Work in Progress - Winbond Flash / Microchip SRAM driver + FemtoBasic

Mike Green · 2009-05-06 05:48

For those interested in experimenting, here's my current working copy of FemtoBasic modified for use with TriBlade#1 and a 2MB Winbond
flash memory chip.

There's a subset VT100 terminal emulator included and tested low level support for a 2MB to 8MB Winbond flash memory chip.
You can erase 4K blocks (FERASE) and read (FREAD) data and write (FWRITE) small amounts of data to flash. The DUMP statement
works with flash as well.

There's the ability to create, open, and delete files and read and write data to files, but this is not tested yet. Once this works, the SAVE,
LOAD, FILES, and COPY statements will be implemented for flash so they look and work much like for an SD card. Have a look at the
"Winbond Driver".

Note that the source is set up for a 6MHz crystal. You can change that and recompile. It's also set up for a VGA display. I've included the
half-height 40 x 30 TV text driver modified to be able to substitute for the VGA driver including the VT100 support. I haven't tested it yet,
but it compiles and looks like it should work.

6 May 09: Corrected error in FWRITE
7 May 09: Corrected error in findFile. Added code for SAVE, LOAD, FILES ... tested & seem to work.
7 May 09: v4.005 - Corrected error in eraseFile and DELETE
9 May 09: v4.008 - Corrected errors in eraseFile, FILES, DELETE, COPY. Sped up some operations.
10 May 09: V4.009 - Corrected errors in readFile and writeFile. Now files >4K work properly.

At this point, everything flash related appears to work except the SPIN statement to load and execute Spin programs.

The Winbond Driver is designed to not use a central directory. This slows things down, but helps reduce the number of times a 4K sector has to be erased or written. Because of the way flash memory is implemented, once you write to a portion of the file, you can't write again to that portion until that 4K block is erased. The Winbond Driver is intended to provide for sequential files where you write sequentially to the file, then read the file sequentially one or more times. Next step is to finish the PASM routines so the SPIN statement will work and the whole Winbond Driver will run faster.

Post Edited (Mike Green) : 7/4/2009 5:03:53 AM GMT

Dr_Acula · 2009-05-06 07:59

So many things happening so quickly!

Right, I've been jolted out of my stupor, and am off to the shed to solder up a triblade.

Re femtobasic, sooner or later any program seems to grow bigger than the available memory. One way around that is Shell and Chain, to pass control to another program. Does femtobasic have these commands? (I must do some more research, it sounds intriguing).

Cluso99 · 2009-05-06 11:44

Great work Mike

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:

· Home of the MultiBladeProps: TriBladeProp, SixBladeProp, website (Multiple propeller pcbs)
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: Micros eg Altair, and Terminals eg VT100 (Index)
· Search the Propeller forums (via Google)
My cruising website is: ·www.bluemagic.biz·· MultiBladeProp is: www.bluemagic.biz/cluso.htm

Mike Green · 2009-05-06 12:40

FemtoBasic has the ability to load and run new Basic programs either from EEPROM or an SD card. Look at the descriptions of the LOAD and RUN statements. It doesn't preserve the values of variables when you do a RUN, but you can write values to EEPROM or an SD card file and reload them in the "overlay". If there's a file "AUTOEXEC.BAS" on the SD card, FemtoBasic will do a LOAD "AUTOEXEC.BAS":RUN when it starts up.

FemtoBasic also has the ability to run a new Spin/PASM program using the SPIN statement. Parallax used FemtoBasic for a kiosk for Propeller demo programs that would let you choose a demo from a menu, then execute it. The demo would do a reboot when it finished, restarting FemtoBasic.

CassLan · 2009-05-06 13:24

Now I have to get a Tri-Blade [noparse]:)[/noparse]

Mike Green · 2009-05-06 13:30

I just posted a correction. I'll continue to post corrected versions to the first message on this thread along with a date and what's corrected.

