|
I think the PIC is plenty fast enough
Thanks for your attention and replies.
I think the PIC is plenty fast enough. I think the SFR05's are responding quickly enough judging from their flashing LED's. My guesses are either SFR inconsistencies and/or weakness in my jury rigged PWM signal.
The servo and the speed control want PWM pulses at 50 per second. I have not been able to do that with my PIC. I get the pulse width correct, but I can not get the rep rate down to once each 20 millisec.
Ken
|
I think I've got it...
The PIC KIT2 programmer has, at the bottom, reference to EEPROM. It has addresses 00 - FF.
I think that if I use the PBP WRITE command at run time I can store data in the inchip EEPROM space. 00 to FF in size. I can then manually read this out with PIC KIT2 programmer.
Sound correct??
Ken
|
Yup, that will work.Sound correct??
Dave
Always wear safety glasses while programming.
|
A prototyping problem
I need to be able to connect my PICkit 2 to three SRF05 sonar sensors. The easiest way would be to solder J or JR (or whatever - I do not know one connector from another) headers to the board.
I looked carefully at the connectors that came with my radio control car. The pin separation does not match the hole separations in my PICkit 2 forty four pin demo board. Do any of you know the name of the connector header - if it exists - that would solder directly into that row of holes along the sides of my proto card? They are definitely not 1/10 inch centers. I have no means to measure their separation accurately.
Ken
|
Servo pwm passthrough using PCPWM
The 18F2431 makes it slightly easier to perform servo PWM because it's PCPWM module can go down to 19 htz while maintaining fast OSC rates, unlike other PIC chips HPWM. Here is an example of using the PCPWM to output 3 servo signals. This specific example senses the pulses (up to 5 channels) by using the CCP1 capture pin. To get 5 channels into one pin for PW capture, channels 1, 3 and 5 are input using small signal diodes. Then, since this drops the signal down to under 3 volts, I used two 2n3904 transistors. One to boost up the voltage, and the other one to invert the signal back to normal. (It was all I had lying around). Somehow, you will need to boost your signal after the diodes for the CCP1 capture to sense the incoming pulses.
Note: Because the PCPWM was used, this will not work on any chip outside of the 18F2431 family.
To see a similar project, but using SPWM, look here: http://www.picbasic.co.uk/forum/showthread.php?t=12657Code:;This only works on a 18F2431 family device with Power Control PWM (PCPWM) ;By Scale Robotics Inc. ; The input signal should look something like this:- ; ; |<------------------ 1 frame (~20mS) -------------->| ; |<1~2mS> ; ______ ______ ______ ______ ; _______| |______| |______| |________________| |____ ; gap ch1 ch2 ch3 ch4 ch5 sync gap ch1 etc define OSC 40 asm __CONFIG _CONFIG1H, _OSC_HSPLL_1H __CONFIG _CONFIG2H, _WDTEN_OFF_2H & _WDPS_512_2H __CONFIG _CONFIG4L, _LVP_OFF_4L endasm clear ADCON0 = %00000000 ADCON1 = %00000000 portb=0 trisb = %11000000 trisc = %00000110 trisa = %00000000 DTCON = %00000101 'dead time for complimentary ouputs PTCON0 = %00001101 '1:1 postscale, Fosc/4 1:64 prescale, Sincle Shot mode PTPERL = 255 ' PTPERH = 251 PWMCON0 =%01010000 'PWM[5:0] ouputs enabled PWMCON1 = 1 'updates enabled, overrides sync w/timebase PTCON1 = %10000000 'PWM timebase is on, counts up FLTCONFIG = %00000010 'disable fault A, cycle by cycle mode CCP1CON = %00000101 'Capture mode; every rising edge duty1 var word 'width of outgoing pulse1 duty2 var word 'duty values 625 to 1250 = 1 to 2 ms pulse (Center at 625) duty3 var word risetime1 var word 'Rise Time for incoming pulse1 risetime2 var word risetime3 var word falltime1 var word 'falltime for incoming pulse1 falltime2 var word falltime3 