I should take the code I did for my one digit clock entry for the RTC on the PPDB and see if we can share the high level code with the different low level code.
Nice, Brian!
I should take the code I did for my one digit clock entry for the RTC on the PPDB and see if we can share the high level code with the different low level code.
You will notice up top where I aliased the low level i/o:
: RTC! ( byte -- ) SI2C! ;
: RTC!? ( byte -- flag ) SI2C!? ;
: RTC@ ( flag -- byte flag ) SI2C@ ;
When it became apparent that the 1307 was 'slow' the sum of the edits were three "S"s one in front of each I2C i/o primitive.
I originally wrote this routine for PUMPTIMER, my entry in the Thanksgiving Challenge. I spent an hour this afternoon, isolating the routine and its few dependent variables, and included some code snippets of usage so it can just be included and used with little or no additional edits ... enjoy!
The neat new 8x8 RGB LED Matrices offer a lot of fun and when several are combined ca produce spectacular results.
I transcribed one of Simply Tronics for the BS2 from PBASIC to TF, added a few things, the made it more Forth-like.
The concept is redirection of Console Output which means all the nice formatting words Peter created are available, as is EMIT to send unprintable values.
Can some one tell me the fundamental difference between version V2 and the original?
Regards,
Rich
I've lost track of that but when I started to change the way the kernel operated I wanted to break that off into a different version. I don't see the sense in zero point zero versions so I started with the first version which is of course V1. So V2 is just the platform for testing new enhancements while leaving V1 stable but V2 has proved stable from the start however.
So to sum up the enhancements:
Loop branch parameters are stacked in cog RAM to allow for minimal looping overhead
Faster dictionary searches
Trimmed VM kernel and modularized special I/O instructions for SPI and SD etc.
Extended direct call to VM opcodes using XOP to access top 256 longs.
Removed possibility of absolute calls (was unused) and instead expanded the vector table handling.
Updated serial receive to preprocess special control characters and sequences including fast-load filtering
Optimized the number preprocessing for compilation speed
Optimized the memory map etc
So don't even bother with V1 as all the extensions and examples are coded for V2 anyway.
If you're not picky about what a language "should be" and like to get interactive with your hardware and get up-and-running as quick as possible as well as having fast compiled code then I can definitely recommend Tachyon.
Peter - could you check into a minor issue. I think that if you have a line with just a backslash and no trailing whitespace, Tachyon may not terminate the comment with the end of line. This had me confused for a while, and it might save someone some time. I didn't look into how comments are handled (are they just another word?), so maybe this is not a surprise.
Also I was curious what you think about DOES> - do you have plans to implement it, or do you find that you don't make use of it in embedded programming?
Peter - could you check into a minor issue. I think that if you have a line with just a backslash and no trailing whitespace, Tachyon may not terminate the comment with the end of line. This had me confused for a while, and it might save someone some time. I didn't look into how comments are handled (are they just another word?), so maybe this is not a surprise.
Also I was curious what you think about DOES> - do you have plans to implement it, or do you find that you don't make use of it in embedded programming?
Cosmetics, cosmetics..... Yes, I should do something about the backslash as TF doesn't use a line input buffer, just a word buffer so it's just as happy to have a backslash and CR as a terminator as it is just to have a space.
I haven't ruled out implementing DOES> as there have been a few times I could have used it effectively, so maybe sometime in the next few weeks after I've finished wrangled with a few things that didn't quite get done last month!
Seems like I've been away for ages from any active development work on the Tachyon kernel. Seeing it's a new year I've fixed up the Intro page and brought it up to date as well as the various files and links. I hadn't realized that the links were still pointing to V2.0 rather than V2.1 which is my current kernel since November so make sure you grab this version. I will also update the binaries and make sure that I snapshot a copy that can always be used with the intro page examples.
I will finish off judging the much delayed challenge and proceed to kick off the 4/4 challenge for 2013 as well which I hope I will have better organized this time
130107
Added , word to compile longs (complements | and || which compile bytes and words)
Peter added the LONG to this for me. I thought that I would repay the favor by running these scripts and copying the console. It should be self-explanatory. I know the visuals one after another like this really brought it home for me ... ENJOY
\ ---- LONG ----------------
TABLE XY ok
ok
1 , 2 , 3 , 4 , ok
ok
XY 20 DUMP
2D44: 01 00 00 00 02 00 00 00 03 00 00 00 04 00 00 00 ................
