Yahoo Groups archive
Thread
2004-09-28 by acetoel
Hello... I have seen boards with 10MHz, 14.xxxMhz and 20Mhz. All with the PLL set so the chip runs at 60Mhz more or less... But really, is there any difference if I use a 20Mhz crystal oscilator than a 10Mhz? IF there is no difference between crytal speed.Why so many people use the 14.xxxMhz crystal osc? Thanks Ezequiel
2004-09-28 by Leon Heller
----- Original Message -----
From: "acetoel" <acetoel@...> To: <lpc2000@yahoogroups.com> Sent: Tuesday, September 28, 2004 12:03 PM Subject: [lpc2000] LPC performance and Clock frecuency > Hello... > I have seen boards with 10MHz, 14.xxxMhz and 20Mhz. All with the PLL > set so the chip runs at 60Mhz more or less... > But really, is there any difference if I use a 20Mhz crystal > oscilator than a 10Mhz? > IF there is no difference between crytal speed.Why so many people > use the 14.xxxMhz crystal osc? It gives accurate baud rates for the UARTs. Leon
2004-09-28 by acetoel
Thanks I noticed that today when looking at the Baud Rate error. With 20Mhz is near 0.12% and with 14.xx is near 0.02%. It's quite a difference and 1 Mhz in the processor clock will not make a significant difference... Thanks Ezequiel --- In lpc2000@yahoogroups.com, "Leon Heller" <leon_heller@h...> wrote: > ----- Original Message ----- > From: "acetoel" <acetoel@y...> > To: <lpc2000@yahoogroups.com> > Sent: Tuesday, September 28, 2004 12:03 PM > Subject: [lpc2000] LPC performance and Clock frecuency > > > > Hello... > > I have seen boards with 10MHz, 14.xxxMhz and 20Mhz. All with the PLL > > set so the chip runs at 60Mhz more or less... > > But really, is there any difference if I use a 20Mhz crystal > > oscilator than a 10Mhz? > > IF there is no difference between crytal speed.Why so many people
> > use the 14.xxxMhz crystal osc? > > It gives accurate baud rates for the UARTs. > > Leon
2004-09-28 by Bill Knight
If the frequencies are exact, a 20MHz crystal should have an error of about 0.16% for common BAUD rates. A 14.7456MHz crystal should have an error of 0.00%. Regards -Bill Knight the ARM Patch On Tue, 28 Sep 2004 13:32:09 -0000, acetoel wrote: Thanks I noticed that today when looking at the Baud Rate error. With 20Mhz is near 0.12% and with 14.xx is near 0.02%. It's quite a difference and 1 Mhz in the processor clock will not make a significant difference... Thanks Ezequiel
2004-09-28 by Micron Engineering
This problem is common to every microcontroller, it is not new; if baud rate error is important you have to choose the correct crystaland this is not difficult since there are a lot of frequencies available. Of course the microcontroller clock will be lower then possible maximum. acetoel wrote: >Thanks >I noticed that today when looking at the Baud Rate error. With 20Mhz >is near 0.12% and with 14.xx is near 0.02%. It's quite a difference >and 1 Mhz in the processor clock will not make a significant >difference... >Thanks >Ezequiel > >--- In lpc2000@yahoogroups.com, "Leon Heller" <leon_heller@h...> >wrote: > > >>----- Original Message ----- >>From: "acetoel" <acetoel@y...> >>To: <lpc2000@yahoogroups.com> >>Sent: Tuesday, September 28, 2004 12:03 PM >>Subject: [lpc2000] LPC performance and Clock frecuency >> >> >> >> >>>Hello... >>>I have seen boards with 10MHz, 14.xxxMhz and 20Mhz. All with the >>> >>> >PLL > > >>>set so the chip runs at 60Mhz more or less... >>>But really, is there any difference if I use a 20Mhz crystal >>>oscilator than a 10Mhz? >>>IF there is no difference between crytal speed.Why so many >>> >>> >people > > >>>use the 14.xxxMhz crystal osc? >>> >>> >>It gives accurate baud rates for the UARTs. >> >>Leon >> >> > > > > > >Yahoo! Groups Links > > > > > > > > [Non-text portions of this message have been removed]
2004-09-28 by Jens Hildebrandt
Right, but when you have to use the same clock for your timer you have to deal with odd counter intervalls (e.g. 67.8ns with 14,7456MHz). So you have to decide what you want - error free UART or clean counter intervalls. Micron Engineering wrote: > This problem is common to every microcontroller, it is not new; if baud > rate error is important you have to choose the correct crystaland this > is not difficult since there are a lot of frequencies available. Of > course the microcontroller clock will be lower then possible maximum. > > acetoel wrote: > > >>Thanks >>I noticed that today when looking at the Baud Rate error. With 20Mhz >>is near 0.12% and with 14.xx is near 0.02%. It's quite a difference >>and 1 Mhz in the processor clock will not make a significant >>difference... >>Thanks >>Ezequiel >> >>--- In lpc2000@yahoogroups.com, "Leon Heller" <leon_heller@h...