Maker Pro
Maker Pro

IGBT ignition coil driver

mmcl055

Feb 26, 2012
33
Joined
Feb 26, 2012
Messages
33
Im getting 4.13V on the 5V rail. the IGBT is warm and the PIC isn't running as far as I can tell - no LED feedback.

If I skip the power circuit and hook 5V up to the 5V rail then chip runs...
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
330R is too much. Try 47R

Also the zener needs to be connected reverse biased from the resistor and connected to ground so that it absorbs transients over 18V.

I just noticed your output capacitor on the 5V regulator is pretty low too. 10 to 100uF may be better.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
If I skip the power circuit and hook 5V up to the 5V rail then chip runs...

Does the IGBT get hot? (make sure you have a common ground)
 

mmcl055

Feb 26, 2012
33
Joined
Feb 26, 2012
Messages
33
Like this then?

attachment.php
 

Attachments

  • schematic_v3.png
    schematic_v3.png
    32.4 KB · Views: 2,269

mmcl055

Feb 26, 2012
33
Joined
Feb 26, 2012
Messages
33
Actually, 100R is the lowest resistor I have to hand is that OK for the 33R and 47R?
 

mmcl055

Feb 26, 2012
33
Joined
Feb 26, 2012
Messages
33
No you mean like this?!

attachment.php


But are 100R resistors OK?
 

Attachments

  • schematic_v3.png
    schematic_v3.png
    32.4 KB · Views: 1,114

mmcl055

Feb 26, 2012
33
Joined
Feb 26, 2012
Messages
33
I've wired it up as above and get the same outcome. I've not changed the 47R and 33R yet but will try 100R if you think that's suitable, or I'll get the correct ones tomorrow.

I'm off to bed; shattered.

But thanks for all your help. It is very much appreciated. As you can tell I know very little about this stuff. And I realise my questions must be damn annoying; sorry.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
You can put two 100 ohm resistors in parallel to get 50 ohms (close enough for 47, and three in parallel for 33 ohms.

The 33 ohm gate resistor is a guess based on nothing much at all. I'm not sure about your frequency of operation or pulse length.

depending on the frequency, the 1k and 0.1uF resistor on the output of the PIC may be a factor too.

But the most important thing is getting a good 5V rail. The output of this needs to be 5V. Do you have any idea of the maximum current required from your 5V rail?

Yes, the resistor and zener are now correct. It is important to ensure that at maximum current the resistor does not drop too much voltage. If you need a current higher than about 100mA, then another approach may be needed.
 

GonzoEngineer

Dec 2, 2011
321
Joined
Dec 2, 2011
Messages
321
Reading the data sheet, I wouldn't use a gate resistor higher than 10 ohms.

You still need a diode in parallel with the inductance. Read this post carefully!

https://www.electronicspoint.com/need-help-wih-snubber-t244850.html

As to the latch-up of the PIC:

What are you doing with the secondary of the coil. Sparks are nasty things, and a PIC sure isn't an EMP hardened device!

Get that spark far away as you can, and be aware I know quite a few folks who tried this same thing and damaged computers and keyboards on the same bench!

Got any pictures of your layout? Are you discharging that sparc directly to the ground rails?

I did a circuit like this many years ago, had to put the drive electronics in a metal box, and the external.

Good luck!

(I really need to find a way t convert schematics from my PCAD so I can post them here.)
 

GonzoEngineer

Dec 2, 2011
321
Joined
Dec 2, 2011
Messages
321
You can put two 100 ohm resistors in parallel to get 50 ohms (close enough for 47, and three in parallel for 33 ohms.

The 33 ohm gate resistor is a guess based on nothing much at all. I'm not sure about your frequency of operation or pulse length.

Driving an IGBT gate is essentially driving a capacitor. You need to turn it on and off to reduce switching losses. I would add a low esr capacitor (100uf) directly on the rails of the driver circuit, and reduce the gate resistor to less than 10 ohms.

depending on the frequency, the 1k and 0.1uF resistor on the output of the PIC may be a factor too.

Absolutely correct.....get rid of both!

But the most important thing is getting a good 5V rail. The output of this needs to be 5V. Do you have any idea of the maximum current required from your 5V rail?