Mike Green · 2009-05-07 05:28

Exciting ... LOAD and SAVE seem to work. OPEN R and OPEN W seem to work and FILE / FILE = seem to work as well as READ, WRITE and FILES. I've not tested these other than on small files (< 4096 bytes). There are also long pauses (500-1000ms) as some of the file lookups go through lots of unused sectors one at a time. I'm trying to avoid having a central directory since that would have to be erased and rewritten every time it's changed. I may have to do something anyway since lookups and space allocation are slow.

Mike Green · 2009-05-09 23:43

At this point, everything seems to work except the SPIN statement, so you can save and load Basic programs, read and write sequential data files, copy 32K EEPROM pages to and from flash files, list the files on the flash. You just can't execute Spin programs stored on flash memory yet.

Ole Man Earl · 2009-05-09 23:46

Great work Mike. I am a fan of yours.

Mike Huselton · 2009-05-10 08:54

I also admire Mike's professional approach, clarity of thought and patience exhibited in his replies to a full spectrum of users, both neophytes and experts.

I want to become more like Mike when I grow up

. I am sure Mike's patients appreciate his approach as well.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
JMH

Mike Green · 2009-05-25 19:02

Here's a later version of the Winbond driver. It's now been moved to a Demo Board since I was unable to get the loader to work on the TriBlade Prop #1. I think the problem has to do with cog I/O pin conflicts between the Winbond loader and the I2C loader cogs and I plan to eventually get back to it, but I wanted to do some specific things with the Winbond flash.

This driver also supports the Microchip 23K256 32K byte SRAM. It allows for up to 2 SRAM chips for a total of 64K. The included version of FemtoBasic is used for development and testing of this driver and has statements for low level access to the flash memory and SRAM. You can save Basic programs to the flash and can use the COPY and SPIN statements to handle Spin programs the same way you'd use an SD card. The only limitation is that you can't replace (overwrite) a file. You have to delete it first, then recreate it. This is a limitation of the flash memory, something the SD card internal controller takes care of.

Added a comment to FemtoBasic.spin describing the connections for the Flash and SRAM memories.

Corrected use of "dev" in Winbond_Driver.spin. Not allowed under Propeller Tool 1.2.6. Works with BradC's BST.

I think I've really corrected it this time. Sorry.

Post Edited (Mike Green) : 5/26/2009 2:54:16 PM GMT

Mike Huselton · 2009-05-25 19:25

Thank you! I have 23K256 chips and I was going to have write the driver myself. Again, thank you!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
JMH

Mike Huselton · 2009-05-25 19:41

Mike, how difficult do you think it would be to add FRAM support to the driver?

My sample devices include:

FM24L256-G
FM24C512-G
FM25V10

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
JMH

Mike Huselton · 2009-05-25 19:59

Compile Error: Block Designator Must Be In The First Column
Module: WinBond Driver
Routine: PRI sendRecv(dev, sD, rCt, rA) | i, m ' Send and possibly receive
..................................^^^ Error cursor at dev

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
JMH

Mike Green · 2009-05-25 21:30

James,
There's some kind of new error in the Propeller Tool 1.2.6 with the use of "dev". It compiles fine with BradC's compiler. Try replacing the 3 uses of "dev" with "d" and recompiling. It should work.

All 3 FRAM devices you mentioned are I2C devices as I recall and they should work fine with the existing I2C drivers as direct replacements for ordinary EEPROM. Some of them are 5V-only devices and you'd need some kind of voltage interface (3.3V to 5V for SCL and SDA). Usually just a 1K series resistor will work.

BradC · 2009-05-25 21:33

Mike Green said...
James,
There's some kind of new error in the Propeller Tool 1.2.6 with the use of "dev". It compiles fine with BradC's compiler. Try replacing the 3 uses of "dev" with "d" and recompiling. It should work.

Now _that_ is interesting. Sounds like the compiler is being updated and they are adding a new block type to it.. time to get out the disassembler again [noparse];)[/noparse]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"VOOM"?!? Mate, this bird wouldn't "voom" if you put four million volts through it! 'E's bleedin' demised!

jazzed · 2009-05-25 21:36

The "dev" thing is also in version 1.1 ..... I just ran into it this morning. OBJ dev[noparse][[/noparse]1] : "somedevice" gave me an error: "Block designator must be in the first column."

