Post by thread:[PIC]: Doing modulated IR with 16F628 internal clo

I wanted to hopefully caution someone against making the same mistake I made. In the interest of reducing part count, I have been working on using a 16F628 to do modulated (38 Khz) IR. I wanted to have everything on one chip, instead of having to use multiple chips. I did the modulation and bit-banging myself, for 1200 baud IR communication. I was using the internal oscillator of the 16F628. Basically, I could get it working fine, and then I would come back the next morning, and it would be slightly off, missing a few bits on certain values, while working fine on others. I messed with this for two days, trying to tweak things to make them work correctly. It was very frustrating. And then i though about the temperature variations, and the internal clock. Arggh! I added a ceramic resonator, and everything worked flawlessly. I though that the internal clock would have been accurate enough for something as coarse as 1200 baud serial, but apparently it is not. I've attached the source code for the IR function below, in case anyone is interested. I am using the CCS compiler, and a 4 MHz clock. The 16/4 loop makes a 38 KHz pulse. Alex void IRWrite(byte b) { disable_interrupts(INT_EXT); #use fast_io(b) // start bit for(i=0;i<27;i++) { output_low(PIN_B2); delay_us(16); output_high(PIN_B2); delay_us(4); } // serials for(j=0; j<8;j++) { if(!shift_right(&b,1,0)) { // output a modulated zero for(i=0;i<27;i++) { output_low(PIN_B2); delay_us(16); output_high(PIN_B2); delay_us(4); } } else // output a one { output_high(PIN_B2); delay_us(760); } } // stop bit output_high(PIN_B2); delay_us(760); #use standard_io(b) enable_interrupts(INT_EXT); } -- http://www.piclist.com hint: To leave the PICList spam_OUTpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu>

Alex wrote: > I wanted to hopefully caution someone against making the same mistake I >made. In the interest of reducing part count, I have been working on >using a 16F628 to do modulated (38 Khz) IR. I wanted to have everything >on one chip, instead of having to use multiple chips. I did the >modulation and bit-banging myself, for 1200 baud IR communication. I >was using the internal oscillator of the 16F628. Basically, I could get >it working fine, and then I would come back the next morning, and it >would be slightly off, missing a few bits on certain values, while >working fine on others. I messed with this for two days, trying to >tweak things to make them work correctly. It was very frustrating. >And then i though about the temperature variations, and the internal >clock. Arggh! I added a ceramic resonator, and everything worked >flawlessly. As a matter of interest, what range of temperature was this over? >I though that the internal clock would have been accurate enough for >something as coarse as 1200 baud serial, but apparently it is not. Just a note on the baud rate. The accuracy required is not related to the baud rate. 5% at 1200 baud is still the same as 5% at 115200 baud as it's still relative to the chosen baud rate. David... ___________________________________________ David Duffy Audio Visual Devices P/L U8, 9-11 Trade St, Cleveland 4163 Australia Ph: +61 7 38210362 Fax: +61 7 38210281 New Web: http://www.audiovisualdevices.com.au ___________________________________________ -- http://www.piclist.com hint: To leave the PICList .....piclist-unsubscribe-requestKILLspam@spam@mitvma.mit.edu>

[snip about change in clock over temperature] > > As a matter of interest, what range of temperature was this over? > Not much. Maybe 70 low to 85 high. Austin, TX. No real winter here, plus I have a bunch of computers. :-) > >I though that the internal clock would have been accurate enough for > >something as coarse as 1200 baud serial, but apparently it is not. > > Just a note on the baud rate. The accuracy required is not > related to the baud rate. 5% at 1200 baud is still the same > as 5% at 115200 baud as it's still relative to the chosen > baud rate. David... > > Hmm. I guess you are right. I was getting confused with the receiver. If the receiver clock is off, then it should be able to handle 1200 baud easier than 115200. Alex -- http://www.piclist.com hint: To leave the PICList piclist-unsubscribe-requestKILLspammitvma.mit.edu>

> -----Original Message----- > From: Alex Kilpatrick [SMTP:.....AlexKKILLspam.....HCITRAINING.COM] > Sent: Wednesday, March 12, 2003 11:57 PM > To: EraseMEPICLISTspam_OUTTakeThisOuTMITVMA.MIT.EDU > Subject: [PIC]: Doing modulated IR with 16F628 internal clock > > > I've attached the source code for the IR function below, in case anyone > is interested. I am using the CCS compiler, and a 4 MHz clock. The > 16/4 loop makes a 38 KHz pulse. > > Alex > Out of interest why didn't you use the CCP module to generate the 38KHz carrier with zero CPU overhead? Regards Mike ======================================================================= This e-mail is intended for the person it is addressed to only. The information contained in it may be confidential and/or protected by law. If you are not the intended recipient of this message, you must not make any use of this information, or copy or show it to any person. Please contact us immediately to tell us that you have received this e-mail, and return the original to us. Any use, forwarding, printing or copying of this message is strictly prohibited. No part of this message can be considered a request for goods or services. ======================================================================= Any questions about Bookham's E-Mail service should be directed to postmasterspam_OUTbookham.com. -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

