nyemi

Hi Gompy
Yes.
Look at this picture.
Original Moki ignition secondary circuit.
(This is not a voltage doubler.)
More information:www.rc-network.de/forum/showthread.php/248219-Der-gro%C3%9Fe-Benzinmotor-Z%C3%BCndkurventhread
Components are mandatory. (R-C-D)


I'm asking you.
What voltage to run in ZVS DC CDI?
What size ferrite cores you use?
The turns values ​​to be changed.
I will help you.

Gompy

@John, only for 2 euro 10 neodeen magnets 5mm a cross.
I only have to remove the gypsy diamonds, but thoose I sell again on ebay

@Neymi, I don't have build the CDI yet.
I'm missing the components name on the PCB.
I've to finf out and follow the schematic to PCB to find out wat component where located.
There are more components on PCB then in schematic !
I drive a Aprilia Classic 80cc, I think there are a lot of them in eastern europe, this one I buy in Austria.

BTW, first I will try the CDI from John on miy Homie and if everthing is OK maybe also on my bike.
Your software is into the PIC so it have to work good [8D]

bluejets

Don't know if you need to go to this extreme but stuck it in just in case.

Gompy

I hope you can see wat happend with the led if I move the hallsensor belong the magnet.

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

COM

Rob,

What is the point? I see the LED come on when the magnet is lined up with the HES and off when its not.

gr8flyer55

Rob, that's exactly what happens here.. led stays on after removal of the magnet..


John

Gompy

The LED goes on but not (allways) off when the hall leave the magnet.
If I turn the hall upside down, the same will happen, the LED goes off but not (allways) back on.
Thats why the SCR or PIC will stay trigger, the hall say you have to trigger I'm on / off.

After some test I mention if the hall sensor is not right in the middle of the magnet, or the magnet is not in the middle of the hall sensor, the problem is offen.
Maybe some one can test this also, put the hall a little out of the middle of the magnet and look if the LED will stay on / off.

Also the distance between the magnet and the hallsensor will do the same, if the distance is to big between them the LED will also stay on / off.

bluejets

I'm assuming the led is from the micro output GP0.

What type of hall switch is everyone using here. I use uA3240 at least if my memory is correct. (unipolar)

They have now been replaced with uA1120.

Never have any problem with them.

COM

Rob,

can we get some more details? What pin is the LED connected to? also doing that my hand with out the HES fixed leaves a lot of room for error.

Gompy

I only use the hallsensor (5 Volt) and a resistor / capacitor like this schematic, there is nothing else on the hallsensor.

Wrong picture, new upload

COM

John,

The GPIO.1 Led stays on all versions and Nyemi's code too. With the pic in my development board, GPIO.1 is on solid and with GPIO.2 toggling GPIO.1 is the only pin that goes on and off, but It still seems to work. I can try and video it to show you if need be.

Gompy

I try 3 differend types of hall, but all the same.
If the hall is not centred to the magnet or there is to much space between the hallsensor and the magnet I get the "problem".
I only switch the LED on off with the magnet, I don't do anything else.
It can't be the PIC or anything else, there is nothing else who can disturb the signal.

COM

OK so the HES is turning the LED of and off by itself. What type of HES are you using? do you have the datasheet? I also think you should secure teh HES to something and try again. Doing that free hand you are prone to errors.

bluejets

One thing I noticed in the data sheets comparing the Allegro A1120 and the Seimens TLE4905 (yours)
is the Siemens unit has very low turn on and turn off gauss requirements i.e. rather sensitive,

A1120 on is typ. 35 "on" and 25 "off"

whereas the seimens is 7.5 on and 9.5 off (min) and 19 on and 17 off (max)

Also wondering what the 100R resistor is for in the supply line..??

The TLE4905 will operate from 3.5 to 24 volt supply.

Gompy

I use HS501 and some others, the hall is fix with a RVS-band.
Now I have set the hall right in the middle I can't generate the errors anymore.
Screw the hall somewhere on the engine and hope it's OK, is not OK.

It doesn't matter if the 100R resistor is in the or not.
I follow only the datasheet of this hallsensor.
But if I remove it, the problem will be the same.
So i think it's only a matter of fix the hall on the right way.

gr8flyer55

Guys, Nyemi's software does not do this. I tried 12 different types of hall sensor with it and no led will stay on with his software. I do the exact same test with Jakes software and ALL of them produced the led error!
It is following Jakes software, even Jake himself tested it and got my results. He thought he fixed it by reflashing the pic chip but worked for only a very short time correctly, and then came back malfunctioning same as before reflashing the chip.
It is a borderline error. Random sometimes during an engine run, but always present during that powered test. Tried 5 separate timers, 2 chips. One chip with jakes program, the other with Nyemi's program. Problem followed Jakes chip every single time.

