Register

If this is your first visit, please click the Sign Up now button to begin the process of creating your account so you can begin posting on our forums! The Sign Up process will only take up about a minute of two of your time.

Page 28 of 46 FirstFirst ... 18262728293038 ... LastLast
Results 676 to 700 of 1127

  1. #676

    Join Date
    Aug 2011
    Location
    Moscow, ID
    Posts
    144
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    I've run the CDI several times today with my software and it seems to be working great!

    All I can figure out is that there might be some problem with specifying the memory address of the table and the pointer scheme I use for referencing them.

    I've changed:
    static const uint16_t AdvanceTable1[256] @ 0x03F0 =
    To:
    static const uint16_t AdvanceTable1[256] =

    This seems to have fixed the problem as far as I can tell. It was always an intermittent problem for me though. Sometimes things worked, sometimes the LED would say on when the magnet was near the sensor. Only once did it actually keep sparking when the magnet was there. Since some curves seemed to work and some didn't I can only think that there was something going on with the memory and addressing. This is my first PIC project, so I'm not sure exactly how the memory paging works and it seems that the compiler is maybe not taking care of this properly for me.

    I also noticed that the LED turns on very easily. Just touching the underside of the board is enough to turn it on.

    In any case I'm working through the code and tuning it up. I'll try to have another version out soon. The spreadsheet hex generator will have to wait for later when I figure out the memory issue.
    http://www.electrofunnel.com/CDI-2012/index.htm

  2. #677
    Gompy's Avatar
    Join Date
    Jul 2010
    Location
    Alkmaar, NETHERLANDS
    Posts
    404
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Have you set pull-up / pull down resistants intern on into your code ?
    CU, Rob
    http://www.electronics.gompy.net

  3. #678
    Gompy's Avatar
    Join Date
    Jul 2010
    Location
    Alkmaar, NETHERLANDS
    Posts
    404
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Something else, I think some of you have a Android GSM.
    https://play.google.com/store/apps/d..._result#?t=W10.
    CU, Rob
    http://www.electronics.gompy.net

  4. #679

    Join Date
    Aug 2011
    Location
    Moscow, ID
    Posts
    144
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Nice link, that seems like a handy program.

    WPU = 0b110101; // Internal pull-ups on GP5+4+2+0

    Pull-ups are on for the three inputs (hall, table switch, kill switch). GP0 pull-up (LED) is also set to on, but since it's an output this is actually disabled.


    I've been running different curves and settings all day long now and haven't had any problems. Table and kill switches are working just like they should, everything seems to check out.

    I really can't figure out what the bug was or if there really ever was one. Some of my problems were likely due to the crappy socket I soldered into the timer board. The only other thing I can think of is that the pro version of the compiler is doing some optimization that fails to generate stable code. I'm running on the free version now that my trial expired.

    Whatever the case I'll put out a new version tonight and we'll see if people still have problems.


    -Jake
    http://www.electrofunnel.com/CDI-2012/index.htm

  5. #680

    Join Date
    Aug 2011
    Location
    Moscow, ID
    Posts
    144
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    I just put v0.99 beta 8 out on the website.

    I eliminated the hex file generation button and am letting the compiler specify the memory locations for the tables.

    I also increased the speed of the main loop to make it respond faster to the table and kill switch. They should now respond in about a quarter sec. instead of half a sec or more.

    I changed the routine for loading table values in order to start the timer at the end of the ISR. This makes it so that the minimum load value is 1 instead of 7. It may cause the RPM measurement to be a little inflated (~6 uS or so), but I think this is not significant.


    I'd also like to change the way we are reporting bugs. Feel free to keep posting them here, but first send me an email report with the files you are using attached.

    So if you find a bug please describe the conditions that cause it in your email. Also tell me which version of the compiler you are using and if you are using free or pro mode. Make sure to attach the spreadsheet with your curves and table values, the actual code you are compiling, and the generated HEX file.

    Having all of this will make it a lot easier on me to figure out what is going wrong. In my early testing I was having some curves work just fine while others were causing problems. Hopefully this is now fixed, but if I can't replicate the problem then I can't figure out what is going on or fix it.

    Everyone please use the free version of the compiler for now. The tiny bit of memory the pro version saves is not important, and certainly not worth any potential headaches. It is not uncommon for compiler optimizations to cause problems with certain programs, and until I can nail down for sure the exact compiler optimizations that will for sure work we don't need to be throwing this potential problem into the mix.