var word pulsewidth1 var word 'pulse width for incoming pulse1 pulsewidth2 var word pulsewidth3 var word pulsewidth4 var word pulsewidth5 var word pulseNumber var byte pulseNumber = 1 'tells which pulse we are reading CCP1CON.0 = 1 timerone var word timerone = 60315 ; INCLUDE "DT_INTS-18.bas" ; Base Interrupt System ASM INT_LIST macro ; IntSource, Label, Type, ResetFlag? INT_Handler TMR5_INT, _PulseOut, ASM, yes INT_Handler CCP1_INT, _PulseMeasure, ASM, yes INT_Handler TMR0_INT, _TimeOut, ASM, yes endm INT_CREATE ; Creates the interrupt processor ENDASM T1CON = %00110001 ;Timer1 used by CCP1 capture T5CON = $01 ;used as pulseout timer T0CON = %11000111 ; Prescaler = 8, TMR1ON @ INT_ENABLE TMR5_INT ; Enable Timer 1 Interrupts @ INT_ENABLE CCP1_INT ; Enable Capture Compare for pulse width measurement @ INT_ENABLE TMR0_INT ; deadtime (sync gap) indicator for lull between pulses Main: pause 10 'do something to these values if you want to filter, center, etc 'Below we are just passing values through duty1 = pulsewidth1 >>1 'send channel 1 out PCPWM on PortB.1 duty2 = pulsewidth3 >>1 'send channel 3 out PCPWM on PortB.3 duty3 = pulsewidth5 >>1 'send channel 5 out PCPWM on PortB.4 GOTO Main '---[TMR5_INT - interrupt handler]------------------------------------------ PulseOut: 'set up pulse width values for pulseout and reset single shot bit TMR5L = timerone.byte0 TMR5H = timerone.byte1 PDC0L = duty1.lowbyte PDC0H = duty1.highbyte PDC1L = duty2.lowbyte PDC1H = duty2.highbyte PDC2L = duty3.lowbyte PDC2H = duty3.highbyte PTCON1.7=1 'resets single shot PTEN bit @ INT_RETURN '---[CCP1_INT - interrupt handler]------------------------------------------ PulseMeasure: if CCP1CON.0 = 1 then ; Check CCP1M0 for rising edge watch select case pulseNumber case 1 TMR0L = 0 'reset timeout timer0 TMR0H = 0 risetime1.lowbyte = CCPR1L risetime1.highbyte = CCPR1H case 2 TMR0L = 0 'reset timeout timer0 TMR0H = 0 risetime2.lowbyte = CCPR1L risetime2.highbyte = CCPR1H case 3 TMR0L = 0 'reset timeout timer0 TMR0H = 0 risetime3.lowbyte = CCPR1L risetime3.highbyte = CCPR1H End select @ BCF CCP1CON, CCP1M0 ; Now capture the trailing edge Else 'check for falling edge time @ BSF CCP1CON, CCP1M0 ; Re-set for trailing edge capture select case pulsenumber case 1 falltime1.lowbyte = CCPR1L falltime1.highbyte = CCPR1H pulsewidth1 = falltime1 - risetime1 case 2 falltime2.lowbyte = CCPR1L falltime2.highbyte = CCPR1H pulsewidth2 = risetime2 - falltime1 pulsewidth3 = falltime2 - risetime2 case 3 falltime3.lowbyte = CCPR1L falltime3.highbyte = CCPR1H pulsewidth4 = risetime3 - falltime2 pulsewidth5 = falltime3 - risetime3 end select pulsenumber = pulsenumber + 1 'get ready for next channel endif @ INT_RETURN '---[TMR0_INT - interrupt handler]------------------------------------------ TimeOut: pulsenumber = 1 'if pause between pulse in exceeds about 7 ms, get ready 'to receive pulse 1 (senses dead time between pulse bursts) @ INT_RETURN
Last edited by ScaleRobotics; - 28th July 2010 at 20:08.
|
Thank you, Dave
Yes, the 18F2431 looks like a winner for my PWM construction. It is included in my PICkit2 programmer.
It comes in a 40 pin package. It would easily replace my 40 pin 16F887 on the solderless protoboard. To use it effectively I need to master ASM coding. I have not succeeded in cracking that nut.
Here is my latest stumbling block. I have an arrangement with some folks who are willing to contribute brand new HPI Racing Sprint cars to my project. I received my first new car last week. HPI has drastically changed the Electronic Speed Control. This is the package that translates incoming PWM signals (from the radio receiver or from my PIC) into outgoing DC current for the motor that drives the wheels. I am not doing well with the HPI WEB site getting technical information about this SC-15 WITH REVERSE SPORT CONTROL.