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve

Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230

BradC · 2009-05-25 21:38

jazzed said...
The "dev" thing is also in version 1.1 ..... I just ran into it this morning. OBJ dev[noparse][[/noparse] 1][noparse][[/noparse] 2] : "somedevice" gave me an error: "Block designator must be in the first column."

Oh, in which case I won't bother. Perhaps its a reserved word we did not know about.

I guess I should make it a reserved word in my compiler to save the incompatibility. Is it documented anywhere?

<edit>
Nyaaaahaaa.. further investigation shows it to be used in constructs similar to this
DEV ARCHIVE "FILENAME"
DEV PRECOMPILE "FILENAME"

I was wondering what those two directives were.. now I need to see if they are actually hooked up in the compiler or just remnants. My gut feeling is they provide some sort of interface between the compiler and the IDE. In any case, I won't hijack Mikes thread any further [noparse]:)[/noparse]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"VOOM"?!? Mate, this bird wouldn't "voom" if you put four million volts through it! 'E's bleedin' demised!

Post Edited (BradC) : 5/25/2009 9:54:15 PM GMT

jazzed · 2009-05-25 21:44

Friggin' tags [noparse]:)[/noparse] OBJ dev[noparse][[/noparse] 1] : "somedevice" ....

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve

Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230

Mike Huselton · 2009-05-26 14:16

Mike,

Mike Green said...
Corrected use of "dev" in Winbond_Driver.spin. Not allowed under Propeller Tool 1.2.6. Works with BradC's BST.

This correction does not appear in my freshly downloaded version. Still uses "dev".

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
JMH

BradC · 2009-05-26 14:21

James Michael Huselton said...

This correction does not appear in my freshly downloaded version. Still uses "dev".

Change all the instances of "dev" to something else. Dev is a reserved word (a block) in the Parallax compiler.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"VOOM"?!? Mate, this bird wouldn't "voom" if you put four million volts through it! 'E's bleedin' demised!

Mike Huselton · 2009-05-26 14:29

BradC,

I have done so yesterday. This was meant to inform Mike that he still needs to change the source for the unsuspecting.
He states that he corrected this, but the code remains the same.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
JMH

Mike Green · 2009-05-28 05:30

There are some minor changes / fixes to the VGA driver and Winbond driver here and some additional comments in the Winbond driver.

Mike Huselton · 2009-05-28 09:41

Mike,

Thanks for the updated code.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
JMH

Mike Green · 2009-07-04 04:41

Attached is a new version of the Winbond flash / Microchip sram driver written in assembly. It's drop-in compatible with the Spin-only version which is also included and quite a bit faster which is to be expected.

Next step is to modify "createFile" so it will erase any existing copy of the specified file. This will make it work more like an SD card.

Mike Green · 2009-07-04 05:04

Here's the next version that erases an existing file on a createFile.

Mike Green · 2009-07-04 18:56

The Winbond Flash / Microchip SRAM Driver has been posted to the Object Exchange. Further updates will be posted directly there.

Mike Green · 2009-07-04 21:17

The Winbond version of FemtoBasic for a VGA display has been posted in the ObEx along with the I/O driver and a minor error has been corrected in the driver.

Mike Huselton · 2009-07-05 15:42

What is the purpose of ouroboros and ouroboros1 in the your software architecture?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
JMH

Mike Green · 2009-07-16 16:17

Ouroboros and ouroboros1 are left-overs from a project to make a Basic to LMM compiler for the Propeller. I forgot to delete them from the archive for the Winbond/SRAM driver. They're not in the Object Exchange version of the driver.

Duane Degn · 2009-10-21 16:27

I'm trying to use Winbond chip W25X16A.· I've written a small demo object with hopes of understanding Mike Green's Winbond driver.

The size of the chip is read correctly but I can't write to the chip (either that or I can't read from the chip).· Here's the code.