    The recent posts here should give people a clue that there are dangers involved in testing beta code. If my code burns out a transistor, LED, SCR, etc. or blows up your engine... I'm truly sorry. All I can tell you is that it works for me, with my setup, with my engine. I would never release code that destroyed my own setup, but that is the ONLY warranty I can make.

    I currently use the hardware setup that John sent me. I run the code on a Homelite 25cc 2-stroke engine. Today I reflashed my chip about 15 times with slightly different settings, they all seemed to work flawlessly.

    The only "problem" I seem to be having is that the LED isn't useful. I haven't found a setting yet that really lets the flashes show. You can play with Trigger_Pulse_Time in the code to find a setting that works. I tried several settings, but I don't use the LED and just ran out of patience trying to find the right setting today. At these speeds turning the LED on and off is essentially just running it off a PWM output. For me it's just getting brighter and dimmer. The best setting is probably less than the 20 uS setting that's currently in the code.


    -Jake
    http://www.electrofunnel.com/CDI-2012/index.htm

  6. #681

    Join Date
    Nov 2005
    Location
    Hamburg, PA
    Posts
    806
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Jake, I have the Pro version of the compiler installed. How do I switch it back to the free version if possible? Is there a command to reset it within the program?

    I made two curves using the new spreadsheet Beta 99B8 version and will be testing monday evening if I get the static tests working.
    Since I recompiled the last version and reprogrammed my PIC with it, I haven't had the problems as before.

    Figured out the transistor problems also by changing the load resistor in the tuning resistors on the HV board. Everything now seems to run much cooler. As a side benefit, current draw from the HV board has gone down, around 60 mah without the engine running, and around 300 mah engine at max rpm.. That is good. Tuning resistors are opposite the .47 cap on the HV board. most boards I built had a 2.4 k resistor in the first position. Using just that resistor works fine alone. I was tuning for more output on the .47 cap and a parallel resistance total should have ended up around 2.2k. I think I got greedy and went down to 2k and that's when I started having the heating problems and excess current draw, so I solved that now by just using one 2.2k resistor, works fine and everything is back to normal.

    John

  7. #682

    Join Date
    Nov 2005
    Location
    Hamburg, PA
    Posts
    806
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Static tests completed...Version 99b8 compiled without a problem. Compiler worked ok and produced the required hex file. Programmed the hex to the PIC and all went well. Put the PIC in the timer, fired it up and it worked perfectly. Switches work as planned. Engine testing of the new curves will be done tomorrow possibly before I go to work and I'll post the results.

    John

  8. #683
    Gompy's Avatar
    Join Date
    Jul 2010
    Location
    Alkmaar, NETHERLANDS
    Posts
    404
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Not everyone was having a flashlight for 8 or 10 years ago, so we make a flashlight into the PIC.
    People can also use the output for a rotationmeter, or......you name it.
    A few years ago I make a flashlight without using the PIC, no wires between CDI and flashlight (safety)
    I say, remove the LED out of the code.

    BTW, it's right if you say you see dim the LED.
    But flash on a disk with one line at TDC and you will see the line flash up somewhere between TDC and 40 degrees BTDC.
    The only problem will be the time the LED is on, at high rpm the "up" time will be shorter and the brigdness will be less but it didn't loose the function.

    [youtube]http://www.youtube.com/watch?v=PJCWWGwa9Cg[/youtube]

    Next, why a killswitch into the software ?
    A killswitch is on the engine, shortcircuit the ignition, in ever motorsport I know.
    We know all the HV-baoard can make sparks even if the hall isn't flyby the magnet.
    How is the killswitch working with low batteries or EMP into the PIC ?
    I say remove the killswitch and make a real one on the engine and short the ignitioncoil with a real switch.

    Software is not allways the holygrail to make solutions.
    CU, Rob
    http://www.electronics.gompy.net

  9. #684

    Join Date
    Jun 2009
    Location
    Ga, GA
    Posts
    396
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Rob,

    The killswitch in software is for remote operation.
    Regards
    Charlie M.

