 40 PIN HEADER
* VGA support, build time option for 64 color Parallax compatible mode or extended 256 color mode
* working 1024
The new "Blue" (rev.2 pcb) Morpheus and Mem+ boards are in stock.
(I still have to update the BOM, tech docs, and write a new build manual)
PCB prices will stay the same as before, however kit and assembled&tested prices will have to go up due to the additional parts used for the new features.
I also have my new 485Plug (blue) production boards in stock, I will be making them available for sale as soon as I make the BOM and figure out pricing.
In other news, the following boards are going in for production runs later this week:
- DIG-8-IN-V ... opto isolated 8 input nominally 24VDC input, 9 screw terminals
- DIG-8-OUT-V ... Solid state relay based 8 output board, nominally 24VDC but will switch up to 60V @ 500mA, 9 screw terminals
- CPUModule ... tiny through hole Propeller module
- Propteus (pcb rev.2) ... new, blue, version of the original Propteus
- miniPro (pcb rev.2) ... new, small prototyping board for the 10/11 pin headers
Sapieha did his usual excellent work, the boards are practically works of art! His PCB layouts rock.
He also made some excellent additions to the new prototype boards... and CPUModule is the result of combining my ideas with his for a small processor module. He also added some nice touches to Morpheus/Mem+ - namely optional battery backup of ram, soic8/tdfn8 patterns for SPI memory chips, and I probably forgot somethings
The new "PLC" board is also going in, hopefully for production, however I am only making a small number of boards just in case another board revision needs to be made. This board has a ton of features, which I will only talk about after the boards are 100% tested and functional as the first prototype needed some modifications.
Sorry for the delays guys, but I was unable to spend as much time on Propeller projects after UPEW as I'd have liked.
More boards will follow the above :-)
FlexMem driver work will resume shortly; I will modify one of the prototypes, as Sapieha suggests adding a resistor between MISO and MOSI - perhaps my problem with data corruption has been due to not having that resistor; it makes a certain amount of sense, as the corruption *always* happens on the last nybble transferred. Then back to VMCOG, writing v2.0 (Large VM) version...
FlexMem driver work will resume shortly; I will modify one of the prototypes, as Sapieha suggests adding a resistor between MISO and MOSI - perhaps my problem with data corruption has been due to not having that resistor; it makes a certain amount of sense, as the corruption *always* happens on the last nybble transferred.
Some manufacturers recommend that BUT it was not only reason I used that!.
I have intensively tested SRAM's 23k256 on production version of Morpheus (We have possibility to 2 that ones.
Without resistors it was not problem run tests with slow drivers BUT in time I started RUNNING Fast driver 20/10MHz that one that use counter it started have unpredictable errors.
As this Driver was configured to use only one line for write/read I prepared one of IC's to that by lifting DI pin and connected it to DO output pin BUT in ViewPort it showed some Reading errors.
So I read all possible info from many manufacturers on connecting DI to DO and some of them recommend that RESISTOR (PropPin ->DI->100R->DO on SPI IC) so I added that to my next Test and my READING problems ware gone.
After that all speed test not have any writing/reading problems (Tested with 5MHz, 6.25MHz and 14.381.180MHz PLL8x) without adjusting counter values to this driver and that give with 14.xxMHz frequency some extra MHz on writing/reading that SRAM's.
Thank you! I will post some pics and more details on CPUModule later today.
Cluso99:
Thank you!
Kuronko:
Sapieha replied before I did :-)
I totally understand you questioning the need for the resistor; frankly I don't think it should be needed, however I trust Sapieha's test results. (Ask Sapieha how long it took him to convince me that the resistor could make a difference)
Based on Sapieha's test results, my best guess is that the 23K256's handling of tristating (and enabling) MISO (DO) after receiving the read command and address may not be up to the 25Mbps data rate of a 100Mhz or faster Prop using the optimal counter based drivers.
The data sheet states that Fclk is 20MHz max, so running the 23K256 at 25MHz+ is out of spec.
What is not available is an updated build manual for the nice new blue PCB's (Sapieha did an amazing job, as usual)
The new boards work well, and you can see them at UPEW.
For those of you skilled with a soldering iron and reading schematics, you should have a pretty easy build, following the silk screening and schematics; with a glance at the old build manual once in a while - the circuit has not changed (except for some new features).
I am afraid I don't have time to offer full kits or assembled and tested boards until after UPEW, but if some wants a Morpheus, Mem+ or combo before then, I will be making one or two mailing runs to the post office before going to UPEW. (I also have the stacking and non-stacking connectors the boards use). The power switch and 1/8" stereo jack are a bit unique, and I have some stock of those.