Ken
|
Hey Ken,Originally Posted by Kenjones1935
Congratulations on getting the donation! It is a great idea you have, and I am glad you have stuck with it. As for your newest stumbling block, what code are you currently using? You might be running into the same problem of not having a standard R/C output pulse going to the controller. And MIBAM will not work, because it is meant for lighting, and will confuse your speed controller even more.
All analog servos (not the newer digital ones) use a standard protocol. A pulse width from about 0.9mS to 2.1mS or so, pulsed every 20 mS. Since this controller does not say digital on it, I would bet that it is analog, and works the same as all the servos in my house. Since it has reverse, as well as forward, I would bet that somewhere in between .9mS and 1.1 mS is neutral. Less is probably reverse, and more is forward.
I don't think you need a manual for it. Just start feeding it some PWM that matches that protocol, and you will be good.
Assembly is a tough nut to crack. It's just one of those things, the more you use it, the more you learn. But one of the things I learned this week from the forum, was that I could have an instant ASM interrupt without using any ASM. No, I didn't know that when I wrote some similar code for the 12f683. So in my example above, I only used two lines of assembly code. a BSF and a BCF. (Bit set and bit clear). So you can have almost no nuts with this code, or you can change them to PBP code if they drive you ... um ..... nuts.
I think you already know this, but I just found the gif file, and I like it, so I'll post it here. - Walter
Last edited by ScaleRobotics; - 29th July 2010 at 17:04.
|
I was always led to believe it was typically 1ms to 2ms with a 20ms frame rate this gave 1.5ms pw for centre stick, at least that's what I've always based my SSC type projects on.
|
I think this ESC is different
I agree with both of you on the size, shape and frequency of standard PWM signals. I have been running my car (an older version of this same HPI Sprint) successfully for the last couple of months. With the exception of frequency my HPWM pulses agree with your specs. I found the correct HPWM commands by trial and error.
This new Transmitter, Receiver, ESC combo is different. Leaving out my PIC for the moment if I on the radio transmitter I give full forward (trigger full back) and suddenly go to full reverse (trigger full forward) the car goes from full forward to stop (neutral in the PWM sense) and stays there until I bring the trigger to neutral for at least two seconds. Then giving it full back will work fine.
Somewhere in the system is protection against slamming the DC motor with reversed polarities. That is good, but when I tried to emulate that system in my code it did not work.
Thank you sooo much for your support and interest. Oh, yes, the electronic speed control wheel driving system is not a classic servo. Steering is. Wheels used to be - in the day - but no longer.
Ken
|
Edit: oops, yes I meant to say 2.1 as the highpoint, not 1.1. Thanks Malc!
Hey Malc,
Yeah, me too! As a basic rule it works. But then I built my first passthrough device, I built it for that range and it got discombobulated when the signals from my transmitter exceeded the 2 mS. My transmitter has endpoint adjustments, so it allows you to go beyond the 1 to 2 mS. Some transmitters will only give 1.1 to 1.9ms.
Check out the HiTech web site regarding the signals:
You are probably right, I just was not sure if a forward and reverse speed control would be centered at 1.5 mS. My planes don't have reverse.
Last edited by ScaleRobotics; - 29th July 2010 at 17:04.
|
Usually an H bridge is used for forward and reverse. It is pretty much what you describe, only using two more fets. Microchip has an application note about one here: http://ww1.microchip.com/downloads/e...tes/00893a.pdf . It shows a nice diagram of the different states of forward and reverse on page 3.
In your post, I don't understand the " It could steer the correct MOSFET signal with a third 'direction' output." part though.... Don't you steer with a servo?
http://www.scalerobotics.com
|
Not the clearest I am
When I wrote, "It could steer the correct MOSFET signal with a third 'direction' output."
I was thinking of simplifying the speed control functionality by separating out the FORWARD and BACKWARD aspect. HPWM can easily create a PWM signal between one and two millisec. A second digital HIGH and LOW could dictate rotation direction. I've seen reference to something like this, but I did not think it through.
Ken
Members who have read this thread : 2You do not have permission to view the list of names.
Posting Permissions
Bookmarks