{{  WinbondDemo1a.spin
  Attempt at a demo program for Michael Green's
  Winbond flash driver.
         Vdd(+3.3V)           Vdd(+3.3V)
           &#61463;                    &#61463;
          &#9484;&#9515;      W25X16A      &#9484;&#9547;&#9472;&#9488;
          &#61628;&#61628;                   &#61628;&#61628; &#9474;       
          &#9474;&#9474; &#9484;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9488;&#9474;&#9474; &#9474; 0.1µF        
    P0 &#61609;&#9472;&#9472;&#9532;&#9532;&#9472;&#9508;1 /CS    Vcc   8&#9500;&#9532;&#9532;&#9472;&#9531;&#9472;&#9472;&#61611;&#9472;&#9472;&#9472;&#9472; Vss
    P3 &#61615;&#9472;&#9472;&#9532;&#9531;&#9472;&#9508;2 DO     /Hold 7&#9500;&#9496;&#9474;
          &#9492;&#9472;&#9472;&#9508;3 /WP    CLK   6&#9500;&#9472;&#9532;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#61610; P1
          &#9484;&#9472;&#9472;&#9508;4 GND    DIO   5&#9500;&#9472;&#9531;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#61620; P2    
          &#61464;  &#9492;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9496;  
         Vss
             Pullup resisters are 10K
                                               
 }}
con
 
   _clkmode  = xtal1 + pll16x
   _xinfreq  = 5_000_000
   _CS = 0
   _Clk = 1
   _DIO = 2
   _DO = 3
   _DebugBaud = 9600
OBJ
  Debug : "FullDuplexSerialPlus"
  Win : "Winbond_Driver"
   
VAR
   long location
   byte buffer[noparse][[/noparse]128]
PUB Main | localTemp
  Debug.Start(31, 30, 0, _DebugBaud)
  location := @buffer[noparse][[/noparse]0]
  localTemp := win.start(_CS, _Clk, _DIO, _DO, -1, -1)
  Debug.str(string(13, "Starting Winbond driver returned = "))
  Debug.dec(localTemp)
  
  localTemp := flashSize
  Debug.str(string(13, "flashSize = "))
  Debug.dec(localTemp)
  eraseData($8000)
  repeat localTemp from 0 to 127
    buffer[noparse][[/noparse]localTemp] := localTemp
  Debug.str(string(13, "buffer filled with squential numbers = "))
  repeat localTemp from 0 to 127
    Debug.hex(buffer[noparse][[/noparse]localTemp], 2)
    Debug.str(string(", "))
   
  Debug.str(string(13, "Write buffer to flash memory.  Then erase buffer from prop memory."))
  writeData($8000, location, 128)
  waitcnt(clkfreq / 10 + cnt)
  byteFill(@buffer, 0, 128)
  Debug.str(string(13, "cleared buffer = "))
  repeat localTemp from 0 to 127
    Debug.hex(buffer[noparse][[/noparse]localTemp], 2)
    Debug.str(string(", "))
   
  Debug.str(string(13, "Restore buffer from flash memory.  Buffer = "))
  readData($8000,location,128)
  waitcnt(clkfreq / 10 + cnt)
  repeat localTemp from 0 to 127
    Debug.hex(buffer[noparse][[/noparse]localTemp], 2)
    Debug.str(string(", "))
  
  repeat
pub winStop
   return win.stop
pub flashSize                     ' Return the size of the flash memory in bytes
  return win.flashSize
pub readData(a,d,c)               ' Read a block of data from Flash
  return win.readData(a,d,c)
pub writeData(a,d,c)              ' Write a block of data to erased Flash
  return win.writeData(a,d,c)
pub eraseData(a)                  ' Erase a block of data
  return win.eraseData(a)

The DO, /WP, DIO, and /Hold pins all have 10K pullup resistors.

What am I doing wrong?

I don't post often but I'm a fequent reading of this forum.· Most of my questions have already been asked and answered here.· Thanks for all of your help.

Duane Degn

Work in Progress - Winbond Flash / Microchip SRAM driver + FemtoBasic

Comments