Post by thread:[PIC] need to shave off half an instruction

I have an existing small PIC project that produces an output on one pin based on the high bit of a register. In order to save an external inverter, I'd like to also output the complement on another pin. For the sake of argument let's say it is an oscillator that blinks an LED or drives a transducer. The target is a 16F627A and any pins are fair game, but glitches are not OK to the two pins, so no math or rotates directly on the ports. It is OK to do anything to any of the other pins. It is OK for there to be a little latency, and either port A or port B is OK. Or even both ports, though I don't see how that could get any faster. Here's the fastest thing I could come up with: movlw 0x55 btfss bigreg,7 movlw 0xAA movwf PORTA This is almost fast enough, if I could shave an instruction off one of the two conditions it would be enough for me. Any suggestions? Thanks, Bob -- http://www.fastmail.fm - Does exactly what it says on the tin

{Quote hidden}> I have an existing small PIC project that produces an output on one pin > based on the high bit of a register. In order to save an external > inverter, I'd like to also output the complement on another pin. For the > sake of argument let's say it is an oscillator that blinks an LED or > drives a transducer. > > The target is a 16F627A and any pins are fair game, but glitches are not > OK to the two pins, so no math or rotates directly on the ports. > > It is OK to do anything to any of the other pins. > > It is OK for there to be a little latency, and either port A or port B > is OK. Or even both ports, though I don't see how that could get any > faster. > > Here's the fastest thing I could come up with: > > movlw 0x55 > btfss bigreg,7 > movlw 0xAA > movwf PORTA > > This is almost fast enough, if I could shave an instruction off one of > the two conditions it would be enough for me. > > Any suggestions? > Let's assume you're using PORTA bits 0 and 1 (and you stated no other pins mattered). You could use swapf and do it in two instructions (if PORTA is initialized as shown below): ;Initialize PORTA with 0x21 (or 0x12). btfss bigreg,7 swapf PORTA -Scott

Stirrings from long ago. Idea starters. Or not. Writing to a control register may affect pin level differently depending on what is in the pin's data register. Without looking at your PIC I don't know if there is any way that this may be applied but it may be adding or removing a pullup, converting from input to output etc. Also, in the definitely naughty department:: - Behaviour of a pin that varies depending on external load level. - RMW attributes in conjunction with any above.

On Sun, Jan 15, 2012, at 12:38 AM, Scott wrote: > Let's assume you're using PORTA bits 0 and 1 (and you stated no other > pins mattered). > You could use swapf and do it in two instructions (if PORTA is > initialized as shown below): > > ;Initialize PORTA with 0x21 (or 0x12). > > btfss bigreg,7 > swapf PORTA PORTA isn't a full byte, but using PORTB instead, it still won't have the desired effect. What it does is leave the port alone in one condition and toggle it in the other, rather than having one pin always follow and the other be the inverse. Thanks for the suggestion, though, it's one I hadn't tried! Cheerful regards, Bob -- http://www.fastmail.fm - Access your email from home and the web

>> Let's assume you're using PORTA bits 0 and 1 (and you stated no other >> pins mattered). >> You could use swapf and do it in two instructions (if PORTA is >> initialized as shown below): >> >> ;Initialize PORTA with 0x21 (or 0x12). >> >> btfss bigreg,7 >> swapf PORTA > > PORTA isn't a full byte, but using PORTB instead, it still won't have > the desired effect. What it does is leave the port alone in one > condition and toggle it in the other, rather than having one pin always > follow and the other be the inverse. > Hi Bob, Now I'm a bit confused. What is the desired behavior, step by step? I'll throw out another suggestion: how about an xor of 0x03? movlw 0x03 btfss bigreg,7 xorwf PORTA -Scot

:: movlw 0x55 :: btfss bigreg,7 :: movlw 0xAA :: movwf PORTA Just in thinking mode. Why not: movlw 0x55 btfss bigreg,7 rrf PORTA,F movwf PORTA Colin -- cdb, spam_OUTcolinTakeThisOuTbtech-online.co.uk on 15/01/2012 Web presence: http://www.btech-online.co.uk Hosted by: http://www.justhost.com.au This email is to be considered private if addressed to a named individual or Personnel Department, and public if addressed to a blog, forum or news article.

On Sun, Jan 15, 2012, at 02:47 AM, Scott wrote: > Now I'm a bit confused. > What is the desired behavior, step by step? One pin follows the high bit of bigreg Another pin is the complement of the first pin. > > I'll throw out another suggestion: how about an xor of 0x03? > > movlw 0x03 > btfss bigreg,7 > xorwf PORTA That also toggles whenever the high bit of bigreg is set, rather than shadowing it. Thanks! Bob -- http://www.fastmail.fm - IMAP accessible web-mail

> Here's the fastest thing I could come up with: > > movlw 0x55 > btfss bigreg,7 > movlw 0xAA > movwf PORTA What's that bigreg, is its format out of your control? Some PICs have a hardware comparator, could that used? -- Wouter van Ooijen -- ------------------------------------------- Van Ooijen Technische Informatica: http://www.voti.nl consultancy, development, PICmicro products docent Hogeschool van Utrecht: http://www.voti.nl/hvu C++ on uC blog: http://www.voti.nl/erblog

Bob Blick wrote 2012-01-15 06:23: > I have an existing small PIC project that produces an output on one pin > based on the high bit of a register. In order to save an external > inverter, I'd like to also output the complement on another pin. Not on the 16F627A, but it seems as the Configurable Logic Cell (CLC) module on other devices would be able to do this in hardware with proper setup. Mirror the invers of one I/O pin on another I/O pin. I guess at least, the module is a bit complex... There was a thread on the use of the CLC module a few days ago, so I thought this was at least semi-OT... Jan-Erik. {Quote hidden}> For the > sake of argument let's say it is an oscillator that blinks an LED or > drives a transducer. > > The target is a 16F627A and any pins are fair game, but glitches are not > OK to the two pins, so no math or rotates directly on the ports. > > It is OK to do anything to any of the other pins. > > It is OK for there to be a little latency, and either port A or port B > is OK. Or even both ports, though I don't see how that could get any > faster. > > Here's the fastest thing I could come up with: > > movlw 0x55 > btfss bigreg,7 > movlw 0xAA > movwf PORTA > > This is almost fast enough, if I could shave an instruction off one of > the two conditions it would be enough for me. > > Any suggestions? > > Thanks, Bob

On Sat, 14 Jan 2012, Bob Blick wrote: {Quote hidden}> I have an existing small PIC project that produces an output on one pin > based on the high bit of a register. In order to save an external > inverter, I'd like to also output the complement on another pin. For the > sake of argument let's say it is an oscillator that blinks an LED or > drives a transducer. > > The target is a 16F627A and any pins are fair game, but glitches are not > OK to the two pins, so no math or rotates directly on the ports. > > It is OK to do anything to any of the other pins. > > It is OK for there to be a little latency, and either port A or port B > is OK. Or even both ports, though I don't see how that could get any > faster. > > Here's the fastest thing I could come up with: > > movlw 0x55 > btfss bigreg,7 > movlw 0xAA > movwf PORTA > > This is almost fast enough, if I could shave an instruction off one of > the two conditions it would be enough for me. > > Any suggestions? > > Thanks, Bob bcf STATUS,C rrf bigreg,w addlw 0x40 movwf PORTA if you can ensure that the carry flag is in a known state as a side effect of previous instructions then you could eliminate the "clear carry". NOTE: if carry is set use "addlw 0xC0" instead Regards Sergio Masc

I don't see the downside of Wouter's suggestion of using a voltage comparator. Output the bit on A0, which is connected to A1. Set comparator1 to compare the voltage on A1 to 2.5V from the voltage reference. Output comparator 1 results to A3. A3 will always the the inversion of A0. Why wouldn't that do what you want? Kerry Bob Blick wrote: {Quote hidden}> I have an existing small PIC project that produces an output on one pin > based on the high bit of a register. In order to save an external > inverter, I'd like to also output the complement on another pin. For the > sake of argument let's say it is an oscillator that blinks an LED or > drives a transducer. > > The target is a 16F627A and any pins are fair game, but glitches are not > OK to the two pins, so no math or rotates directly on the ports. > > It is OK to do anything to any of the other pins. > > It is OK for there to be a little latency, and either port A or port B > is OK. Or even both ports, though I don't see how that could get any > faster. > > Here's the fastest thing I could come up with: > > movlw 0x55 > btfss bigreg,7 > movlw 0xAA > movwf PORTA > > This is almost fast enough, if I could shave an instruction off one of > the two conditions it would be enough for me. > > Any suggestions? > > Thanks, Bob > >

Yes, I think that is the perfect solution. Sergio's code is good but the same number of instructions. Using Wouter's idea of the comparator I can use the same code I always used, but save an external inverter by using the internal comparator. Thanks to everyone! Best regards, Bob On Sun, Jan 15, 2012, at 12:03 PM, Kerry Wentworth wrote: {Quote hidden}> I don't see the downside of Wouter's suggestion of using a voltage > comparator. > Output the bit on A0, which is connected to A1. > Set comparator1 to compare the voltage on A1 to 2.5V from the voltage > reference. > Output comparator 1 results to A3. > A3 will always the the inversion of A0. > Why wouldn't that do what you want? > > Kerry > > > > Bob Blick wrote: > > I have an existing small PIC project that produces an output on one pin > > based on the high bit of a register. In order to save an external > > inverter, I'd like to also output the complement on another pin. For the > > sake of argument let's say it is an oscillator that blinks an LED or > > drives a transducer. > > > > The target is a 16F627A and any pins are fair game, but glitches are not > > OK to the two pins, so no math or rotates directly on the ports. > > > > It is OK to do anything to any of the other pins. > > > > It is OK for there to be a little latency, and either port A or port B > > is OK. Or even both ports, though I don't see how that could get any > > faster. > > > > Here's the fastest thing I could come up with: > > > > movlw 0x55 > > btfss bigreg,7 > > movlw 0xAA > > movwf PORTA > > > > This is almost fast enough, if I could shave an instruction off one of > > the two conditions it would be enough for me. > > > > Any suggestions? > > > > Thanks, Bob -- http://www.fastmail.fm - The professional email service

Also, you could double the processor speed, and get twice the instruction resolution. Jim -----Original Message----- From: .....piclist-bouncesKILLspam@spam@MIT.EDU [piclist-bouncesKILLspamMIT.EDU] On Behalf Of IVP Sent: Sunday, January 15, 2012 4:55 PM To: Microcontroller discussion list - Public. Subject: Re: [PIC] need to shave off half an instruction > Using Wouter's idea of the comparator Internal comparator briefly flicked through my head. Got me to re-read the datasheet anyway ;-) I've used the comparators only once, a long time ago, as a motor stall detect. Not a feature that gets mentioned much. Less so than the ADC Jo