2D54: B5 02 74 20 B8 24 03 03 00 00 00 00 00 00 00 00 ..t .$..........
ok
\ -------- WORD -------------
TABLE XX ok
ok
1 || 2 || 3 || 4 || ok
ok
XX 20 DUMP
2D58: 01 00 02 00 03 00 04 00 B6 03 74 20 B8 24 03 03 ..........t .$..
2D68: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
ok
\ --------- BYTE --------------
TABLE XZ ok
ok
1 | 2 | 3 | 4 | ok
ok
XZ 20 DUMP
2D64: 01 02 03 04 B5 03 74 20 B8 24 03 03 00 00 00 00 ......t .$......
2D74: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
ok
This is a neat little project I started today for my wife's car. I just installed a new double DIN 6.5" touchscreen stereo in my car and I also had a brand new single DIN unit for my wife's car but the new unit didn't support the steering wheel interface (SWI). I thought, this is a job for the Prop and Tachyon! (with the same sort of emphasis as "Batman and Robin"). A few hours later I'm just about ready to install it in the car although I have tested it out on the bench. Here's a sneak peek at part of the code, I will put up a project page in the next day or so and link it from the Intro page.
[FONT=courier new]
[COLOR=#000000]{ *** STEERING WHEEL INTERFACE ***[/COLOR]
[COLOR=#000000]Measure the resistance of the steering wheel interface[/COLOR]
[COLOR=#000000]Resistance is a maximum of 6K8 and open if no switches are pressed[/COLOR]
[COLOR=#000000]RC time constant method is used to detect the resistance but to help there is[/COLOR]
[COLOR=#000000]a 100K resistor added in parallel. [/COLOR]
[COLOR=#000000]The resistance is referenced to ground and discharges the capacitor as this method offers good noise immunity.[/COLOR]
[COLOR=#000000]PORT ------220R------|-------|-----> to SWI resistors[/COLOR]
[COLOR=#000000] 104C[/COLOR]
[COLOR=#000000] |------------>[/COLOR]
[COLOR=#000000] _|_[/COLOR]
[COLOR=#000000]}[/COLOR]
[COLOR=#000000]#12 CONSTANT =scale [/COLOR]
[COLOR=#000000]#P8 |< CONSTANT =swi[/COLOR]
[COLOR=#000000]\ The maximum value read after conversion is 325 for the 100K resistor which is converted to zero[/COLOR]
[COLOR=#000000]\ 10K reads around 15[/COLOR]
[COLOR=#000000]pub [/COLOR][B][COLOR=#000000]SWI[/COLOR][/B][COLOR=#000000] ( -- val )[/COLOR]
[COLOR=#000000] =swi OUTSET \ charge the capacitor[/COLOR]
[COLOR=#000000] #100 us \ for long enough (prototype is using a series 220R resistor)[/COLOR]
[COLOR=#000000] =swi INPUTS \ now float the line[/COLOR]
[COLOR=#000000] CNT@ =swi WAITPNE \ and wait for it to go low[/COLOR]
[COLOR=#000000] CNT@ SWAP - \ calculate the time constant [/COLOR]
[COLOR=#000000] =scale SHR \ justify[/COLOR]
[COLOR=#000000] DUP #32 > IF DROP 0 THEN[/COLOR]
[COLOR=#000000] ;[/COLOR]
[COLOR=#000000]\ IR Code [/COLOR]
[COLOR=#000000]\ LOW (carrier) 12.5ms [/COLOR]
[COLOR=#000000]\ HI (no carrier) 4.5ms[/COLOR]
[COLOR=#000000]\ pulse low then 32 bits msb to lsb[/COLOR]
[COLOR=#000000]#P24 CONSTANT =irled[/COLOR]
[COLOR=#000000]pri [/COLOR][B][COLOR=#000000]IRLED[/COLOR][/B][COLOR=#000000] ( units -- )[/COLOR]
[COLOR=#000000] A =irled APIN #38 KHZ \ Generate a 38kHz carrier[/COLOR]
[COLOR=#000000] #500 * us[/COLOR]
[COLOR=#000000] MUTE \ mute or stop the carrier[/COLOR]
[COLOR=#000000] ;[/COLOR]