> >>wrote: >> >> >> >>>----- Original Message ----- >>>From: "acetoel" <acetoel@y...> >>>To: <lpc2000@yahoogroups.com> >>>Sent: Tuesday, September 28, 2004 12:03 PM >>>Subject: [lpc2000] LPC performance and Clock frecuency >>> >>> >>> >>> >>> >>>>Hello... >>>>I have seen boards with 10MHz, 14.xxxMhz and 20Mhz. All with the >>>> >>>> >> >>PLL >> >> >> >>>>set so the chip runs at 60Mhz more or less... >>>>But really, is there any difference if I use a 20Mhz crystal >>>>oscilator than a 10Mhz? >>>>IF there is no difference between crytal speed.Why so many >>>> >>>> >> >>people >> >> >> >>>>use the 14.xxxMhz crystal osc? >>>> >>>> >>> >>>It gives accurate baud rates for the UARTs. >>> >>>Leon >>> >>> >> >> >> >> >> >>Yahoo! Groups Links >> >> >> >> >> >> >> >> > > > > [Non-text portions of this message have been removed] > > > > > > Yahoo! Groups Links > > > > > > -- -------------------------------------------------- Dipl.-Ing. Jens Hildebrandt Universität Rostock, FB Elektrotechnik u. Informationstechnik Institut für Angewandte Mikroelektronik und Datentechnik R.-Wagner-Str. 31 18119 Rostock-Warnemünde Tel.: +49 381 4983537
2004-09-28 by Jens Hildebrandt
Right, but when you have to use the same clock for your timer you have to deal with odd counter intervalls (e.g. 67.8ns with 14,7456MHz). So you have to decide what you want - error free UART or clean counter intervalls. Micron Engineering wrote: > This problem is common to every microcontroller, it is not new; if baud > rate error is important you have to choose the correct crystaland this > is not difficult since there are a lot of frequencies available. Of > course the microcontroller clock will be lower then possible maximum. > > acetoel wrote: > > >>Thanks >>I noticed that today when looking at the Baud Rate error. With 20Mhz >>is near 0.12% and with 14.xx is near 0.02%. It's quite a difference >>and 1 Mhz in the processor clock will not make a significant >>difference... >>Thanks >>Ezequiel >> >>--- In lpc2000@yahoogroups.com, "Leon Heller" <leon_heller@h...> >>wrote: >> >> >> >>>----- Original Message ----- >>>From: "acetoel" <acetoel@y...> >>>To: <lpc2000@yahoogroups.com> >>>Sent: Tuesday, September 28, 2004 12:03 PM >>>Subject: [lpc2000] LPC performance and Clock frecuency >>> >>> >>> >>> >>> >>>>Hello... >>>>I have seen boards with 10MHz, 14.xxxMhz and 20Mhz. All with the >>>> >>>> >> >>PLL >> >> >> >>>>set so the chip runs at 60Mhz more or less... >>>>But really, is there any difference if I use a 20Mhz crystal >>>>oscilator than a 10Mhz? >>>>IF there is no difference between crytal speed.Why so many >>>> >>>> >> >>people >> >> >> >>>>use the 14.xxxMhz crystal osc? >>>> >>>> >>> >>>It gives accurate baud rates for the UARTs. >>> >>>Leon >>> >>> >> >> >> >> >> >>Yahoo! Groups Links >> >> >> >> >> >> >> >> > > > > [Non-text portions of this message have been removed] > > > > > > Yahoo! Groups Links > > > > > > -- -------------------------------------------------- Dipl.-Ing. Jens Hildebrandt Universität Rostock, FB Elektrotechnik u. Informationstechnik Institut für Angewandte Mikroelektronik und Datentechnik R.-Wagner-Str. 31 18119 Rostock-Warnemünde Tel.: +49 381 4983537
2004-09-28 by Micron Engineering
Yes this is right and common to every microcontrollers sold in the world, so it is not a problem to solve but simply to deal with it. You have no chance to change this functionality simply because it isn't possible. If you need some sort of timing related to seconds or fractions of seconds you may use a microcontroller with internal real time clock or use an external real time clock. You may find rtc with 10 milliseconds resolution. Massimo Manca, Micron Engineering Jens Hildebrandt wrote: >Right, but when you have to use the same clock for your timer you have to deal >with odd counter intervalls (e.g. 67.8ns with 14,7456MHz). So you have to decide >what you want - error free UART or clean counter intervalls. > >Micron Engineering wrote: > > > >>This problem is common to every microcontroller, it is not new; if baud >>rate error is important you have to choose the correct crystaland this >>is not difficult since there are a lot of frequencies available. Of >>course the microcontroller clock will be lower then possible maximum. >> >>acetoel wrote: >> >> >> >> >>>Thanks >>>I noticed that today when looking at the Baud Rate error. With 20Mhz >>>is near 0.12% and with 14.xx is near 0.02%. It's quite a difference >>>and 1 Mhz in the processor clock will not make a significant >>>difference... >>>Thanks >>>Ezequiel >>> >>>--- In lpc2000@yahoogroups.com, "Leon Heller" <leon_heller@h...