Pic reset and following lock up problem

[zerosvarka]

Hi Jerson,

I put a schottky diode across the coil primary terminals. I’ve attached the scope traces taken at the coil negative (switching point). It does fix the problem completely. It proves the need for suppression diode whenever inductive load is switched on and off. Though in my case it makes the ignition system inoperative. Absence of the voltage spike at the primary side means that no high voltage will be generated at the secondary side and there will be no spark. Nevertheless thanks for the advise. I have learned something about suppression from this.

I also tied to use switch mode step-down voltage regulator LM2574N-5. It does not make much difference and requires several external components to operate.

When I touch any oscillator pin the PIC stops. I tried to remove oscillator capacitors one at a time and tried different brand oscillator. It didn’t change a lot. I tied to use external clock source. I used 16f628 at 20MHz powered from and independent power supply. Using High and LOW command in a tight loop gave me 500kHz signal. So I adjusted my original code to keep the length of the driving signal the same and changed configuration fuses accordingly. It made the system less stable.

I also changed the MOSFET (had to because it stopped working )to IGBT 12N60A4. I think the reason for the MOSFET to fail was the clamping voltage. So I lowered it to 200v intending to rise it later to 300v. It improved the stability. The problem is still there though.

[zerosvarka]

Hi Everyone,

My IGBT stopped functioning as well as the mosfet. It measures short circuit between Collector and Emitter. I replaced IGBT with vb921 dedicated coil driver. It made no difference. After a while the mosfet driver TC1427 stopped working and got really hot. I increased the value of the resistors at its input and output to 10 kOm. Installed the other TC1427 and after several firings of the coil it got really hot and stopped working. So I ordered more parts

[Acetronics2]

Hi, Zerosvarka

First of all, you should use an ignition dedicated Ibgt, like

- ISL 9V5063 P3
- IRGS 14 C 40
- MGP 15 N 40
- FGB 3040 CS

...

they include their own overvoltage protections ( your Zeners are Wayyyyyyy tooooooo slow !!! ).

Second, you'll have to take great care of the 12 > 5 v section, serial inductors HAVE to be used to kill the spikes from the coil. Two cascaded regulators are not a view of mind here.

Third, the locking problems ...
ahaaaaaa, here, generally, the ground connections are made via the car body ... and are not good at all ( rust, oxydation, paint relieves ... ) . Try "star" connections to a same point. ( like HiFi !!! )

- and remember a breadboard doesn't like high currents or high slew rate signals.


NOW ... I see you use a NiCad 1500-1700 mah as a primary battery ... That's way too weak to drive a Car ignition coil AND give a "clean" 12 v ... ( see Osc pict. ...)





and most of all : ...

don't tell me your Ibgt is the little black thing with a heatsink, on the upper right part of your breadboard ...
Connected to the coil with thin gauge wires ...



begin by a "real world" wiring with ONLY the 5 volts section on the Breadboard ... at least use "veroboard" for the rest of the components ...

The problem is in the wiring and supply ... nowhere else.



This kind of design should work ... and have been working fine for years aboard the ignition we designed for Alpine A110 cars ...

The "Alpine A110 AEPL" ... it's it !!!

Alain

[Archangel]

Hi, Zerosvarka,
I have to agree with Acetronics, an automotive coil uses about 6 - 7 amps RMS, so in reality it is using probably 6 times that in very short bursts, making everything Acetronics said make sense. Big also, what do you expect to keep cool with a little Chicken S_ _ T heatsink like that, maybe a 7805 regulator, but never your switching transistor. Really good connections will lose some voltage at those currents and the really bad connection associated with breadboards will cause heating and big time voltage loss. Cars require ROBUST parts, and then even then they fail. Good luck, and keep us posted, I would like to see the results when you have it working.
JS