  10. #685
    Gompy's Avatar
    Join Date
    Jul 2010
    Location
    Alkmaar, NETHERLANDS
    Posts
    404
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    I know Charlie, but what happens when the switch is tripped above the audience if there is suddenly a radio interference.
    An emergency switch must be mechanically and should never be written in software.
    Software can not be sure enough to switch and operate the hardware for continue working.
    Hardware must disable by hardware with a real switch, a shut-off switch.
    CU, Rob
    http://www.electronics.gompy.net

  11. #686

    Join Date
    Nov 2005
    Location
    Hamburg, PA
    Posts
    806
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Charlie, have you experimented with the "on" time of the led?
    When I get home from work today ill try changing the delay time to see if the brightness can be adjusted enough to bring that function back. On the dyno my friend built to test engines, we were making use of that function. I have one of Rob's strobeoscope's and will have to use it seperately I guess if the built in one won't work bright enough. It was not bad using it just to check that the timing was actually changing with rpm as programmed.

    John

  12. #687

    Join Date
    Nov 2005
    Location
    Hamburg, PA
    Posts
    806
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Forgot to mention that the engine run test of the curves worked great this morning. Switches function also, very positive engine kill.
    I programmed two greatly different curves in the pic so I could tell if the curve switch worked, it does!

    John

  13. #688

    Join Date
    Feb 2010
    Location
    szarvas, HUNGARY
    Posts
    132
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Hi Gompy
    I say, remove the LED out of the code.
    No way.
    For my hardware(zvs cdi), it will not be good for the code.
    I use 7 pin scr control.
    The inverted output is required.
    Alternatively, control can be used HV generator.(on-off)


    I know Charlie, but what happens when the switch is tripped above the audience if there is suddenly a radio interference.
    An emergency switch must be mechanically and should never be written in software.
    Software can not be sure enough to switch and operate the hardware for continue working.
    Hardware must disable by hardware with a real switch, a shut-off switch.
    I agree with you.
    Regards nyemi.

  14. #689

    Join Date
    Aug 2011
    Location
    Moscow, ID
    Posts
    144
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Hopefully with some testing we can find the right on time for the LED so that it flashes well, and I'll make it the default. I just didn't have time to mess with it enough. My GF was getting mad about me running the engine in the doorway.

    I've also read that LEDs can handle a lot higher current than they are rated for if it is in short pulses, which is exactly what we have. Perhaps someone can come up with a circuit that dumps a small cap through the LED when it is triggered.


    The reason I put a kill switch in the software is that it just seemed to make sense. I figured why short out your circuit with a mechanical kill switch when you can just stop firing it? I also figure that it will be easy to tie the mechanical kill switch to the kill switch pin. That way it will increase the safety and reliability of the kill switch. The kill switch activates when it is grounded, so you can take the mechanical kill switch and also hook it to the kill switch pin. You can do that in a way that it grounds the ignition output AND turns it off in the software at the same time. Then you aren't really shorting the circuit and wasting power. You would have to hook this up to make sure you aren't sending the ignition spark to the kill switch pin, but if your switch has 4 pins on it I would think it would work.

    In any case, I've been mainly thinking of RC and UAV/drone applications. Having a kill switch that doesn't require a relay or mechanical means is easier to control and takes less parts. I'm not saying it works this way right now, because I haven't tested it, but with the way I debounced the switch you should be able to run a PWM signal to the kill switch. If you have a +50% duty cycle it should keep the kill switch from activating, with -50% duty cycle it should activate the kill switch. I don't know yet if this actually works, it may take some tuning of the code, but some people might find it handy to have a easy kill switch on their TX.

    I don't think the kill switch is vulnerable to EMI/noise/interference. It is only using the internal weak pull-up resistor in the PIC, but the main loop runs at around every 10ms and it takes 10 reads of the switch as activated to make it kill the ignition. So 10 * 10 = 100ms of grounded kill switch pin to activate it. That doesn't include the ISR overhead, so it is probably more like 200ms to activate the switch.

    So it would take very low frequency interference, or very continuous interference, to activate it. Usually interference also induces voltage in the circuit, but the kill switch responds to being grounded. So I think if anything the kill switch might be vulnerable to NOT activating if there is a lot of noise in the ground.

    In any case, it's easy to disable...
    Simply delete lines 200-213 in the source code. They should look something like this...
    Code:
    if(GPIObits.GP4 == 0){   // If kill switch is on (grounded)
             if(KillSwitch < 10){ // Less than debounce value
                 KillSwitch++;    // Increment value
                 if(KillSwitch == 10){   // Debounce value reached
                     INTCONbits.GIE = 0; // Disable interrupts
                     sGPIO.GP1 = 1;      // Make sure HV trigger is off (turn GP1 on)
                     GPIO = sGPIO.ALL;   // Update GPIO
                 }
             }
         }
     else if(KillSwitch > 0){ // Kill switch off (high)
             KillSwitch;       // Decrease counter
     if(KillSwitch == 0){INTCONbits.GIE = 1;} // If debounced back to 0, enable interrupts
     }
    Without these lines it will not check the kill switch pins or have any lines that disable the interrupts. The compiler may complain that some variables are not used, but you can ignore this.
    BTW I've noticed that this forum munges all sorts of things for no obvious reason, even when pasted inside the "code" BBcode tag. So never trust any code posted as text. I noticed it deleted a couple characters from the code above, but hopefully things are clear enough.


