 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Microcontroller and PC projects : PicoMite V5.09.00 release candidates
  Page 5 of 13   |
|||||
| Author | Message | ||||
| matherp Guru  Joined: 11/12/2012 Location: United KingdomPosts: 9110 |
Painting a screen takes time. If you have a short first repeat then the keyboard handler times out the repeat timer before seeing the key-up and therefore adds another character. I don't understand the reason for a fast first repeat timer as it just encourages spurious extra characters. If you want a fast repeat for a game then there are better ways of doing this (keydown function) BLIT, transparent text Edited 2024-04-11 22:37 by matherp |
||||
| Bleep Guru Joined: 09/01/2022 Location: United KingdomPosts: 509 |
Yes it was primarily for Petscii where the players movement would stall after the initial step, but then fine once the repeat started, I'll look at using key up etc. I couldn't use that when the keyboard was connected through a terminal, so I used to set the terminal repeat to faster, to help. I've now set the resistor to 6.8k and don't now get a white screen, but there still appears to be something strange going on, which I realise doesn't matter when using noscroll, but I'm reporting it incase it can affect anything else. Using 'option lcdpanel console 7' when the cursor gets to the bottom of the screen, when the screen scrolls, because I hit return, two lines scroll not just the one, the first line scroll is reasonably fast the second is slower, similarly if the cursor is at the bottom of the screen and I type 'cls' it takes almost a second before the scrern clear happens, it appears that the command isn't actioned until the second line scroll is performed. Similarly if I set NOSCROLL, when the screen clears because it was full, I get two blank lines at the top of the screen, or if I do 'cls' again I get two blank lines at top of screen, to be sure I have set keyboard repeat to 500,200, otherwise all options as before. Regards Kevin |
||||
thwill Guru Joined: 16/09/2019 Location: United KingdomPosts: 4036 |
Hello Peter, As a result of my recent experimentation with the Wii Nunchuk can I suggest the following for your consideration: 1. Add Nunchuk support to the DEVICE command since you already know how to do it  .2. and/or add MM.INFO(OPTION SYSTEM I2C) so that a program can determine whether OPTION SYSTEM I2C is I2C or I2C2. 3. and/or add command I2CSYS which behaves as I2C or I2C2 as appropriate to the setting of OPTION SYSTEM I2C. 4. Have DEVICE WII OPEN check the id of any attached device and report an error if it isn't correct, like you do on the CMM2. 5. Revisit the "debug style" messages in cmd_Classic: "Wii device not connected<N>" Thanks, Tom Game*Mite, CMM2 Welcome Tape, Creaky old text adventures |
||||
| Volhout Guru Joined: 05/03/2018 Location: NetherlandsPosts: 4223 |
Hi Peter, I have a few remarks on the PicoMite code. Maybe you can enlighten me  1/ Reading from I2C is limitted to 255 values in one command. Any reason why not more? I am reading pixel data from a thermal camera, 256 bytes.. And just for the last byte I need an extra read. 2/ MATH WINDOW shows strange behaviour when input matrix and output matrix are not the same type . dim a!(1,1) = (1.1,1.2,1.3,1.4) dim b(1,1) dim c%(1,1) math window a!(),10,20,b() math window a!(),10,20,c%() print b(0,0),b(1,0) print b(0,1),b(1,1) print c%(0,0),c%(1,0) print c%(0,1),c%(1,1) The conversion to integer gives false results. I wonder if this is changed, since I recall this was used to scale ADC values (0-3.3V) to the LCD window, in the oscilloscope. LCD pixels are integer locations. Regards, Volhout PicomiteVGA PETSCII ROBOTS |
||||
stanleyella Guru Joined: 25/06/2022 Location: United KingdomPosts: 2122 |
OPTION BASE 0 | 1 Set the lowest value for array subscripts to either 0 or 1. This must be used before any arrays are declared and is reset to the default of 0 on power up. 0 to 255 is 256 |
||||
| Amnesie Guru Joined: 30/06/2020 Location: GermanyPosts: 396 |
@ Volhout a thermal camera? This sounds interesting. I suppose you have one with a low resolution (8x8 or 16x16) and therefore it even makes sense to visualize it in colour or even interpolate? If you have some pictures or tests maybe it is worth a new topic for this (as there are more people interested in this)  Greetings Daniel Edited 2024-04-25 01:37 by Amnesie |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 9110 |
1. is a limitation relating to the size of strings currently in all versions of MMbasic but I can't see why it needs to be there for integer or float arrays 2. is a bug What version are you using? Let me know and I'll post a test version |
||||
| Volhout Guru Joined: 05/03/2018 Location: NetherlandsPosts: 4223 |
Peter, I am using the VGA version, so I can easily display the thermal image. On 640x480 using tiles. There may be a LCD version derivative possible. It is using a grove module (seed studio) Volhout Edited 2024-04-25 02:08 by Volhout PicomiteVGA PETSCII ROBOTS |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 9110 |
Try this PicoMiteVGA (3).zip |
||||
| Volhout Guru Joined: 05/03/2018 Location: NetherlandsPosts: 4223 |
Hi Peter, This works fine. I see that you are doing the INT, and not a CINT on the values when converting from float to integer, but that is fine. I can wokr with that. Thank you for fixing, Volhout PicomiteVGA PETSCII ROBOTS |
||||
| Volhout Guru Joined: 05/03/2018 Location: NetherlandsPosts: 4223 |
Hi Peter, In playing with the PIO assembler, I noticed some quirks. Maybe something you can fix for 5.09.00. 1/ The codeword "pins" in the move instruction needs a trailing <space>. Otherwise the assembler will not assemble it. pio assemble 1,"mov Y, pins " 2/ The PUSH instruction can have a BLOCK or NOBLOCK extension. pio assemble 1,"push" compile correctly to &h8000 pio assemble 1,"push noblock" compiles to &h8020 (meaning it blocks). this should be &h8000, same as above. pio assemble 1,"push block" should compile to &h8020. Thank you for your attention, Volhout PicomiteVGA PETSCII ROBOTS |