[COLOR=#000000]pub [/COLOR][B][COLOR=#000000]IRTX[/COLOR][/B][COLOR=#000000] ( code -- )[/COLOR]
[COLOR=#000000] 0 REV \ flip msbs into lsbs for right shifts[/COLOR]
[COLOR=#000000] #25 IRLED \ 25 units of carrie[/COLOR]
[COLOR=#000000] #4500 us [/COLOR]
[COLOR=#000000] 1 IRLED[/COLOR]
[COLOR=#000000] BL FOR [/COLOR]
[COLOR=#000000] DUP 1 AND 2* 1+ #500 * us \ "high" period [/COLOR]
[COLOR=#000000] 1 IRLED \ the "low" pulse [/COLOR]
[COLOR=#000000] 2/[/COLOR]
[COLOR=#000000] NEXT[/COLOR]
[COLOR=#000000] DROP[/COLOR]
[COLOR=#000000] ;[/COLOR]
[COLOR=#000000]\ Test it out[/COLOR]
[COLOR=#000000]$817EF00F IRTX[/COLOR]
[COLOR=#000000]pub [/COLOR][B][COLOR=#000000]PWR[/COLOR][/B][COLOR=#000000] $817EF00F IRTX ;[/COLOR][/FONT]
I've refined the IR transmission a little by reducing the 32-bit code down to a family code byte plus the IR code byte. So the digits 1 to 9 are simply values 1 to 9 and so on. Except for different family codes this seems to work with all the remotes I have lying around so this may just become a useful object on it's own.
IR remote transmission format (this is the way I look at it):
PREAMBLE = 20 times the synch pulse duration
4.2ms gap
"SYNCH" PULSE
FAMILY CODE BYTE (lsb first)
FAMILY CODE BYTE INVERTED
IR CODE BYTE
IR CODE BYTE INVERTED
[FONT=courier new]
#P24 CONSTANT =irled
BYTE family \ IR family code byte
$81 family C! \ Set the default family code
pub IRLED ( units -- )
A =irled APIN #38 KHZ \ Generate a 38kHz carrier
#460 * us
MUTE \ mute or stop the carrier
;
pub IRTX8 ( byte -- )
8 FOR
DUP 1 AND IF #1530 ELSE #480 THEN us \ "high" period - a 1 is 3 times the length of a 0 high
1 IRLED \ the "low" pulse
2/
NEXT
DROP
;
\ Transmit the IR code byte as code + inverted
\ First transmit the preamble then the family code + invert and then the IR code + invert
pub IRTX ( code -- )
#20 IRLED \ HEADER PREAMBLE
#4200 us
1 IRLED \ "start" bit
family C@ DUP IRTX8 INVERT IRTX8
DUP IRTX8 INVERT IRTX8
;
[/FONT]
Now I'm going to mangle the way the steering wheel buttons operate so that it does more than it does at present and it will sense sequences and duration for extra control.
Fired Tachyon V2 on the quickstart propeller board earlier this week. What a delight! Worked as soon as realizing that the binary image is loaded on EEPROM at 56kb, but the board talks at 230kb when booted after that. Runs great on minicom under linux. The introduction webpage by Peter Jakacki, the author of Tachyon, reveals enough to study the source code and its comments and play with a few unfamiliar Forth words as well as the propeller-specific Tachyon words to start making things happen. The performance is so good, I believe it will encourage folks who have never tried Forth to get on with it, because it will open many new possibilities in both functionality and execution speed, including immediate debugging, as only Forth can.
At last I started coding with Tachyon.
I'm playing with the cheap stepper motors, and the interactive environment is great.
I'm using the wifly modules, and I ordered the BT ones too, remote programming is a nice option. I'm really having fun, I only need some time to adapt to the different approach and be a little more fluent in the language.
Thanks for the effort.