>Hmm. I guess you are right. I was getting confused with >the receiver. If the receiver clock is off, then it should >be able to handle 1200 baud easier than 115200. I suspect your real problem was not the 1200 baud drift, but the 38KHz carrier drift going outside the IR receiver passband. -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

>[snip about change in clock over temperature] > > > > As a matter of interest, what range of temperature was this over? Alex: >Not much. Maybe 70 low to 85 high. Austin, TX. No real winter here, >plus I have a bunch of computers. :-) That should be fine. I have a design that used a 12C509 (now a 12C672) and runs quite happily at 9600 baud with the internal RC oscillator. {Quote hidden}> > >I though that the internal clock would have been accurate enough for > > >something as coarse as 1200 baud serial, but apparently it is not. > > > Just a note on the baud rate. The accuracy required is not > > related to the baud rate. 5% at 1200 baud is still the same > > as 5% at 115200 baud as it's still relative to the chosen > > baud rate. David... > >Hmm. I guess you are right. I was getting confused with the receiver. >If the receiver clock is off, then it should be able to handle 1200 baud >easier than 115200. No, it still doesn't matter. The percentage error is no different with the different baud rates no matter how you look at it. You can probably tolerate a couple of percent at each end of the link and still be ok. Maybe like Alan said, the modulation is not really 38KHz although you can usually be a little off and only suffer reduced range as you start to get out of the receiver's pass band. See if you can get a look at the IR or demodulator output on the CRO to check your timing. Heating the board with a hair drier to see if it wanders. David... ___________________________________________ David Duffy Audio Visual Devices P/L U8, 9-11 Trade St, Cleveland 4163 Australia Ph: +61 7 38210362 Fax: +61 7 38210281 New Web: http://www.audiovisualdevices.com.au ___________________________________________ -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> I though that the internal clock would have been accurate enough for > something as coarse as 1200 baud serial, but apparently it is not. Where did you get the idea that asynch is 'coarse'? It requires a clock that is accurate at least to a few %, preferrably better than 1%. And this requirement is the same regardless of the baudrate. Wouter van Ooijen -- ------------------------------------------- Van Ooijen Technische Informatica: http://www.voti.nl consultancy, development, PICmicro products -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> I suspect your real problem was not the 1200 baud drift, but the 38KHz > carrier drift going outside the IR receiver passband. Not likely, a 5% shift will kill asynchronous communication, but the common IR receivers that I know [ http://www.vishay.com/docs/82030/82030.pdf ] a 5% shift will reduce the sensitivity to 60%, hence the effective distance to 77%. Wouter van Ooijen -- ------------------------------------------- Van Ooijen Technische Informatica: http://www.voti.nl consultancy, development, PICmicro products -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> Strange. I did just the same application (transmitter only] a > couple weeks > ago without any problem. I calibrated the oscilator by myself > (used 12C508 > with a window), the speed I use is 2400 baud with 1%... I guess YMMV depending on the PIC type, revision, maybe even the production lot, the temperature, Vcc variations, the phase of the moon and whether you have properly waved a dead fish. Wouter van Ooijen -- ------------------------------------------- Van Ooijen Technische Informatica: http://www.voti.nl consultancy, development, PICmicro products -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