||||
| Volhout Guru Joined: 05/03/2018 Location: NetherlandsPosts: 4223 |
Peter, I am not sure if this is simple or hard to do. But I have a question about the ADC OPEN command. It uses GP26 and further. But GP26 is the ADC start channel. The GEEK unit I have has GP28 and GP29 on the connector pins. I tried "surgery" to re-wire to GP26 and 27, but it is not do-able. The traces on the board run under the LCD that is glued to the board. Is there any way to add the starting channel to the command (as an option)? Or does this impact all other commands (ADC START/ADC RUN), and adds more risk than benefit. This is no priority at all, I do not have a project waiting for it(*), but when you feels this helps MMBasic... Volhout (*) I always have a pico lying on my desk that I could use with GP26/27/28.... Edited 2024-04-26 17:48 by Volhout PicomiteVGA PETSCII ROBOTS |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 9110 |
Harm Thanks for the various reports and suggestions - do keep them coming. I won't be able to look at anything for a week or so but will see what is possible then. I also need to get the documentation up-to-date with the USB and other 5.09.00 changes  |
||||
| Volhout Guru Joined: 05/03/2018 Location: NetherlandsPosts: 4223 |
Hi Peter, I found another quirk with the PIO assembler (see also thread about the quadrature decoder). A program starts and ends with PIO ASSEMBLE x,".program NAME" - - - PIO ASSEMBLE x,".end program" The last line can also be PIO ASSEMBLE x,".end program list" In this case the console shows a listing of the hex codes the assembler produced. In case you use the "list" option, the program works as designed. In case you omit the "list" option, the program does not work. Apparently there is something wrong when the list option is not used. Volhout PicomiteVGA PETSCII ROBOTS |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 9110 |
I see in the other thread you have padded the assembler to the full 32 commands. What happens if you remove all the spurious padding now you have fixed the Y initialisation? |
||||
| Volhout Guru Joined: 05/03/2018 Location: NetherlandsPosts: 4223 |
Peter, When I check the memory adresses where the code has to be written, there are only 0x00's. It feels like code is only actually written to PIO memory when you add the "list" option. The code can be found by (first instruction , should be &h001A) print hex$(peek(word &h50300048)) Successive instructions words are found at &h5030004c &h50300050 /54/58/5c etc... Y is not initialized. The first instruction is that Y (2bits) are copied to ISR. And when it is random, anything can be put in ISR (but only 2 bits). So I think the padding is not really needed. Volhout Edited 2024-04-30 17:53 by Volhout PicomiteVGA PETSCII ROBOTS |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 9110 |
OK, Another one to add to the list - thanks |
||||
| Volhout Guru Joined: 05/03/2018 Location: NetherlandsPosts: 4223 |
Hi Peter, Now I understand why this problem has hidden for so long. It also exists in 5.08.00. When you develop code, you use the "list" option to see where the assembler errors. And when the code works, the "list" option is removed, and you focus on the MMBasic code. But as long as you do not power down, the PIO code is still in memory. Once you power down, or do a PIO CLEAR, the code vanishes. And then you have no code in memory anymore. I have proven this to work. Run the program with "list" option. Stop. Remove "list" option. Comment the PIO CLEAR out. Run again, and the program still runs. Uncomment the PIO CLEAR, and RUN. The program does not work anymore. Add the "list" option: program runs correct. Volhout Edited 2024-04-30 18:05 by Volhout PicomiteVGA PETSCII ROBOTS |
||||
| Bleep Guru Joined: 09/01/2022 Location: United KingdomPosts: 509 |
Hi Peter, Some advice if you can, I am experimenting with sound, using wav files, I am playing a compressed mono wav file, but all variants behave the same, from both A: or B:, while using a background frame copy, at 320x240, it is just about ok a few glitches, however at 480x320 the sound runs slow. CPU speed doesn't seem to make any difference. Is this a limitation of keeping the PWM outputs running, while also keeping the screen updated? or is there anything that can be done to help? this is using a usb version and ILI9488, which I know you said was slow. Any insights appreciated. Regards Kevin. 2 sample wav files, one stero one mono wav.zip Test program, search for 'Experimenting' comment, to change wav/screen size. clockclock.zip Edited 2024-05-02 01:36 by Bleep |
||||
| Volhout Guru Joined: 05/03/2018 Location: NetherlandsPosts: 4223 |
Dear Peter, I would like to issue 2 more items for the 5.09.00 release 1/ When writing to PIO FIFO you have to supply a 32 bit integer value. Avoiding to do so (i.e. a 64 bit negative integer value) MMBasic shows following error. [270] PIO write 1,0,1,value% Error : xxx is invalid (valid is 0 to -1) Maybe this can be fixed, by masking the 32 lsb ? 2/ When adding "MATH C_MUL a(),a(),a()" we created the opportunity to square all values in each cell of the matrix a(). That is very helpfull. Is there a similar way to take the square root, or just any root (3,4,5,6..) using these fomula's? Maybe a command like "MATH POWER a(),factor,b()" could be created, where the contect of each cell is raised to the power "factor". That would allow square, but also square root (factor is =0.5) or 3-root (factor is 0.33333) ???? This would be helpfull for the thermal camera. There each pixel has to be raised to the power 3, then multiplied, then the cube root must be taken. For each pixel. Thank you for your attention, Volhout Edited 2024-05-16 18:02 by Volhout PicomiteVGA PETSCII ROBOTS |
||||
| Page 5 of 13 |
|||||
 Print this page |