Massimo
Fired Tachyon V2 on the quickstart propeller board earlier this week. What a delight! Worked as soon as realizing that the binary image is loaded on EEPROM at 56kb, but the board talks at 230kb when booted after that. Runs great on minicom under linux. The introduction webpage by Peter Jakacki, the author of Tachyon, reveals enough to study the source code and its comments and play with a few unfamiliar Forth words as well as the propeller-specific Tachyon words to start making things happen. The performance is so good, I believe it will encourage folks who have never tried Forth to get on with it, because it will open many new possibilities in both functionality and execution speed, including immediate debugging, as only Forth can.
Thanks Fidolo, and welcome to the Propeller forums. You might find those binaries have passed their "best before" date and the memory map has changed enough that the old binary dump methods don't work for now. In the meantime I have downloaded the binary image of the current kernel (Tachyon V2.1-130110 kernel-5MHZ.binary) compiled for a 5MHz crystal and defaulted to 230400 baud into the Dropbox account. When you have the kernel running then set Minicom to a 10ms line delay (^A T D) and copy and paste the EXTEND.fth file, either from the Dropbox or the online document. Of course you can compile the kernel from the source as well, it's very easy and it's only a single file
Some of the Forth words are unfamiliar but not so much if you are familiar with the Propeller chip. There are other words which I would probably change now but they're okay and so they've stuck.
If you have any suggestions for improving Tachyon or the intro page then please let me know.
If you have any suggestions for improving Tachyon or the intro page then please let me know.
PETER, you have probably had enough suggestions and requests from me to last two lifetimes, but I have been meaning to tender this request for a while ... modify WORDS and QWORDS to take a simple search string
" S" SEARCH or " IN" QSEARCH searches for left justified match of cmds to string. SEARCH ---> WORDS QSEARCH ---> QWORDS'
This will help the apprentices/rookies and will definitely help us journeymen.
I probably need both hands full of fingers to count the number of times in the last 5 weeks I have sent you some code only to have it returned with an "oh by the way, you can use this neat word in place of these 5!" like, for example "+!" .
QWORDS is helpful but a simple filter will go a long way to making Tachyon easier to learn.
PETER, you have probably had enough suggestions and requests from me to last two lifetimes, but I have been meaning to tender this request for a while ... modify WORDS and QWORDS to take a simple search string
" S" SEARCH or " IN" QSEARCH searches for left justified match of cmds to string. SEARCH ---> WORDS QSEARCH ---> QWORDS'
This will help the apprentices/rookies and will definitely help us journeymen.
I probably need both hands full of fingers to count the number of times in the last 5 weeks I have sent you some code only to have it returned with an "oh by the way, you can use this neat word in place of these 5!" like, for example "+!" .
QWORDS is helpful but a simple filter will go a long way to making Tachyon easier to learn.
Oh, by the way I have been wanting to put a filter into place but I just never got around to it
Rather than passing a string which would mean you would always have to do it or else it might pick up garbage from the stack, I would rather have WORD look for another word or string to describe further actions. This is fine for interactive terminal use whereas the string passing method is better for words which are meant to be called from other words.
So I would have WORDS operate this way if I wanted to find all the words with a + symbol for instance:
WORDS +
In this case wildcards are assumed without having to specify, this makes it easier to get down to business.
If I was looking for anything that started with + then it would be entered this way:
WORDS +**
Where star star (seems unique enough) is the wildcard but also forces a search only for words that begin with + in which case you would find +!
Now this brings me back to my original intention of storing help comments for each word which would be stored if EEPROM > 32K or SD or SPI Flash were available, this is what would happen.
WORDS +!
CFA CODE ATRS NAME DESC
6BD3 0061 ...$ +! ( long addr -- ) add long to long at hub addr
6BC6 005D ...$ W+! ( word addr -- ) add word to word at hub addr
6BB1 0059 ...$ C+! ( byte addr -- ) add byte to byte at hub addr
If I was looking for anything that started with + then it would be entered this way:
WORDS +**
Where star star (seems unique enough) is the wildcard but also forces a search only for words that begin with + in which case you would find +!
Peter, I would suggest just using +* with * being the wildcard. And the additional default rule would be if there is no wildcard, a * will be added before and after, which gives you the intended behaviour.
