 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 : I2C LCD
  Page 3 of 4   |
|||||
| Author | Message | ||||
| MOBI Guru  Joined: 02/12/2012 Location: AustraliaPosts: 819 |
I think it is time to call it quits on this one. No point in bashing one's head against the wall. I have tried every trick in the book, both logical and il but can not get the second 8 bit port to do anything nor can I get either port to read in pin status. Anything I try and do on port b shows up only on port a. I have scoured the web - many people having problems but no solutions that made sense. One chap alluded to a vital command that is not well known but the solution was not code so couldn't work out from it what to do. The MCP23018 looks like a mongrel chip. If anyone has used one successfully, I'd love to know. It all looks so simple. Every other I2C device I have whether bought or home grown all work fine. Just Dunno. David M. |
||||
jman Guru Joined: 12/06/2011 Location: New ZealandPosts: 711 |
No don't quit yet MCP23018 which is basically the same as the 23017 except it has an open collector and 5v tolerant pins so take a look here MCP23017 And here More MCP23017 Regards John |
||||
| MOBI Guru Joined: 02/12/2012 Location: AustraliaPosts: 819 |
Hi John, That initialise routine is much the same as I have been trying. Anyway, I set it all up the same as per your routine and get absolutely zip. All the commands return MM.I2C "0". Beginning to think it is a hardware problem. I made a new chip carrier board and soldered in the second chip with the same results. I set up the circuitry to accommodate the open drain ports. I'll download the mcp23017 datasheet and see if there is any difference. Last resort. I'M fair flummoxed! David M. |
||||
| MOBI Guru Joined: 02/12/2012 Location: AustraliaPosts: 819 |
Just as an aside, The device is in default bank 0 on POR or /reset. So, writing &HA0 to register &H0A will set the Bank bit (7) to 1 which will put it in bank 1 mode. Then writing to register &H0B appears to do nothing as there is no register &H0B in bank 1. puzzled. ====== I wrote &HA0 to all three iocon regs 0A, 0B and 05 and now I'm getting activity on port b. Haven't had that before. Didn't see that in the datasheet. Too late at night - continue tomorrow. David M. |
||||
MicroBlocks Guru Joined: 12/05/2012 Location: ThailandPosts: 2209 |
Aha progress! That little chip must obey! Unfortunately i am not much of help as i did not use one of those chips in my own coding. Thanks Mobi for figuring this thing out. I am sure many will benefit from your efforts. Microblocks. Build with logic. |
||||
| jman Guru Joined: 12/06/2011 Location: New ZealandPosts: 711 |
Hi To get that little chip to obey :) try this Init23017: I2CSEND i2caddr, 0, 2, &H0A, &B10100000 I2CSEND i2caddr, 0, 2, &H0B, &B10100000 I2CSEND i2caddr, 0, 2, &H05, &B10100000 I2CSEND i2caddr, 0, 2, &H00, &B00111111 I2CSEND i2caddr, 0, 2, &H06, &B00111111 I2CSEND i2caddr, 0, 2, &H10, &B00000000 I2CSEND i2caddr, 0, 2, &H19, &B00000000 Return Then use this to send and check the status I2CRCV i2caddr, 0, 1, Buttons, 1, &H09 ' This will check port A inputs I2CSEND i2caddr, 0, 2, &H19, temp ' This will write to port B I2CSEND i2caddr, 0, 2, &H09, temp ' This will write to port B Regards John |
||||
Grogster Admin Group  Joined: 31/12/2012 Location: New ZealandPosts: 9308 |
Obey! Obey or you will be exterminated! Smoke makes things work. When the smoke gets out, it stops! |
||||
| MOBI Guru Joined: 02/12/2012 Location: AustraliaPosts: 819 |
Now where did I leave that sonic screwdriver? John, thanks for the code - I had a look at your earlier posting - I must confess I didn't look at your links at the time. I extracted the appropriate section last night and that is when I started getting activity on port b (for the first time). The only thing I was doing differently was doing all operations in bank 0 (I'm convinced it still can) Anyway one headache at a time - get it working first. It all appears to be doing what it should but the open drain system means lots of resistors and leds etc. I've also come to the conclusion that the chip does not like "noise". Floating pins seem to be a problem and the system locks up. I commented out lines 2 and 3 and it all still works. Port b seems to NEED pullups on. I'm going to redo the mock-up so that I can write data to a port and read it back from the other port and display on screen (or i2c display?). Fingers crossed. I2CSEND i2caddr, 0, 2, &H0A, &B10100000 'select bank 1 mode (separate registers) and non-sequential read. 