pic microcontroller discussion list <> wrote on Thursday, March 13, 2003 2:10 PM: > I've had success with 19531bps masquerading as 19200bps. > 1.7% worked well A typical UART looks for the rising edge of the start bit, then samples the signal a couple of times around where the centre of each bit should be. For example, it could look at 25%, 50% and 75% of the way through each bit. That helps to get rid of glitches. However it means that by the end of the byte (about 10 bit slots later), the receivers guess of when the bit starts and stops has to be within 25% of a bit period, so each bit can slip by 2.5% of a bit period. For the next byte, the UART will re-synchronise on the next start bit, so errors aren't accumulated. In other words, for the hypothetical UART described above, the clocks at the receiver and transmitter have to be in agreement within 2.5%, so each one needs to be within 1.25% of the correct value to make comms reliable. Different UARTs use different sampling schemes, and could even resynchronise during the byte (do any do this?), but assuming the clocks have to be accurate to less than 1% is a good place to start. If both clocks are 30ppm crystals, their maximum combined error at room temperature is 0.06%, so you can get away with a bit rate error of 1.7%. Add a significant temperature difference between the crystals (some of what we do here has > 150 deg C between them!) and your 19531 might not work any more :-) Nigel -- Nigel Orr, Design Engineer @spam@nigelKILLspamaxoninstruments.co.uk Axon Instruments Ltd., Wardes Road,Inverurie,Aberdeenshire,UK,AB51 3TT Tel:+44 1467 622332 Fax:+44 1467 625235 http://www.axoninstruments.co.uk -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> > > > Strange. I did just the same application (transmitter only] > a couple > > weeks ago without any problem. I calibrated the oscilator by myself > > (used 12C508 > > with a window), the speed I use is 2400 baud with 1%... > > I guess YMMV depending on the PIC type, revision, maybe even > the production lot, the temperature, Vcc variations, the > phase of the moon and whether you have properly waved a dead fish. > Right. And the problem is multiplied if you are doing PIC to PIC communication, because the variations in clocks may add. I do think the internal clock on the 16F628 is great for non-critical timing applications, though. Alex -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> -----Original Message----- > From: Wouter van Ooijen [KILLspamwouterKILLspamVOTI.NL] > Sent: Thursday, March 13, 2003 6:48 AM > To: RemoveMEPICLISTTakeThisOuTMITVMA.MIT.EDU > Subject: Re: [PIC]: Doing modulated IR with 16F628 internal clock > > > > I suspect your real problem was not the 1200 baud drift, > but the 38KHz > > carrier drift going outside the IR receiver passband. > > Not likely, a 5% shift will kill asynchronous communication, > but the common IR receivers that I know [ > http://www.vishay.com/docs/82030/82030.pdf ] a > 5% shift will > reduce the sensitivity to 60%, hence the effective distance to 77%. > Right. I was checking the signal on the decoder, and all the bits were there, it was just variations in the bit times messing up the receiver's ability to interpret the serial stream. Alex -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> -----Original Message----- > From: Michael Rigby-Jones [spamBeGoneMichael.Rigby-JonesspamBeGoneBOOKHAM.COM] > Sent: Thursday, March 13, 2003 1:57 AM > To: TakeThisOuTPICLISTEraseMEspam_OUTMITVMA.MIT.EDU > Subject: Re: [PIC]: Doing modulated IR with 16F628 internal clock > > > > I've attached the source code for the IR function below, in case > > anyone is interested. I am using the CCS compiler, and a 4 > MHz clock. > > The 16/4 loop makes a 38 KHz pulse. > > > > Alex > > > Out of interest why didn't you use the CCP module to generate > the 38KHz carrier with zero CPU overhead? > Umm, I wanted to build character? :-) I guess I will have to learn what the CCP module is now. I just never looked at it before, coming from the 16F84 world. Alex -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

{Quote hidden}> -----Original Message----- > From: David Duffy [RemoveMEdavidTakeThisOuTAUDIOVISUALDEVICES.COM.AU] > Sent: Thursday, March 13, 2003 4:50 AM > To: PICLISTEraseME.....MITVMA.MIT.EDU > Subject: Re: [PIC]: Doing modulated IR with 16F628 internal clock > . > > > >Hmm. I guess you are right. I was getting confused with > the receiver. > >If the receiver clock is off, then it should be able to handle 1200 > >baud easier than 115200. > > No, it still doesn't matter. The percentage error is no > different with the different baud rates no matter how you > look at it. You can probably tolerate a couple of percent at > each end of the link and still be ok. Maybe like Alan said, > the modulation is not really 38KHz although you can usually > be a little off and only suffer reduced range as you start to > get out of the receiver's pass band. See if you can get a > look at the IR or demodulator output on the CRO to check your > timing. Heating the board with a hair drier to see if it > wanders. David... > > I did check the output of the decoder on a logic analyzer. All the bits were there. But I was getting glitches on the receiver. (f -> e, 1 -> 8, etc.) Here were the symptoms: 1) Tweak the timing. Everything works. Get up the next morning, nothing changed, but now I get glitches. 2) Got glitches. Didn't change code Replaced the internal oscillator with a ceramic resonator. Everything works perfectly. That certainly seems like a internal clock problem to me. ALex -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

Neat. I'll be sure to avoid using my Palm IIIxe at 150C. :) My purpose was debugging and troubleshooting a hard real-time device the company is developing for a customer. Just an extra trace to the ICSP connector, and I can hook in an LED and see which fail-safe has kept the device from smoking. I was already updating a DAC at 19531, so bit-banging the IR at that rate was no problem; getting it exact would have caused no end of troubles. How hard would it be to bit-bang the receive, with absolutely minimal hardware? Jason Neudorf {Original Message removed}