Peter, I would suggest just using +* with * being the wildcard. And the additional default rule would be if there is no wildcard, a * will be added before and after, which gives you the intended behaviour.
of course much better :-)
Yes but if I'm searching for U* words where the star is also what I'm looking for then I am stuck. Forth words can use any mix of characters so we need to be able to find them. A double combination of the star should be suitable for practically all cases, we don't want to get too crazy (too late).
(Tachyon V2.1-130110 kernel-5MHZ.binary) compiled for a 5MHz crystal and defaulted to 230400 baud into the Dropbox account. When you have the kernel running then set Minicom to a 10ms line delay (^A T D) and copy and paste the EXTEND.fth file, either from the Dropbox or the online document. Of course you can compile the kernel from the source as well, it's very easy and it's only a single file
Some of the Forth words are unfamiliar but not so much if you are familiar with the Propeller chip. There are other words which I would probably change now but they're okay and so they've stuck.
If you have any suggestions for improving Tachyon or the intro page then please let me know.
Thanks Peter, I was just wondering about how to install your latest build, and you just answered my question before I even asked it!
I may have suggestions at some point, but I will not attempt to contribute to your effort until I have a decent familiarity with the extraordinary Propeller-Tachyon combination, which will take a while. I remember when it was estimated that there were about 50,000 Forth progammers in the world, roughly in the days of the FORTH-83 standard. Not being a software engineer by trade, Forth has nevertheless been essential to my professional career, which required me to build a lot of fast experimental hardware, to which Forth is so well adapted.
Since these days, I have been feeling a bit nostalgic, worried about Forth seemingly being progressively eclipsed, quite unfairly if one considers its unique ability to get the best out of hardware. Now I wonder whether Forth on the propeller chip may reverse the trend by inspiring robotic enthusiasts to get the most out of the propeller chip without going all the way down to the tedium exclusive assembly programming, and without any formal training in software engineering, thanks to the inner simplicity and the interactiveness of Forth?
I mentioned that I would make the project page for the Steering Wheel Inteface to Infrared remote control available. So here it is.
Just to make this project more visible I will post a separate thread as well.
I occasionally use the PINS? word to tell me about the state of the port pins. Because I have so many different hardware configurations I can't always keep track of the I/O unless I have a schematic in front of me. To assist a little in this regard I have added an extra function to PINS? to also tell me if there is some kind of load on the pin and whether its a pullup or pulldown load. So if for instance a pin is connected to an NPN (via a resistor) then this will show up as a D or pullDown whereas pullup resistors display as a U for pullUp. X indicates a floating pin.
In this hardware P5 is connected to an NPN and P8 has a pulldown with pullups on P16..31 except for P26 which has a pulldown for reset control of a Silabs CPU..So an analysis of what PINS? reports and a comparison with the schematic indicates that this word is very useful.
Nothing fancy, but I really like the interactive approach and the possibility to remotely program the propeller.
Ordered a couple of BT modules too.. :-)
I received the BT modules!!
I was able to connect to Tachyon from an Android device, but I have problems finding a nice BT terminal for the phone.
Blueterm for instance is not sending the correct end of line.
I also cannot find a tool with the possibility to set the mandatory line delay.
Any suggestion?
Thanks in advance,
Massimo
I received the BT modules!!
I was able to connect to Tachyon from an Android device, but I have problems finding a nice BT terminal for the phone.
Blueterm for instance is not sending the correct end of line.
I also cannot find a tool with the possibility to set the mandatory line delay.
Any suggestion?
Thanks in advance,
Massimo
Blueterm is the right one but you will find that the return problem has to do with the keyboard. Load "Terminal IDE" keyboard as this seems to be the best for this kind of stuff. I modified Tachyon the other week to accept a DEL key as well as the BKSP key because I couldn't find an Android keyboard that would actually send a $08 instead of a $7F or escape sequence.
Thanks, on the telephone I have handy there is not room for the Terminal IDE. I'll tried with hacker's keyboard but no luck. I'll test with a more decent android pad soon.
Thanks,
Massimo
Comments
I should take the code I did for my one digit clock entry for the RTC on the PPDB and see if we can share the high level code with the different low level code.
