nyemi

Link:www.elektroworld.info/ftp/index.php

Bigboat

New hardware (again), but (still) no new software.
I don't make new pcb's befor we have the software too.

There is no progress from Ton or Elmar of the Dutch-forum.
I think it was to diffecult to make a new program, or the boys don't like it anymore.

Rallyfinnen

Nice to see things are happening here again!

Is there somebody who has extra pcb's for the current (old?) version?
I would like to make a few controllers for the car engines, but i will not make pcb's. Tried that with mixed results before, so if soembody wants to sell a few, send me a PM!

SparkyGSX

Jpanhalt is absolutely correct on the MOSFET body diodes. There are, however, MOSFETs with enhanced body diodes, and for applications like this, I always prefer MOSFET with proper avalanche rating. This means the MOSFET can break down (avalanche), and survive, provided you designed the circuit in such a way that the integral of the current through the MOSFET, while in avalanche mode, is sufficiently small.

EDIT:Iforgot; zener diodes are completely useless for limiting large voltage spikes, because they are simply to slow to react to a high voltage condition. Transorb diodes are much better in this application, and are specificially designed for such duties.

I truely don't have a clue as to what Nyemi is trying to say. It is just me, or are his posts completely devoid of anything sensible to all of you?


As I already told bigboat on the dutch forum, such things tend to take a while, because I always have many concurrent projects.

I have a nearly complete design in text, and I think I have everything covered. It's just waiting to be implemented now, which shouldn't be to hard; the design is the important and difficult part, the implementation should be relatively easy, if the design is good and complete.

With a little luck, I'll receive some PIC development tools tomorrow, otherwise I'll probably get them on monday.

Bigboat

Where is everybody ?
In the mean time I learn to play a instrument, a kazoo

gr8flyer55

I've been converting my cdi boards back to using the CH Synchro Spark timer since we no longer have software for our now obsolete timer boards..a pity since I built 12 of them and wasted all my time.

John

COM

Hello John,

Can't you use the existing code?


Charlie

gr8flyer55

I have been using it to test my engines Charlie...the thing is it jumps timing all over the place after about 2000 rpms... making the engines run hot..and I get preignition noise. When I went back to the Synchro Spark, all the problems go away. So it's either not tracking rpm correctly or the table values are wrong, which I can't change since I am not that good at the excell sheet with copy and paste. It's got to be a steady increase in timing advance up to mid rpm range..just seems odd that it jumps back and forth at a small rpm increase.

John

COM

Hi John,

I wasn't sure if it worked or not.

Bigboat

If I read the comment of John, not the software is a problem but he can't copy/past.

@John, do can try a straight curve to test.
If you know the normal ignitionpoint of the engine, I can make a straight curve for you.
I use the ignition on 3 engines, but I don't have thoose problems you have.
Maybe the magnets or sensors are not the one you need.
Don't use *supermagnets* and don't use bipolair sensors.
The *supermagnets* have a to big magneticfeeld and bipolair sensors need a south- and northmagnet.

I hope Ton can start to day to test his program.
If the program runs, it takes maybe one or two weeks to get the little vaults out.

Rallyfinnen

The old code worked for me, the only problem was lag and bad presicion on high rpm, but should be usable on low rpm plane engines.
I think you have a code/compilation problem?

PM me if you want to sell the old timer boards.

Evert1969

I've started to build this cdi for using it at my converted saito FA80 four stroke engine.
In the excel sheet for setting the curve i noticed that the point of pick up is set at 40 degrees. Why is that a fixed value?
What is meant by the point of pick up? Is it the point were the magnet passes the hall sensor? So at 40 degrees means when the piston is at tdc the magnet is 40 degrees before the hallsensor?

Bigboat

The magnet must be placed at TDC, the sensor at 40 degrees BTDC.
The old program is never finished and now we hope Ton will do the job.
This is not mean the old program will do the job, but it's not the program wat can do we want.

SparkyGSX

Hey everybody,

Some good news: I have received a new logic analyser today, and with it came a load of PIC stuff. Some test boards, a PICKIT 1 and 2 (programmer/debuggers), and a whole lot of different controllers. This means I can finally start writing code.

I also got a license for a professional compiler, but I think I'll stick to the freely available one from Microchip, so that others can change and recompile the code if they need to.