> of a bit period, so each bit can slip by 2.5% of a bit But you might also need to take into account that - the uart might only sample the input at for instance 4 times the baudrate, which might cause an additional error of 1/4 baud - the electrical part of the communication might add timing errors Wouter van Ooijen -- ------------------------------------------- Van Ooijen Technische Informatica: http://www.voti.nl consultancy, development, PICmicro products -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> > > > of a bit period, so each bit can slip by 2.5% of a bit > > But you might also need to take into account that > - the uart might only sample the input at for instance 4 > times the baudrate, which might cause an additional error of 1/4 baud > - the electrical part of the communication might add timing errors > > Wouter van Ooijen > I'm using the CCS compiler, which has a 'getc()' function. I'm not using the built-in UART, because I need to invert the signal. The CCS docs seem to imply that each bit is only sampled once, but that may be a simplification of the real process. Alex -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

Jason Neudorf wrote: >Neat. I'll be sure to avoid using my Palm IIIxe at 150C. :) > >My purpose was debugging and troubleshooting a hard real-time device the >company is developing for a customer. Just an extra trace to the ICSP >connector, and I can hook in an LED and see which fail-safe has kept the >device from smoking. I was already updating a DAC at 19531, so bit-banging >the IR at that rate was no problem; getting it exact would have caused no >end of troubles. > >How hard would it be to bit-bang the receive, with absolutely minimal >hardware? > You could use the PWM output for this. Stef Mientki > >Jason Neudorf > >{Original Message removed}

I appologize not reading all messages if I repeat someone notice. I would like to stress that Vishay's receivers don't like any discontinuity of a modulation.... The only way how to obtain a reliable communication is to modulate a transmitter by a way the time of modulation pulses multiply by a constant gives you a bps.... Using PWM is a little confusing by my meaning but useable. Igor {Original Message removed}

> >How hard would it be to bit-bang the receive, with absolutely minimal > >hardware? > > > You could use the PWM output for this. > Stef Mientki Um, you're thinking sending TV-remote style IR. It has a ~38kHz (or so) carrier frequency. I was asking about receiving the SIR IrDA stuff that you might get from a PalmOS device or laptop. Anyone have any ideas? Can I hook up a 1 or 2 component detector, with filtering in software, or do I need to use some expensive transducer? Jason Neudorf -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

Jason Neudorf wrote: >>>How hard would it be to bit-bang the receive, with absolutely minimal >>>hardware? >>> >>> >>> >>You could use the PWM output for this. >>Stef Mientki >> >> > >Um, you're thinking sending TV-remote style IR. It has a ~38kHz (or so) >carrier frequency. I was asking about receiving the SIR IrDA stuff that >you might get from a PalmOS device or laptop. > >Anyone have any ideas? Can I hook up a 1 or 2 component detector, with >filtering in software, or do I need to use some expensive transducer? > TSOP1138 around $2 expensive ? Did you take a look into the functional circuit of such a device ? If you've seen that, I think the $2 isn't even worth of even trying to find a solution of your own. Just a $0.02 opinion ;-) Stef Mientki > >Jason Neudorf > >-- >http://www.piclist.com hint: PICList Posts must start with ONE topic: >[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads > > > -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

{Quote hidden}> Jason Neudorf wrote: > > >>>How hard would it be to bit-bang the receive, with absolutely minimal > >>>hardware? > >>> > >>> > >>> > >>You could use the PWM output for this. > >>Stef Mientki > >> > >> > > > >Um, you're thinking sending TV-remote style IR. It has a ~38kHz (or so) > >carrier frequency. I was asking about receiving the SIR IrDA stuff that > >you might get from a PalmOS device or laptop. > > > >Anyone have any ideas? Can I hook up a 1 or 2 component detector, with > >filtering in software, or do I need to use some expensive transducer? > > > TSOP1138 around $2 expensive ? > Did you take a look into the functional circuit of such a device ? > If you've seen that, I think the $2 isn't even worth of even trying to > find a solution of your own. > Just a $0.02 opinion ;-) > > Stef Mientki Not useful for SIR IrDa (I sense blank looks). The infrared sent by a laptop or PDA is not modulated by 38kHz; rather, it consists of single, high intensity pulses, LSB first, one per bit. A "0" bit consists of a pulse, and a "1" bit has no pulses. It's very much like RS-232, but with short high-intensity pulses instead of continous high values. A quick look at digikey mentions TI's TIR1000PSR, at $CDN 4.11 quantity 1. This thing requires an external clock, but STILL doesn't mention what you receive with. It looks like it's functionality could be emulated in software. The question is what sort of IR detector can be used, and what sort of noise rejection to put into software/hardware. (the TSOP1138 does look useful for TV-remote type stuff). Jason -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics