PIC16F690 PWM关闭时PWM峰值？

颠簸，只是人们看到这个消息，因为花了这么长时间才得到批准。



      以下为原文

    Bump, just so people see this message as it took so long to get approved.
 

嗨，还没有通过所有的代码发布，但不是清除CCP1CON，也许尝试使用外围设备中的关机控制机制：ECPAAS＝0B100000；/*设置位7在ECPAAS寄存器中。*与此相关的选项很多，请参阅数据表。问候，Mysil。



      以下为原文

    Hi,
Haven't worked thru all the code posted,
but instead of clearing the CCP1CON,
maybe try to use the shutdown control mechanism in the peripheral:
ECCPAS |=  0b10000000;   /* Set bit 7 in ECCPAS register. */
There is a lot of options in connection with this, see the datasheet.
 
Regards,
   Mysil

当OP使用汇编程序时



      以下为原文

   
As the OP is using assembler, that would be
BANKSEL ECCPAS
BSF     ECCPAS,ECCPASE

嘿，首先谢谢大家的推送和反馈。我试图用ECCPAs禁用PWM，它也工作，但是在PWM恢复之前不久，尖峰仍然在信号上。我在ISR中改变了Dababeland Enabelin到这个：所以我禁用了ECPAs，然后等待暂停，改变PWM输出的极性并启用PWM。再次CCPA。我仍然不知道什么东西会产生PWM信号的尖峰，因为它应该被关闭。还是我错过了什么？



      以下为原文

    Hey, first of all thanks for the push and the Feedback. I tried to disable PWM with the ECCPAS and it works too, but the Spike still are on the signal shortly before PWM gets turned back on. 
 
I changed the disabeling and enabeling in the ISR to this:

 banksel CCP1CON
 movlw b'10000000' ; deaktivate Halfbridge, ECCPAS,7 to high
 movwf ECCPAS
 ;movlw .198
 ;movwf TMR2
 
 movlw .50
 movwf counter
 decfsz counter
 goto $-1
 
 movlw .194 ; starte TMR2 mit 0
 movwf TMR2
 incf counter1
 movlw b'10001111' ; activate Halfbridge, 2LSB from 100 = 00, P1A active high, P1B active low
 btfsc counter1,0 ; if counter uneven, P1A P1B active high
 movlw b'10001100' ; activate Halfbridge, 2LSB from 100 = 00, P1A active high, P1B active high
 movwf CCP1CON
 movlw b'00000000'
 movwf ECCPAS
 movlw .62 ; start TMR0 with decimal 65
 movwf TMR0 ; with 4 Prescaler and 20MHZ interrupt after 156,8µs
  
 bcf INTCON, T0IF ; clear T0IF
 goto exit_ISR
 
So i disable with ECCPAS, then wait the pause, change the polarity of the PWM outputs and enable PWM with ECCPAS again. I'm still not sure what could generate the spikes on the PWM Signal since it should be tourned off. Or do i miss something? 

一些外围设备控制TIS和端口引脚，并覆盖它们所设置的任何外围设备。但是当外围设备关闭时，TRIS和端口引脚恢复到外围设备控制之前的状态。我在PIC24部分注意到了这一点。也许这部分是这样做的。在启用外围设备之前，将TIS和端口设置为一些定义的状态。



      以下为原文

    some peripherals take control of the tris and port pins and override whatever they have been set to make the peripheral work.  but when the peripheral is turned off, the tris and port pins revert to what they were before the peripheral took control.  I noticed this on PIC24 parts.  perhaps this part does it as well.
 
set the tris and ports to some defined state before enabling the peripheral.

如果我能理解的话，肯定不会100%。这个想法是在停用PWM之前再次设置端口和TIS引脚。还是在再次启用它之前？一定会尝试的，谢谢！因此，我尝试设置TISISC、PORTC和安塞尔/ANSELH在停用PWM之前，在配置CCP1CON之后，在CCP1CON配置之后以及在ECPCAS中写入之前和之后，对PWM进行去激活，但结果基本相同。只有在ECPAS之后，它根本不起作用，而且不确定我把它放在哪里，我得到了2个尖峰而不是一个。但是，这与TMR设置相关，因为在TrPUT中改变TMRP0的值或TMR之前的TMR值得到类似的结果。



      以下为原文

   
Not 100% sure if i understood it corretly. The idea is to set the Port and Tris pins again before deactivating the PWM? Or before enabeling it again? Will try that for sure, thanks! 
 
So i tried to set the TRISC, PORTC and ANSEL / ANSELH before deactivating PWM, after deactivating PWM right before configuring CCP1CON again, right after CCP1CON configuration and before and after writing in the ECCPAS but the result where mostly the same. Only after the ECCPAS it didn't work at all and once, not sure where i placed it, i got 2 spikes instead of one. But that is correlated to the TMR settings since changing the value in TMR0 for the Interrput or the Value for TMR before stoping PWM gets a similar result. 

不，刚开始的时候，当你的外围设备被禁用时，你希望他们恢复状态。我怀疑你已经正确地设置了它们。



      以下为原文

   
No, just once at the start, to the state you want them to revert to when the peripheral is disabled.
I suspect you were already setting them correctly.
 

我在主程序开始时这样做，万一这是什么？



      以下为原文

    ;*******************************************************************************
; INITIALISING PWM
;*******************************************************************************

 banksel TRISC
 clrf TRISC
 banksel PORTC
 clrf PORTC
 banksel ANSEL
 clrf ANSEL
 clrf ANSELH
 banksel T2CON
 movlw B'00000101' ; no Postscaler, Timer 2 on, Prescaler 4
 movwf T2CON
 banksel PR2
 movlw .197 ; .197 in Pr 2 register for 160µs Period.
 movwf PR2
 
I do this at the start of the main Programm, in case this is what was ment?

是的。有没有其他代码运行，您没有告诉我们，访问PORTC？由于您使用的是缺少LAT寄存器的旧PIC，如果在同一端口上使用其他引脚的位集或位清除指令，则总是存在读取修改写入问题的机会。



      以下为原文

    Yes.
Is there any other code running that you have not shown us, that accesses PORTC ?
As you are using an old PIC that lacks LAT registers, there is always a chance of Read-Modify-Write issues if you use bit set or bit clear instructions on other pins in the same port.

第一个帖子中的代码是正在运行的整个代码。主要目标是只需要不需要任何其他东西就可以生成PWM。



      以下为原文

    The code in the first Post is the whole code that is running. The main goal was just to generate the PWM needed without anything else. 

嗨，已经尝试过在消息“1”中显示的代码，看来你正在使用定时器0的中断来预测定时器2和CCP1何时起作用。为这两个定时器选择了预分频器4，所以应该可以预测定时。记住，有几个指令周期。从定时器转换中断中断直到输入中断，然后在中断Prolog、EXEOG和中断源测试中使用每个指令，在到达动作之前使用1个或2个指令周期。然后，我认为禁用中断处理程序中的ECP，然后初始化。在消息3中，我的建议是基于从ISR内部删除CCP的新初始化。如Qub所描述的：应用停机状态时要取的值，在ECPCAS寄存器中的其他位中指定，这些应该预先初始化。EN要继续操作，清除相同的比特，不要干扰CCP模块的其他设置。当计时器2返回到0时，CCP应该继续下一个周期循环。另一种暂停CCP信号的方式，可能会冻结定时器2MysIL。



      以下为原文

    Hi,
Have tried to step thru the code shown in message #1
Seems you are using interrupts from Timer 0 to predict when Timer 2 and the CCP1 are acting.
There is prescaler 4 selected for both these timers, so it should be possible to predict the timing.
 
Keep in mind that there is a few instruction cycles interrupt latency from timer transition until interrupt is entered, and then each instruction during interrupt prolog, epilog and testing for interrupt source, use 1 or 2 instruction cycles, before getting to the action.
 
Then, I think disabling the ECCP inside interrupt handler, and then initializing it again, is a risky undertaking.
My suggestion in message #3  was based on removing the new initialization of CCP from inside the ISR.
As described by qub:
   banksel ECCPAS
    bsf     ECCPAS, ECCPASE 
The values to be taken when Shutdown state is applied, is specified in other bits in ECCPAS register,
these should be initialized beforehand..
 