But, because of the huge number of posts in this topic, I'm having some trouble getting a few things clear: which processor and hardware design are (most of you) using? I see a lot of people referring to "old designs", but which is the correct "old design" they are referring to?

Ofcourse, the software can be ported to other controllers (even non-PIC, or a different PIC series controller) later, but for initial development and testing, it would be very unpractical if we're using different controllers.

Bigboat

This is the CDI we all use http://www.modelbouw.gompy.net/newcdi/index.htm
Only the HV-boards can be differend http://www.electronics.gompy.net/cditci/

SparkyGSX

Ok, that's nice, because I got two 16F628 PIC's in the package today.

I just found out the 16F62x doesn't support bootloaders, so it won't be possible to load the code without a programmer. It should, however, be possible to load parameters (like the spark advance table) trough the serial port in the future. The 16F628 only has 128 bytes of EEPROM, so the size of the table might be a bit limited.

For good and consistent results, I think hardware timing periferals should be used (input capture, output compare modules). Unfortunately, the 16F628 only seems to have 1 timer that can do either, but only one of them at a time. Also, even more unfortunate is the fact that the documentation for this rather ancient controller ROYALLY SUCKS (sorry, I just had to get that out).

I'll try to get it to work in this controller, because quite a few people have already build the hardware, but I do thing it would be a good idea to design a new board, and fix some other mistakes in the hardware as well.

gr8flyer55

The normal ignition setting should be 28 degrees of advance. I tried to set the low at 6 degrees for idle per TKG. He mentioned it in the engine conversion forum. When I do that with the Synchro Spark module, everything is perfect and I get easy starts...so it should be possible to set ours for the same advance / retard. I'm not sure what I did but as I said earlier, mine jumps all over when I tested it with the timing light, causing problems. Most if not all my engines were using the same exact settings..so it should be fine.
Thanks Rob...

John

COM

Hi Sparky,

You should look at a 18F1220 or a 18F1320. (2) 16 bit timers and lots of code space. Or if you want to stick with a 16F pic look at a 16F88. Its the 628s big brother.

nyemi

I want you to simulate the program on and write an opinion. Thank you nyemi.

Bigboat

Wat program do you mean ?

SparkyGSX

I'm not really worried about code space, the EEPROM size is more of alimitation, because that determines the maximum size of the sparkadvance table, if we want to be able to update the table through theUART without reprogramming the chip. Alternatively, a bootloader couldbe used, but the smaller and older PICs can't support bootloaderfunctionality.

If we're going to use a different board anyway (in the future), wecould also use a different controller (not even necessarily a PIC, itcould also be an AVR, or something completely different), and use theon-board peripherals.

I think I'll try to use the input capture function, but also provide a(less accurate) polling-based alternative, and use a timer interruptfor the output timing. I could correct for the interrupt latency when setting up the timer.

Does anyone know if there is a PIC that is pin compatible (for the pins that matter; mostly, the power supply) with the 16F628? If so, and if the alternative has the features we want, the existing boards could still be used, with a replacement controller.

Bigboat

This one I think is pincompattible.(PIC16F/LF1826/27)
http://ww1.microchip.com/downloads/e...Doc/41391B.pdf

COM

the 16F88 is pin compatable.

Bigboat

16F1826
Enhanced Mid-range Core with 49 Instruction, 16 Stack Levels
Flash Program Memory with self read/write capability
Internal 32MHz oscillator
Integrated Capacitive mTouch Sensing Module
Data Signal Modulator Module
MI2C, SPI, EUSART w/auto baud
2 ECCP & 2 CCP (Enhanced/Capture Compare PWM)
Comparators with selectable Voltage Reference
12 Channel 10b ADC with Voltage Reference
25mA Source/Sink current I/O
Four 8-bit Timers (TMR0/TMR2/TMR4/TMR6)
One 16-bit Timer (TMR1)
Extended Watchdog Timer (EWDT)
Enhanced Power-On/Off-Reset
Brown-Out Reset (BOR)
In Circuit Serial Programming (ICSP)
On Board In-Circuit Debug
Wide Operating Voltage (1.8V – 5.5V)
Low Power PIC16LF182x variants (1.8V – 3.6V)
Standby Current (PIC16LF182X): 30 nA @ 1.8V, typical

16F88 is hard to get and cost twice the price a 1826

COM

The 16F1826 and so on Some compilers don't support them yet. It would be a good choice though.