You will notice up top where I aliased the low level i/o:
When it became apparent that the 1307 was 'slow' the sum of the edits were three "S"s one in front of each I2C i/o primitive.
Thanks for the RotaryEncoder words.
I transcribed one of Simply Tronics for the BS2 from PBASIC to TF, added a few things, the made it more Forth-like.
The concept is redirection of Console Output which means all the nice formatting words Peter created are available, as is EMIT to send unprintable values.
Can some one tell me the fundamental difference between version V2 and the original?
Regards,
Rich
I've lost track of that but when I started to change the way the kernel operated I wanted to break that off into a different version. I don't see the sense in zero point zero versions so I started with the first version which is of course V1. So V2 is just the platform for testing new enhancements while leaving V1 stable but V2 has proved stable from the start however.
So to sum up the enhancements:
- Loop branch parameters are stacked in cog RAM to allow for minimal looping overhead
- Faster dictionary searches
- Trimmed VM kernel and modularized special I/O instructions for SPI and SD etc.
- Extended direct call to VM opcodes using XOP to access top 256 longs.
- Removed possibility of absolute calls (was unused) and instead expanded the vector table handling.
- Updated serial receive to preprocess special control characters and sequences including fast-load filtering
- Optimized the number preprocessing for compilation speed
- Optimized the memory map etc
So don't even bother with V1 as all the extensions and examples are coded for V2 anyway.If you're not picky about what a language "should be" and like to get interactive with your hardware and get up-and-running as quick as possible as well as having fast compiled code then I can definitely recommend Tachyon.
Also I was curious what you think about DOES> - do you have plans to implement it, or do you find that you don't make use of it in embedded programming?
I haven't ruled out implementing DOES> as there have been a few times I could have used it effectively, so maybe sometime in the next few weeks after I've finished wrangled with a few things that didn't quite get done last month!
I will finish off judging the much delayed challenge and proceed to kick off the 4/4 challenge for 2013 as well which I hope I will have better organized this time
Peter added the LONG to this for me. I thought that I would repay the favor by running these scripts and copying the console. It should be self-explanatory. I know the visuals one after another like this really brought it home for me ... ENJOY
Brian, thanks for visually demonstrating how clever Peter is!!
[FONT=courier new] [COLOR=#000000]{ *** STEERING WHEEL INTERFACE ***[/COLOR] [COLOR=#000000]Measure the resistance of the steering wheel interface[/COLOR] [COLOR=#000000]Resistance is a maximum of 6K8 and open if no switches are pressed[/COLOR] [COLOR=#000000]RC time constant method is used to detect the resistance but to help there is[/COLOR] [COLOR=#000000]a 100K resistor added in parallel. [/COLOR] [COLOR=#000000]The resistance is referenced to ground and discharges the capacitor as this method offers good noise immunity.