    To switch the compiler between free and pro modes go to...
    File -> Project Properties -> Conf: -> XC8 global options -> XC8 compiler
    Then select "Option categories: Optimizations" from the drop down list box.
    Then choose "Free" in the "Operation Mode" list box.

    My free trial expired, so I'm stuck with the free version for the moment. It really doesn't seem to matter, pro mode saves only a tiny shred of memory. When I switch over to my laptop or upgrade one of my computers I'll probably get another 60 day free trial, but for the moment I can only compile in free mode so we should all stay on the same page.

    I really think it was that memory location issue with the tables. I've noticed that the way the PIC memory/banks are setup and paged it will roll addresses over. So in some cases by using the pointer arithmetic that I did it probably was screwing things up. I didn't change the memory scheme, but by letting the compiler place the table in memory it is probably picking a better spot that won't run into this problem.


    -Jake


    P.S. We really need to come up with a name for the project! I've just been calling it "CDI-2012" for now, but it would be nice to have a catchy and distinctive name for the project.
    http://www.electrofunnel.com/CDI-2012/index.htm

  15. #690
    Gompy's Avatar
    Join Date
    Jul 2010
    Location
    Alkmaar, NETHERLANDS
    Posts
    404
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    I think you don't call it.......Project-X. [&o]

    But wat do you think of Jake-in-the-box ?
    CU, Rob
    http://www.electronics.gompy.net

  16. #691

    Join Date
    Nov 2005
    Location
    Hamburg, PA
    Posts
    806
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Jake, things are working great now except for the led issue that you are working on. Switches are fine, ran long wires right by my receiver to a switch, no interference at all. Also the curve switch done the same way with no problems noted. I must mention that this a 72mhz setup which normally everything will affect, but still no problem.

    Rob, let's have a name contest ! I like the "Jake in the box" name. Lol

    John

  17. #692

    Join Date
    Aug 2011
    Location
    Moscow, ID
    Posts
    144
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Jake-in-the-box is out! I actually worked at the dang place once and heard that plenty.

    "Project-X" was a movie that came out fairly recently, and I think it pretty well used up.

    This is also the work of a lot of people. I've just written one software for the device, others have designed and produced the hardware, and figured all the things out to this point.

    I'm sure I can figure out a name for the software at some point, but it would be cool to have a name for the whole project. It will help people find it and know what everyone's talking about.

    I kind of liked "Vortex" since it seems to go well with my "Electrofunnel" domain name, but I looked and there already is a "Vortex" electronic ignition system.

    I'm sure with all the creative minds here we can come up with a cool name. Don't be afraid to throw something out there, we're just brainstorming so there's no wrong answers!

    Some random stabs here...

    Calculon CDI
    Twister "
    Turbo Tap "
    Electro Tap "
    Time-O-Matic (TOM)
    Blaster
    Snapper
    Mega-blast
    Electro-blast
    Combust-O-Matic (COM)
    Electro-blaster
    Master blaster
    Big Wheeler Cap Peeler
    Cap Snap
    Snap Cap

    So on and so forth I guess. An acronym type name would be good also.


    -Jake
    http://www.electrofunnel.com/CDI-2012/index.htm

  18. #693

    Join Date
    Feb 2010
    Location
    szarvas, HUNGARY
    Posts
    132
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Hi Jake
    I found a bug in your code.
    ****** GP 543210 *******/
    ****** Pin 234567 *******/
    TRISIO = 0b111100; // 0/1=Output/Input, Pin 234567, Label (GP543210)
    WPU = 0b110101; // Internal pull-ups on GP5+4+2+0

    WPU 0Bit Why is the value 1?
    The output is defined.(GP0-Pin 7)

    Regards nyemi.

  19. #694
    Gompy's Avatar
    Join Date
    Jul 2010
    Location
    Alkmaar, NETHERLANDS
    Posts
    404
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Projext X is a Facebook party.......and we don't like it.