John

Gompy

OK John, but it's stranges I get the same without using any software, only a hallsensor and a LED.
I just take a brand new hallsensor, unipolar, but it will do like the rest.
But as I say befor, only if the hall is not in line with the magnet or the distance is to big between magnet and sensor.
Tomorrow I will try again, I've two other type hallsensors, now I go to bed it's allmost 2 AM.

COM

Thats not true. I will take a video and email it to you.

COM

OK Let me work on it.

gr8flyer55

Ok, so everyone is telling me that there is no truth to the notion about the software being flawed and that nothing burned up all of my transistors on my hv boards because the SCR just happened to mysteriously latch up?
I just ain't see in' the big picture I guess.
There's no sense arguing about it. I'll wait till my neighbor fixes it and just use it, although Nyemi's software still works perfectly, I'd just like to have a working changeable setup like the excel spreadsheet.

If this hall sensor equipped ignition is no better than an old transistor and points ignition where if the ignition was on an the points were closed, engine not running, that would overheat the transistor, then why are all of us wanting to use it if it has the exact same problem?


John

bluejets

John,
What Rob is trying to say is, he is testing JUST the hall switch on it's own and it is behaving unexpectedly. i.e. he says it is staying on after being removed from the magnet. ( no micro...just the hall and an LED)

Now this not usual for a hall switch.
(Unipolar) will turn on with magnet and off without it, so where is the flux coming from to give this turning on or holding on effect.

I suggested it may be because the hall switches you are using are too sensitive but maybe just barking up the wrong tree. It may be that the magnets you are using are far to strong or too large.

I use 3mm x 1mm magnet and Allegro hall effect uA3240 and have never had occurences like what Rob is explaining above.

There may well be some other problem in the software setup but more important to get the input that the micro is expecting to see correct first, which is what I tried to explain last week when I posted a circuit that would give just the one and only pulse.

As a further note, I have experienced in the past that if the gap between the hall switch and the magnet is too small, then weird things can happen such as turning on too early and staying on too long.

I usually set to about 2mm with the above hardware.

gr8flyer55

I always use the same number sensors on my test stand, Honeywell ss443R. The thing I don't understand is with Jakes software in the chip, I have the problem, same sensor same timer board and ignition board, switch just the chip to Nyemi's software, turn it back on and it works perfectly, no led staying on no matter how long I hold the magnet there or what position I change to, one spark, one blink of the led. Put jakes back in and nothing but problems, buzzing spark, dimly lit led and overheating of the hv board mje521 transistor.
Results are always exactly the same, Nyemi's wins!

John

bluejets

Yes John, understand what you are saying.

However, Rob has found something that needs to be addressed firstly in my opinion, as one clean single pulse is what the micro is, or should be, expecting to see.

After Rob's discovery has been rectified, it would then be time to look further. Just basicly eliminating problems, or potential problems in order.

gr8flyer55

By the way, we found the error. In process of rewriting the whole sensor input routine, foolproof this time. There are conditions with the error in the software which retriggered the output, a sort of resonant oscillation he called it, Ringing.
Will not show up on any simulator, only in the live circuit. Stopping it from happening in the first place is what he needs to rewrite the routine for. The length in u-seconds it is in limbo, trying to decide should i be on or off is what causes it. Changing the time it's on is the cure for it to settle itself back to a secure operation. Any hall sensor can have this problem, they are not as perfect as we would like to think!!!
What Jake did was cut the time in half, which threw it into the ringing problem. Also, he is running the software at 8 MHz which doubles the likely hood of the problem, think about it for a second and you will decide what he found to be 100% true.

John

jakestew

I should have some time to do some testing tomorrow.

I'm not all that sure that trying to address the ringing in software is going to be the best way to go. It will be good to get things working in the short term, but we really should figure out the hall issues that Gompy demonstrated and figure out any other problems.

If we're getting inductive ringing in the inputs and outputs I think that will continue to be a problem. One quick fix would be putting some simple filtering in the wires/connectors from the sensor and to the HV board.

When you mentioned ringing I googled up a good article on the issue. I probably should have known that with a noisy circuit with no filtering that uses unshielded wires going to both inputs and outputs that there could be problems like these. I'm just not good enough with electronic theory to figure these out without blundering into them first.

http://www.digikey.com/us/en/techzon...ectronics.html

If we're getting inputs like in the first scope picture, that's going to be a problem.