[/COLOR] [COLOR=#000000]PORT ------220R------|-------|-----> to SWI resistors[/COLOR] [COLOR=#000000] 104C[/COLOR] [COLOR=#000000] |------------>[/COLOR] [COLOR=#000000] _|_[/COLOR] [COLOR=#000000]}[/COLOR] [COLOR=#000000]#12 CONSTANT =scale [/COLOR] [COLOR=#000000]#P8 |< CONSTANT =swi[/COLOR] [COLOR=#000000]\ The maximum value read after conversion is 325 for the 100K resistor which is converted to zero[/COLOR] [COLOR=#000000]\ 10K reads around 15[/COLOR] [COLOR=#000000]pub [/COLOR][B][COLOR=#000000]SWI[/COLOR][/B][COLOR=#000000] ( -- val )[/COLOR] [COLOR=#000000] =swi OUTSET \ charge the capacitor[/COLOR] [COLOR=#000000] #100 us \ for long enough (prototype is using a series 220R resistor)[/COLOR] [COLOR=#000000] =swi INPUTS \ now float the line[/COLOR] [COLOR=#000000] CNT@ =swi WAITPNE \ and wait for it to go low[/COLOR] [COLOR=#000000] CNT@ SWAP - \ calculate the time constant [/COLOR] [COLOR=#000000] =scale SHR \ justify[/COLOR] [COLOR=#000000] DUP #32 > IF DROP 0 THEN[/COLOR] [COLOR=#000000] ;[/COLOR] [COLOR=#000000]\ IR Code [/COLOR] [COLOR=#000000]\ LOW (carrier) 12.5ms [/COLOR] [COLOR=#000000]\ HI (no carrier) 4.5ms[/COLOR] [COLOR=#000000]\ pulse low then 32 bits msb to lsb[/COLOR] [COLOR=#000000]#P24 CONSTANT =irled[/COLOR] [COLOR=#000000]pri [/COLOR][B][COLOR=#000000]IRLED[/COLOR][/B][COLOR=#000000] ( units -- )[/COLOR] [COLOR=#000000] A =irled APIN #38 KHZ \ Generate a 38kHz carrier[/COLOR] [COLOR=#000000] #500 * us[/COLOR] [COLOR=#000000] MUTE \ mute or stop the carrier[/COLOR] [COLOR=#000000] ;[/COLOR] [COLOR=#000000]pub [/COLOR][B][COLOR=#000000]IRTX[/COLOR][/B][COLOR=#000000] ( code -- )[/COLOR] [COLOR=#000000] 0 REV \ flip msbs into lsbs for right shifts[/COLOR] [COLOR=#000000] #25 IRLED \ 25 units of carrie[/COLOR] [COLOR=#000000] #4500 us [/COLOR] [COLOR=#000000] 1 IRLED[/COLOR] [COLOR=#000000] BL FOR [/COLOR] [COLOR=#000000] DUP 1 AND 2* 1+ #500 * us \ "high" period [/COLOR] [COLOR=#000000] 1 IRLED \ the "low" pulse [/COLOR] [COLOR=#000000] 2/[/COLOR] [COLOR=#000000] NEXT[/COLOR] [COLOR=#000000] DROP[/COLOR] [COLOR=#000000] ;[/COLOR] [COLOR=#000000]\ Test it out[/COLOR] [COLOR=#000000]$817EF00F IRTX[/COLOR] [COLOR=#000000]pub [/COLOR][B][COLOR=#000000]PWR[/COLOR][/B][COLOR=#000000] $817EF00F IRTX ;[/COLOR][/FONT]IR remote transmission format (this is the way I look at it):
[FONT=courier new] #P24 CONSTANT =irled BYTE family \ IR family code byte $81 family C! \ Set the default family code pub IRLED ( units -- ) A =irled APIN #38 KHZ \ Generate a 38kHz carrier #460 * us MUTE \ mute or stop the carrier ; pub IRTX8 ( byte -- ) 8 FOR DUP 1 AND IF #1530 ELSE #480 THEN us \ "high" period - a 1 is 3 times the length of a 0 high 1 IRLED \ the "low" pulse 2/ NEXT DROP ; \ Transmit the IR code byte as code + inverted \ First transmit the preamble then the family code + invert and then the IR code + invert pub IRTX ( code -- ) #20 IRLED \ HEADER PREAMBLE #4200 us 1 IRLED \ "start" bit family C@ DUP IRTX8 INVERT IRTX8 DUP IRTX8 INVERT IRTX8 ; [/FONT]Now I'm going to mangle the way the steering wheel buttons operate so that it does more than it does at present and it will sense sequences and duration for extra control.
I'm playing with the cheap stepper motors, and the interactive environment is great.
I'm using the wifly modules, and I ordered the BT ones too, remote programming is a nice option. I'm really having fun, I only need some time to adapt to the different approach and be a little more fluent in the language.
Thanks for the effort.
Massimo
Thanks Fidolo, and welcome to the Propeller forums. You might find those binaries have passed their "best before" date and the memory map has changed enough that the old binary dump methods don't work for now. In the meantime I have downloaded the binary image of the current kernel (Tachyon V2.1-130110 kernel-5MHZ.binary) compiled for a 5MHz crystal and defaulted to 230400 baud into the Dropbox account. When you have the kernel running then set Minicom to a 10ms line delay (^A T D) and copy and paste the EXTEND.fth file, either from the Dropbox or the online document. Of course you can compile the kernel from the source as well, it's very easy and it's only a single file
Some of the Forth words are unfamiliar but not so much if you are familiar with the Propeller chip. There are other words which I would probably change now but they're okay and so they've stuck.