> >>>wrote: >>> >>> >>> >>> >>> >>>>----- Original Message ----- >>>>From: "acetoel" <acetoel@y...> >>>>To: <lpc2000@yahoogroups.com> >>>>Sent: Tuesday, September 28, 2004 12:03 PM >>>>Subject: [lpc2000] LPC performance and Clock frecuency >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>>>Hello... >>>>>I have seen boards with 10MHz, 14.xxxMhz and 20Mhz. All with the >>>>> >>>>> >>>>> >>>>> >>>PLL >>> >>> >>> >>> >>> >>>>>set so the chip runs at 60Mhz more or less... >>>>>But really, is there any difference if I use a 20Mhz crystal >>>>>oscilator than a 10Mhz? >>>>>IF there is no difference between crytal speed.Why so many >>>>> >>>>> >>>>> >>>>> >>>people >>> >>> >>> >>> >>> >>>>>use the 14.xxxMhz crystal osc? >>>>> >>>>> >>>>> >>>>> >>>>It gives accurate baud rates for the UARTs. >>>> >>>>Leon >>>> >>>> >>>> >>>> >>> >>> >>> >>>Yahoo! Groups Links >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >> >>[Non-text portions of this message have been removed] >> >> >> >> >> >>Yahoo! Groups Links >> >> >> >> >> >> >> >> > > > > [Non-text portions of this message have been removed]
2004-09-29 by Jean-Rene David
* acetoel <acetoel@...>: > But really, is there any difference if I use a 20Mhz crystal > oscilator than a 10Mhz? Aside from the baud rate, there is another difference between selecting a 10MHz and 20MHz crystal; a reason very specific to the LPC2000 parts that have an older version of the bootloader software. Versions before 1.63 (I think, but check Philips' site for the exact version) of the bootloader software fail about 10% of the time when programming the internal Flash when cclk is higher than about 12-15MHz. To work around this problem, simply disable the PLL and work straight off the crystal frequency while programming the Flash. That is, if your crystal is below 12MHz. Otherwise, there is no way to slow down the system enough to program without error for that 1 part in 10. Of course, you could simply update the bootloader with a fixed version (provided by Philips on their site and uploadable in the part using the Flash utility), but sometimes this is not practicable.
2004-09-30 by Lasse Madsen
Hi All I'm planning to try out the LPC2114 and I have a question about the bootloader and the state of the processors pins during this... I have a multiplexer on UART 0 that can switch between a MAX232 for boot loading or another serial device. Channel A = RX Channel B = TX A+B Zero condition = serial device A+B High condition = MAX232 for boot loading A and B select lines are coupled directly to A = Trace sync (P1.20) B = Trace PKT.3 (P1.19) My philosophy is that when one holds the reset line low and drive P.14 low (boot enable) the processor will enter the bootloader with all its pins set for tristate causing them to float except the Trace packet, Trace sync, Pipe Status, etc. pins which has an internal pull up ... The internal pull up should then be able to drive the multiplexers A+B select lines high selecting the MAX232 chip for boot loading. According to the datasheet pins are tristated when reset is low but what are their state when entering the bootloader? I would assume they where all inputs? And that the internal pull ups would be enabled automatically? Any thoughts? Best regards Lasse
2004-09-30 by Karl Olsen
--- In lpc2000@yahoogroups.com, "Lasse Madsen" <lasse.madsen@e...> wrote: > Hi All > > I'm planning to try out the LPC2114 and I have a question about the > bootloader and the state of the processors pins during this... > > I have a multiplexer on UART 0 that can switch between a MAX232 for boot > loading or another serial device. > > Channel A = RX > Channel B = TX > > A+B Zero condition = serial device > A+B High condition = MAX232 for boot loading > > A and B select lines are coupled directly to > > A = Trace sync (P1.20) > B = Trace PKT.3 (P1.19) > > My philosophy is that when one holds the reset line low and drive P.14 > low (boot enable) the processor will enter the bootloader with all its > pins set for tristate causing them to float except the Trace packet, > Trace sync, Pipe Status, etc. pins which has an internal pull up ... > > The internal pull up should then be able to drive the multiplexers A+B > select lines high selecting the MAX232 chip for boot loading. > > According to the datasheet pins are tristated when reset is low but what > are their state when entering the bootloader? I would assume they where > all inputs? And that the internal pull ups would be enabled > automatically? > > Any thoughts? I think your idea should work. Karl Olsen