Then to continue operation, clear the same bit, do not mess with other settings for the CCP module
    banksel ECCPAS
    bcf     ECCPAS, ECCPASE  
The CCP should then continue with the next period cycle when Timer 2 go back to 0.
 
Another way to pause the CCP signal, might to freeze the Timer 2
    banksel T2CON
    bcf     T2CON, TMR2ON
//  ...
//    And then to resume timer operation:
    bsf     T2CON, TMR2ON  
 
Mysil

嘿，很抱歉周末的回复，周末已经过去了。我试着用TMR2接口进行这项工作，但这并不适用于PWM发电机的工作方式。因此，从我理解它，PWM发生器试图完成它的当前PWM信号，以不产生未完成的波。因此，如果我在波浪结束之前到达代码的不可分割部分，这就足够了。我试图降低和增加TMR0启动时间，如果我中断之前，我仍然有相同的结果，而当我中断后，我得到了真正奇怪的混合。由于事实上，他想产生全波，这也是必要的设置TMR2不久之前完成，使启用是完成后，波完整性保持完整。关于禁用和启用，我不知道我是否理解你错了，或者如果不清楚我想做什么。我在附件中添加了一个PICUTUT，使它更直观地显示我想要生成什么样的PWM。因此，理论上，我需要80秒（160秒的160秒）重复最后一部分。在每个通道的78个部分之后，我需要增加一个5-10μs的中断，这两个通道都是关闭的。为了实现PWM，我希望将PWM发生器配置为80秒和78秒。现在，我开始PWM，并运行第一个周期正常，然后我中断后78秒。我计数我的计数器，以产生必要的休息，在这两个PWM关闭。然后，我相应地改变PWM的高电平和低电平设置，将PWM重新打开，使信号在中断之前或第二路的另一条路上再次接通。这就是为什么我重新初始化IpUPT中的CCP1CON的原因。他们之所以这么做，主要是因为，至少在目前，当我将PWM发生器设置为160个周期时，我不能可靠地触发78秒后的中断。希望这能使这个概念更清楚吗？再次感谢大家的反馈和帮助，我真的很感激。



      以下为原文

    Hey,
 
sorry for the late reply, was gone over the weekend. 
 
 
Regarding the interrupts. I tried to do it with the TMR2 interupts, but that didn't work witht he way the PWM Generator works. So from how i understand it, the PWM Generator tries to finish it's current pwm signal to not generate unfinished waves. Therefor it was apparently enough if i arrive at the dissable part of the code before the wave was finished. I treid to lower and increase the TMR0 start time and if i interrupted earlier i still had the same result, while when i interrupted later i got really strange mixes. 
 
Due to the fact that he wants to generate full waves it is also neccesary to set TMR2 shortly before finishing so that the enable is done cloes after and the wave integrity stays intact. 
 
 
Regarding the disable and enable, i'm not sure if i understood you wrong, or if it is not clear what i want to do. I added a Picuter in the attachment to make it more visual what kind of PWm i want to generate. 
 
So in theory i want 80µS on (160µs off 160µs on) repeat last part. after 78µs of the OFF part for each Channel, i need to add a break of 5-10µs were both channels are off. 
 
To Realize the pwm i want i configured the PWM generator to 80µS on and 78µs off. Now i start the PWM and run through the first cycle normaly, then i interrupt after the 78µs. i count my counter down to generate the needed break where both PWM is turned off. Then i change the high and low active settings for the PWM accordingly and turn the PWM back on so that the signal gets ON again if it was ON before the break or the other way arround for the 2nd channel. Thats why i reinitialize the CCP1CON in the interupt. 
 
They main reason i do it this way, is because, atleast for now, i can't reliably trigger a break after 78µS when i set the PWM generator to 160µS cycles. 
 
Hope this make the idea behind this more clear? 
 
Thanks everyone again for all the feedback and help, i really appreciate it. 
   Attached Image(s)