If you have any suggestions for improving Tachyon or the intro page then please let me know.
PETER, you have probably had enough suggestions and requests from me to last two lifetimes, but I have been meaning to tender this request for a while ... modify WORDS and QWORDS to take a simple search string
" S" SEARCH or " IN" QSEARCH searches for left justified match of cmds to string. SEARCH ---> WORDS QSEARCH ---> QWORDS'
This will help the apprentices/rookies and will definitely help us journeymen.
I probably need both hands full of fingers to count the number of times in the last 5 weeks I have sent you some code only to have it returned with an "oh by the way, you can use this neat word in place of these 5!" like, for example "+!" .
QWORDS is helpful but a simple filter will go a long way to making Tachyon easier to learn.
Oh, by the way I have been wanting to put a filter into place but I just never got around to it
Rather than passing a string which would mean you would always have to do it or else it might pick up garbage from the stack, I would rather have WORD look for another word or string to describe further actions. This is fine for interactive terminal use whereas the string passing method is better for words which are meant to be called from other words.
So I would have WORDS operate this way if I wanted to find all the words with a + symbol for instance:
WORDS +
In this case wildcards are assumed without having to specify, this makes it easier to get down to business.
If I was looking for anything that started with + then it would be entered this way:
WORDS +**
Where star star (seems unique enough) is the wildcard but also forces a search only for words that begin with + in which case you would find +!
Now this brings me back to my original intention of storing help comments for each word which would be stored if EEPROM > 32K or SD or SPI Flash were available, this is what would happen.
WORDS +!
CFA CODE ATRS NAME DESC
6BD3 0061 ...$ +! ( long addr -- ) add long to long at hub addr
6BC6 005D ...$ W+! ( word addr -- ) add word to word at hub addr
6BB1 0059 ...$ C+! ( byte addr -- ) add byte to byte at hub addr
Now that would be even better wouldn't it?
Thanks Peter, I was just wondering about how to install your latest build, and you just answered my question before I even asked it!
I may have suggestions at some point, but I will not attempt to contribute to your effort until I have a decent familiarity with the extraordinary Propeller-Tachyon combination, which will take a while. I remember when it was estimated that there were about 50,000 Forth progammers in the world, roughly in the days of the FORTH-83 standard. Not being a software engineer by trade, Forth has nevertheless been essential to my professional career, which required me to build a lot of fast experimental hardware, to which Forth is so well adapted.
Since these days, I have been feeling a bit nostalgic, worried about Forth seemingly being progressively eclipsed, quite unfairly if one considers its unique ability to get the best out of hardware. Now I wonder whether Forth on the propeller chip may reverse the trend by inspiring robotic enthusiasts to get the most out of the propeller chip without going all the way down to the tedium exclusive assembly programming, and without any formal training in software engineering, thanks to the inner simplicity and the interactiveness of Forth?
Just to make this project more visible I will post a separate thread as well.
In this hardware P5 is connected to an NPN and P8 has a pulldown with pullups on P16..31 except for P26 which has a pulldown for reset control of a Silabs CPU..So an analysis of what PINS? reports and a comparison with the schematic indicates that this word is very useful.
Thanks.
Massimo
http://forums.parallax.com/showthread.php/138125-Erco-s-Figure-Eight-Challenge?p=1157894#post1157894
Nothing fancy, but I really like the interactive approach and the possibility to remotely program the propeller.
Ordered a couple of BT modules too.. :-)
Massimo
I was able to connect to Tachyon from an Android device, but I have problems finding a nice BT terminal for the phone.
Blueterm for instance is not sending the correct end of line.
I also cannot find a tool with the possibility to set the mandatory line delay.
Any suggestion?
Thanks in advance,
Massimo
Blueterm is the right one but you will find that the return problem has to do with the keyboard. Load "Terminal IDE" keyboard as this seems to be the best for this kind of stuff. I modified Tachyon the other week to accept a DEL key as well as the BKSP key because I couldn't find an Android keyboard that would actually send a $08 instead of a $7F or escape sequence.
Thanks,
Massimo