Actually, I just ordered more Morpheus and Mem+ (pcb rev2) boards, so I'll have more stock in time for UPEW.
Click on the Morpheus link in my signature, it will take you to the product page on the site, with pics of the old PCB's. I'll try to update the pics and pricing before UPEW, but right now I am working on demo's and the last of the new boards for UPEW ... (Sapieha again made some crazy nice layouts - but I won't announce the new designs until I test them - and no, they are not more variations of Morpheus or PropCade)
The new boards work well, and you can see them at UPEW.
Darn! I wish I could come. That is the worst possible weekend for me. My daughter is graduating from college and I have a choral concert to sing in. Maybe Parallax could postpone it for a week? :-)
Anyway, good luck on the introduction of your new version of Morpheus!
Congrats on your daughters's graduation! Of course that is more important,.
After UPEW, I will have enough of a product line that I will actually start to advertise... as in google and other places :-)
There are always videos of UPEW, and product announcements will be posted by everyone - but it is a GREAT networking event, it will be my third in three years.
Hopefully we'll meet at UPEW 2012.... OBC is liable to start the thread for it the day after UPEW 2011 (OBC, teasing)
Darn! I wish I could come. That is the worst possible weekend for me. My daughter is graduating from college and I have a choral concert to sing in. Maybe Parallax could postpone it for a
week? :-)
Anyway, good luck on the introduction of your new version of Morpheus!
Ah, Morpheus! I haven't heard about that for a while. Your video looks pretty cool although the text scrolling looks kind of jerky. Is that a YouTube artifact?
I have the rev2 boards ready; I just have to write an addendum to the build manual to cover the changes - mind you, all experienced builders will need is the schematic.
The text scrolling is a bit jerky, but given that the bitmap is being scrolled in 512KB of external ram, its not half bad
While the documentation is for rev1.0, with the rev2.0 schematics intermediate to advanced propeller users should have no trouble building it
I've cut the prices for the bare boards, and if I sell out, I can get more boards made.
Note: the 512Kx8 DIP32 SRAM is has long lead times, but the pin-compatible 128Kx8 is still fairly readily available, and I also have some SOJ-DIP adapter PCB's.
Comments
Note: the NTSC video and SVideo build options won't be tested until I build a board configured for that setup - but my 256 color VGA drivers worked just fine.
Mem+ is also almost fully tested, only the SD card interface and the MCP23S17 I/O ports remain to be tested.
I hope to make them available for sale by Monday next week.
In particular, are you running external memory for the video buffer? If so, how are you getting the speeds, are you using serial or parallel memory, and are you reading in the data alternate lines at a time?
I did some back-of-an-envelope calculations and I wasn't sure that external video ram was possible. Are you overclocking as well?
You are NOT a dummy!
Yes, I am using the external memory for the video frame buffer. Catalina also supports it for LARGE programs (thanks Ross!)
You are correct, my drivers do implement two scan-line sized buffers in the hub for the bitmap data.
I designed a parallel bus, using parallel SRAM's that allows up to 16MB of ram.
Here is a list of my currently available Morpheus video drivers - all are fully bitmapped!
- 640x480 with 4 colors per pixel from 256 color palette
- 800x600 with 4 colors per pixel from 256 color palette
- 1024x768 with 4 colors per pixel from 256 color palette
- 256x192 (with 1024x768 timing) with 256 colors per pixel
Coming soon:
- 800x480 with 4 colors per pixel from 256 color palette (for the OEM 7" WVGA/video/svideo module I sell)
- 320x240 with 256 colors per pixel, with sprites (255 colors + transparent) per pixel, the driver supports horizontally splitting the screen in half *with pixel accurate horizontal scrolling* and also pixel accurate independent vertical scrolling
NOTES:
1) All 4 color modes can have a unique palette for each scan line
2) All of my drivers have a common API that provides easy to use API calls for manipulating the palette, point plot, line drawing, rectangles and circles (which may optionally be filled)
Hub-only text driver:
- 128x64/100x50/80x40 RossH's port to Morpheus of Chip's high-rez text driver
My design supports 20MB/sec (MegaBytes per sec) burst access with up to 2.5MB of ram, I need to use faster SOJ ram's to get 20MB/sec with more than 2.5MB due to increased bus capacitance (which explains why I also designed and sell SOJ32->DIP32 PCB adapters)
While I have not tested it yet, I expect that I can get 25MB/sec by using faster memory and my 100MHz crystals - once that is verified, I expect to start selling "Turbo Morpheus" - that is Morpheus with faster memory and 6.25Mhz crystals.