'I2CSEND i2caddr, 0, 2, &H0B, &B10100000 'this would do nothing as address &H0B now is now not in bank 1 'I2CSEND i2caddr, 0, 2, &H05, &B10100000 'same as for line 1 - because of iocon register pairing, so not needed I2CSEND i2caddr, 0, 2, &H00, &B00111111 'pin data direction port a 0=output 1=input I2CSEND i2caddr, 0, 2, &H10, &B00000000 'pin data direction port b I2CSEND i2caddr, 0, 2, &H06, &B11111111 'port a weak pullups I2CSEND i2caddr, 0, 2, &H06, &B11111111 'port a weak pullups I2CSEND i2caddr, 0, 2, &H19, &B00000000 'clear port b outputs Return Then use this to send and check the status I2CRCV i2caddr, 0, 1, Buttons, 1, &H09 ' This will check port A inputs I2CSEND i2caddr, 0, 2, &H19, temp ' This will write to port B I2CSEND i2caddr, 0, 2, &H09, temp ' This will write to port A David M. |
||||
| MOBI Guru Joined: 02/12/2012 Location: AustraliaPosts: 819 |
Hi Jman Thanks for the inspiration. I think all my problems were noise related.I've swept the floor and collected all the hair. Now who is that hair transplant mob? I commented out the weak pullups as they shouldn't be needed as port b is output with resistor and led external pull ups and port a is tied to logic levels from port b. IOCON only needs initial set up at &H0a as the registers concerned are shared. The numbers on the 7seg display are just 1 2 4 8 and nothing to do with the expander, just an indicator that the i2c bus was working so I left it in. The expander is in that block of connectors in the foreground. 
Mode 4
Cls 4 x7segdisp = &h38 'i2c device address (LED) expander = &H20 'i2c address expander Dim inpina 'data read from 8bit wide port a Dim inpinb 'in case needed dim number 'value 0 to 255 entered at keyboard 'Bank 1 mode iodira = 0 iodirb = &H10 iocon = &H0a 'ioconb = &H0b gppua = &H06 gppub = &h16 gpioa = &H09 gpiob = &H19 olata = &H0a olatb = &H1a I2CEN 100,100 'enable the i2c bus I2CSEND expander,0,2,iocon,&Ha0 'set config bank 1 I2CSEND expander,0,2,iodira,&Hff 'port a inputs I2CSEND expander,0,2,iodirb,&H00 'port b outputs 'I2CSEND expander,0,2,gppua,255 'port a pullups 'I2CSEND expander,0,2,gppub,255 'port b pullups I2CSEND expander,0,2,gpiob,&Hff 'clear port b output I2CSEND disp,1,5,0,1,2,4,8 start: Input "Enter number 0 to 255 ",number Cls I2CSEND expander,0,2,gpiob,number I2CRCV expander,0,1,inpina,1,gpioa Print "The number sent is " number 'send byte to port b Print "The number received is " inpina 'receive echo byte from port a ? ? GoTo start David M. |
||||
| MOBI Guru Joined: 02/12/2012 Location: AustraliaPosts: 819 |
The expander also works in bank 0 mode just by substituting the corresponding bank0 register addresses. Moral of the story, NEVER leave pins floating. I got lazy and tried to use just the one pull up resistor and LED on the expander port pins one at a time. I guess in the past I just got lucky. The same thing applies to decoupling capacitors ALWAYS fit one to the vdd vss pins. I have been caught with that one even though mostly I've got away with it. The long and the short of it - all works on the expander just as it looked like it should have. Just call me baldy now.   David M. |
||||
| jman Guru Joined: 12/06/2011 Location: New ZealandPosts: 711 |
Glad to hear it's all good John |
||||
| MicroBlocks Guru Joined: 12/05/2012 Location: ThailandPosts: 2209 |
Great news! Like a movie with a satisfying end. :) Thanks for the hard work! Microblocks. Build with logic. |
||||
| MOBI Guru Joined: 02/12/2012 Location: AustraliaPosts: 819 |
@TZA Now that has put the data expander to bed, we can start on turning the PIC16F1503 into an I2C slave and see if we can get it to act as an I2C LCD display. I cleaned up (syntax wise) a F88 programme I had and successfully dumped it via MPLABX IDE to the 1503. Of course, it doesn't do anything yet. First I'll flash a led to let me know it is working and then I'll fix up the I2C slave mode and flash another led. Different registers, different location and names. Let you know how I go along the way. 