    *** V1 (Table Ignition Timer Version 1).
    CU, Rob
    http://www.electronics.gompy.net

  20. #695
    Nav-Aids's Avatar
    Join Date
    Sep 2012
    Location
    Winnipeg, MB, CANADA
    Posts
    50
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55



    I agree that we need both a hardware and software kill feature. I for one want software kill.
    John:
    Are you talking about R8 on the schematic?




    Jake:
    I was looking at the schematic and thinking about what you said about the LED turning on when you touch the back of the board. In this case you are introducing stray voltage from your fingers and it is high enough to turn the transistor on. I suggest that instead of using a resistor to limit current between the PIC and the base of the transistor, switch the resistor to a small signal diode and place the diode as close the ‘base’ as possible. The diode will help to ensure that a signal ofatleast.4vto 1v only gets through to the base. (.4 – 1v) depending on the diode you use (Schottky or Standard).
    BAT42, 43, 85, 113 or 1N4148,1N914 my favorite.

    For the other circuit I use a zener or diode and a resistor for filtering and to make sure the sgnal is at the level I want.

    The ceramic decoupling capacitor between signal and groundis extra filteringor better yetuse an electrolytic cap between the diode and the 'Base' then it will filter and keep the LED turned on a little longer. If you use one of the circuits the ‘on’ time of the LED may need to be increased slightly. Try the first circuit first it should be sufficient. I’ll show the files as soon as I figure out how to upload pics and PDFs.




    Ray

    Everything is easy......Once you know how.

  21. #696

    Join Date
    Aug 2011
    Location
    Moscow, ID
    Posts
    144
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55


    ORIGINAL: nyemi
    WPUΒ* 0BitΒ* Why is the value 1?
    The output is defined.(GP0-Pin 7)
    I'm not sure why I set that on. I noticed that when I was looking at it recently, but I didn't want to mess with things so I left it alone.

    The pin is set as an output so the pull-up setting doesn't matter. I'm not sure if it just doesn't matter electrically or if the hardware actually turns it off. I'll probably change it in the next version, but I didn't want to risk introducing bugs or changing things once it was working right.


    On another note, I found that switches from old computers work well for the table and timer switches. Many old computers have momentary (reset) and toggle (power or turbo) switches right next to each other. I cut one out and wired a servo connector to it and it works pretty slick. I took a saw and just cut the whole thing right out of the computer face piece. Now I have both switches in a handy chunk of plastic that's easy to hold on to.


    -Jake
    http://www.electrofunnel.com/CDI-2012/index.htm

  22. #697

    Join Date
    Feb 2010
    Location
    szarvas, HUNGARY
    Posts
    132
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Hi Jake
    Thank you for your response.
    Yes I know it does not matter.

    I still have problems with the code.
    Therefore, I need some modification, your code.
    Interrupts are not treated, Hall inputs.
    I hope you succeed.
    I do not know C coding.
    I accept help.



    What do you think?
    Why did you chose? The input interrupt handling.

    Regards nyemi.

  23. #698
    Gompy's Avatar
    Join Date
    Jul 2010
    Location
    Alkmaar, NETHERLANDS
    Posts
    404
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    I hope it will be dry tomorrow John !

    BTW, If you want upload pics aso, you have to go to *post reply* left down under.
    At the bottom you will see attachments upload more files aso.
    Don't use fast reply, it doesn't work.
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	Eb86597.jpg 
Views:	22 
Size:	93.2 KB 
ID:	1813942  
    CU, Rob
    http://www.electronics.gompy.net

  24. #699

    Join Date
    Feb 2010
    Location
    szarvas, HUNGARY
    Posts
    132
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    I hope . You make a video.
    I would love to hear engine sounds.
    I congratulate you. Precise work.
    Regards nyemi.

  25. #700
    Nav-Aids's Avatar
    Join Date
    Sep 2012
    Location
    Winnipeg, MB, CANADA
    Posts
    50
    Gallery
    My Gallery
    Models
    My Models
    Ratings
    My Feedback

    RE: CDI gr8flyer55

    Ok here are my ideas for the LED, I hope they help.

    Ray

    www.unimatrixtech.com/CDI-2012/LED Fix.pdf


    Everything is easy......Once you know how.


Page 28 of 46 FirstFirst ... 18262728293038 ... LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
All times are GMT -6. The time now is 01:27 AM.

SEO by vBSEO 3.6.1 ©2011, Crawlability, Inc.