Please find the feature lists below:
[code]
Morpheus Features (rev.2 PCB)
CPU #1:
* “Designed for Largos” which has its root file system on an SPI flash chip and has an SPI ram chip for an edit buffer
* I2C boot EEPROM
* I2C Phillips PCF8563 Real Time Clock
* P0-P7 available for I/O expansion on ProtoBoard compatible 10 pin header
* Compatible with SpinStudio modules from ucontroller.com except for the “Game Stacker”
* P8-P11 normally used for 32KB SRAM or FRAM, required for Largos
* P12-P15 for second 32KB SRAM or FRAM, or WinBond or compatible 4K page erasable FLASH, available up to 8MB (64Mb) in capacity, required for Largos - VMCOG supported
* P8-P15 may optionally be used to provide a second ProtoBoard compatible header if the SPI memory is not populated
* PS/2 Keyboard support (demo board compatible)
* PS/2 Mouse support (demo board compatible)
* Line level stereo audio out (may be used as two DAC’s for other purposes)
* DC input via pads for a DC 2.1mm/5mm jack
* space for 5V voltage regulation based on 7805
* space for 3.3V regulation based on LM1117 or AS2810 3.3V regulators
* Dual filter capacitor support (standard .1uF ceramic, and tantalum/electrolytic capacitor support as per Saphiea’s findings)
* 4 bit bus connecting to CPU #2
* controls RESET line of CPU #2
* can optionally drive XINPUT of CPU #2 from dedicated I/O pin instead of crystal (dial-a-cpu-frequency)
* PropPlug compatible programming header, adds optional fifth pin for 3.3V
CPU #2:
* 4 bit bus to CPU #1 on P28-P31
* normally loaded by CPU #1
* optional Prop plug compatible 5 pin programming header
* optional EEPROM socket for independent boot
* 512KB 32 pin DIP SRAM socket
* 24 bit address / 8 bit data FULL EXPANSION BUS ON 40 PIN HEADER
* VGA support, build time option for 64 color Parallax compatible mode or extended 256 color mode
* working 1024
(I still have to update the BOM, tech docs, and write a new build manual)
PCB prices will stay the same as before, however kit and assembled&tested prices will have to go up due to the additional parts used for the new features.
I also have my new 485Plug (blue) production boards in stock, I will be making them available for sale as soon as I make the BOM and figure out pricing.
In other news, the following boards are going in for production runs later this week:
- DIG-8-IN-V ... opto isolated 8 input nominally 24VDC input, 9 screw terminals
- DIG-8-OUT-V ... Solid state relay based 8 output board, nominally 24VDC but will switch up to 60V @ 500mA, 9 screw terminals
- CPUModule ... tiny through hole Propeller module
- Propteus (pcb rev.2) ... new, blue, version of the original Propteus
- miniPro (pcb rev.2) ... new, small prototyping board for the 10/11 pin headers
Sapieha did his usual excellent work, the boards are practically works of art! His PCB layouts rock.
He also made some excellent additions to the new prototype boards... and CPUModule is the result of combining my ideas with his for a small processor module. He also added some nice touches to Morpheus/Mem+ - namely optional battery backup of ram, soic8/tdfn8 patterns for SPI memory chips, and I probably forgot somethings
The new "PLC" board is also going in, hopefully for production, however I am only making a small number of boards just in case another board revision needs to be made. This board has a ton of features, which I will only talk about after the boards are 100% tested and functional as the first prototype needed some modifications.
Sorry for the delays guys, but I was unable to spend as much time on Propeller projects after UPEW as I'd have liked.
More boards will follow the above :-)
FlexMem driver work will resume shortly; I will modify one of the prototypes, as Sapieha suggests adding a resistor between MISO and MOSI - perhaps my problem with data corruption has been due to not having that resistor; it makes a certain amount of sense, as the corruption *always* happens on the last nybble transferred. Then back to VMCOG, writing v2.0 (Large VM) version...
Thanks,
David
What's his reasoning for this?
Some manufacturers recommend that BUT it was not only reason I used that!.
I have intensively tested SRAM's 23k256 on production version of Morpheus (We have possibility to 2 that ones.
Without resistors it was not problem run tests with slow drivers BUT in time I started RUNNING Fast driver 20/10MHz that one that use counter it started have unpredictable errors.
As this Driver was configured to use only one line for write/read I prepared one of IC's to that by lifting DI pin and connected it to DO output pin BUT in ViewPort it showed some Reading errors.
So I read all possible info from many manufacturers on connecting DI to DO and some of them recommend that RESISTOR (PropPin ->DI->100R->DO on SPI IC) so I added that to my next Test and my READING problems ware gone.