Picture is a bit fuzzy but that's hardly surprising. David M. |
||||
| MOBI Guru Joined: 02/12/2012 Location: AustraliaPosts: 819 |
Pic Slave. Ok, so I didn't flash a LED but a square wave on the cro is as good as a nod to a blind horse. The internal clock tripped me up a bit, needing a register and config words both set up right. Now for the I2C bit. ....... David M. |
||||
| MOBI Guru Joined: 02/12/2012 Location: AustraliaPosts: 819 |
Has anyone ever tried to make sense of the Microchip spi/i2c section of the data sheet for 8 bit pic micros and came away with a positive result and their sanity? Whoever writes their stuff should be stuffed. I went through it a few years ago and got it it working (after many days) but didn't document my source code as intensely as I should have, so currently doing a re-learn. The source code works ok, but if I want someone else to understand it too, I gotta go back to the registers concerned and work out what each bit/byte does. When writing their data sheet, Microchip tried to do it in narrative form to make it simple, but in the process made it too long, too complex and vague/ambiguous. I guess it is not the last time I'll growl at Microchip! David M. |
||||
| Grogster Admin Group Joined: 31/12/2012 Location: New ZealandPosts: 9308 |
They certainly like their datasheet writing! I suppose that is a good thing, as PERHAPS(?) better to have an overdose of information, then a lack of it... But I hear what you are saying. I seem to recall that the PDF for the PIC32 used in the full MM is about 500 pages - heck! 
Billy Connolly has mentioned in his act, that Americans like to give an overdose of informations. He said: "If you buy a transistor radio in America, there's 18 pages about how to turn the f***ing thing on!" Smoke makes things work. When the smoke gets out, it stops! |
||||
vasi Guru Joined: 23/03/2007 Location: RomaniaPosts: 1697 |
Nope! I never liked that mix. When I saw the rivals (I won't name them) counterpart, I thought that is impossible to be that simple and still functional. A good solution for me is to use ready made libraries and examples. Hobbit name: Togo Toadfoot of Frogmorton Elvish name: Mablung Miriel Beyound Arduino Lang |
||||
| MOBI Guru Joined: 02/12/2012 Location: AustraliaPosts: 819 |
I guess you are talking about Atmel. If so, I found them much easier to understand and usually had C code examples - much easier (for me anyway) to understand. I really should have taken up with AVR in the first place but as much as I don't like Microchip's abysmal documentation, I really am quite fond of the PIC and its physical and instruction simplicity. David M. |
||||
| JohnS Guru Joined: 18/11/2011 Location: United KingdomPosts: 3805 |
Over 1000, but split into about 30 sections - one each for the major features (PMP, UART, SPI, USB, DMA, and so on). The many pages explaining every peripheral port's bit settings seem useful and form most of it. It could be much smaller if it failed to tell you what is what (many companies follow this philosophy) or didn't provide so many peripherals (which would be silly). SPI is 60 pages. Examples are generally in C. I'm not going back to PIC16/18 without a very compelling reason. John |
||||
| MOBI Guru Joined: 02/12/2012 Location: AustraliaPosts: 819 |
I don't blame you, but some of us (maybe not on this forum) actually like playing with assembly language and to keep it simple, an 8bit pic does nicely if you don't mind masochistically bashing your head against Microchip documentation brick walls. I had another little hiccup over the last 24hours - in translating the F88 pic source to F1503 as an I2C slave. I foolishly chose an I2C address for the chip of 255 (hex FF) and wondered why the MM wouldn't talk to it. Forgot MM does range checking. Changed the address to &b00111111 and voilla! it works. Now to see if I can make the code into a LCD slave. Haven't checked if I have enough pins as the pic is only 14 pin. David M. |
||||
| Page 3 of 4 |
|||||
 Print this page |