After that all speed test not have any writing/reading problems (Tested with 5MHz, 6.25MHz and 14.381.180MHz PLL8x) without adjusting counter values to this driver and that give with 14.xxMHz frequency some extra MHz on writing/reading that SRAM's.
Thanks,
David
Thank you! I will post some pics and more details on CPUModule later today.
Cluso99:
Thank you!
Kuronko:
Sapieha replied before I did :-)
I totally understand you questioning the need for the resistor; frankly I don't think it should be needed, however I trust Sapieha's test results. (Ask Sapieha how long it took him to convince me that the resistor could make a difference)
Based on Sapieha's test results, my best guess is that the 23K256's handling of tristating (and enabling) MISO (DO) after receiving the read command and address may not be up to the 25Mbps data rate of a 100Mhz or faster Prop using the optimal counter based drivers.
The data sheet states that Fclk is 20MHz max, so running the 23K256 at 25MHz+ is out of spec.
Since David Betz was asking in Bean's retro thread (see http://forums.parallax.com/showthread.php?131145-Proposal-for-a-Propeller-based-Retro-computer) I thought I'd mention that Morpheus IS available.
What is not available is an updated build manual for the nice new blue PCB's (Sapieha did an amazing job, as usual)
The new boards work well, and you can see them at UPEW.
For those of you skilled with a soldering iron and reading schematics, you should have a pretty easy build, following the silk screening and schematics; with a glance at the old build manual once in a while - the circuit has not changed (except for some new features).
I am afraid I don't have time to offer full kits or assembled and tested boards until after UPEW, but if some wants a Morpheus, Mem+ or combo before then, I will be making one or two mailing runs to the post office before going to UPEW. (I also have the stacking and non-stacking connectors the boards use). The power switch and 1/8" stereo jack are a bit unique, and I have some stock of those.
Actually, I just ordered more Morpheus and Mem+ (pcb rev2) boards, so I'll have more stock in time for UPEW.
Click on the Morpheus link in my signature, it will take you to the product page on the site, with pics of the old PCB's. I'll try to update the pics and pricing before UPEW, but right now I am working on demo's and the last of the new boards for UPEW ... (Sapieha again made some crazy nice layouts - but I won't announce the new designs until I test them - and no, they are not more variations of Morpheus or PropCade)
Anyway, good luck on the introduction of your new version of Morpheus!
After UPEW, I will have enough of a product line that I will actually start to advertise... as in google and other places :-)
There are always videos of UPEW, and product announcements will be posted by everyone - but it is a GREAT networking event, it will be my third in three years.
Hopefully we'll meet at UPEW 2012.... OBC is liable to start the thread for it the day after UPEW 2011 (OBC, teasing)
First up is a YouTube 800x600 mp4 video of my graphics demo running at 800x600 with four colors (out of 256 color palette) per pixel.
I don't know how to embed YouTube videos in the forum, so here is the link http://youtu.be/FOpnMffHmxM
Updated URL for HD upload: http://youtu.be/XQoVsVe_Yac
(See message #280 later on the page for the embedded high quality video)
Please excuse the compression artifacts; YouTube re-compressed my MP4.
p.s.
I now ship rev.2 PCB's, I am out of the original rev.1 green PCB's.
Rev2's feature Sapieha's art-like layout, IR in, and a few more features that I will detail later.
Put the link between [noparse][video][/video] tags, like this:
[video]
[video=youtube_share;FOpnMffHmxM]
-Phil
I have the rev2 boards ready; I just have to write an addendum to the build manual to cover the changes - mind you, all experienced builders will need is the schematic.
The text scrolling is a bit jerky, but given that the bitmap is being scrolled in 512KB of external ram, its not half bad
I will try to improve the video quality.
I am about to try to upload 720p and 1080p versions of the video, hoping that will still leave much higher quality after YouTube re-compresses it.
This one looks MUCH better if you set your playback to 1080p or 720p, especially full screen
[video=youtube_share;XQoVsVe_Yac]
URL: http://youtu.be/XQoVsVe_Yac
(thanks Phil!)
p.s.
I deleted the original lower resolution version.
I am quite happy with what I was able to accomplish with the two Propellers (1's) on Morpheus.
I've had some email inquiries asking if Morpheus was still available.
rev1.0 boards are sold out, but I have a small stock of rev2.0 boards.
http://www.mikronauts.com/parallax-propeller/morpheus/
While the documentation is for rev1.0, with the rev2.0 schematics intermediate to advanced propeller users should have no trouble building it
I've cut the prices for the bare boards, and if I sell out, I can get more boards made.
Note: the 512Kx8 DIP32 SRAM is has long lead times, but the pin-compatible 128Kx8 is still fairly readily available, and I also have some SOJ-DIP adapter PCB's.