Yes, the resistor and zener are now correct. It is important to ensure that at maximum current the resistor does not drop too much voltage. If you need a current higher than about 100mA, then another approach may be needed.

Really could use some pictures of the actual circuit.:D
 

Rleo6965

Jan 22, 2012
585
Joined
Jan 22, 2012
Messages
585
Test DC voltage on LM7805 input terminal. It should not be lower than 8V.
 
Last edited:

jackorocko

Apr 4, 2010
1,284
Joined
Apr 4, 2010
Messages
1,284
(I really need to find a way t convert schematics from my PCAD so I can post them here.)

Forget that, windows 7 has a snipping tool. Best thing since sliced bread, else there is always paint. Ctrl + Print screen and then Ctrl + P in paint, crop edit save and post.
 

duke37

Jan 9, 2011
5,364
Joined
Jan 9, 2011
Messages
5,364
You do NOT need a snubber. The whole idea of this circuit is to develop a high votage pulse of about 300V. With a transformer (coil) of ratio 20:1 you should get 6kV out.

A diode across the coil will stop it working.
 

twister

Feb 12, 2012
172
Joined
Feb 12, 2012
Messages
172
Can you program your pic to use a open collector or open drain? That way you won't need the extra transistor. I have built a couple using slotted interrupters and they worked good. You need a pull up resistor and your pic will ground the gate. I didn't use a capacitor on the gate. What is that for? I saw a circuit with a cap across the gate resistor somewhere. They said it helped to turn the transistor off, faster. You will need a gate resistor to ground, to turn the transistor off. I used a 4 meg ohm. I didn't use the diode across the transistor, but that might work better. You need a .1 silver mica cap across the transistor to ground. This will give a hotter spark, not sure why, maybe it stores energy and releases it when the transistor turns off. You will need a 1.2k resistor between the pic and the gate to protect the pic. 12/.010=1.2k Mine puts out 5000 volts which is the same as my old Toyota car. You could also build a capacitive discharge system if you wanted a hotter spark. You can buy a voltage tester for electric fences at a farm store. They aren't too expensive or you could build one.
I am going to build one using the hall effect device too, when I get time, if I could stay off this computer! I like your idea of using a pic, but I have never used one before. Will you be able to advance and retard the spark? You would need a throttle position sensor and a rpm sensor to do that right.
Oh, you will only want to turn the transistor on for a short time, just long enough for the coil to charge, so it doesn't get too hot.
 
Last edited:

GonzoEngineer

Dec 2, 2011
321
Joined
Dec 2, 2011
Messages
321
Wait a minute......it's an ignition coil!

I feel like a an idiot for not catching this on the first post.

Show us a schematic of the ignition coil. (How many wires does it have?)

:D:D:D:D:D
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
A diode across the coil will stop it working.

It certainly will, but it will also tell us if the IGBT is getting hot because of the way you're driving the load (insufficient gate drive, switching speed, or gate voltage), or if it's caused by avalanche within the device.

Gonzoengineer is better equipped than me for this, but I recall a dv/dt issue as well.
 

Rleo6965

Jan 22, 2012
585
Joined
Jan 22, 2012
Messages
585
@GonzoEngineer

Got any pictures of your layout? Are you discharging that sparc directly to the ground rails?

My Multiple Spark CDI project have intermittent misfiring problem and I can't avoid using ground because car battery - terminal was connected car body chassis or ground.

Here's picture of waveform


5pulsewithmisfiring.jpg
 

mmcl055

Feb 26, 2012
33
Joined
Feb 26, 2012
Messages
33
330R is too much. Try 47 R

Also the zener needs to be connected reverse biased from the resistor and connected to ground so that it absorbs transients over 18V.

I just noticed your output capacitor on the 5V regulator is pretty low too. 10 to 100uF may be better.

OK I've done all of these recommendations and the circuit looks like this. Although I have just noticed that in place of the 1N4007's I have 1N4004's does that matter?

attachment.php


The 5V rail is now 4.94V. Good enough?

The gate of the IGBT now switches from 0 to 7.08V, however with just the multimeter in place of the coil I get 12.1V to 13.1V when the gate switches and the IGBT get's quite warm. I'll recheck everything I guess....

Here's a couple of photos of my setup.

IMAG0125.jpg
IMAG0123.jpg
 

mmcl055

Feb 26, 2012
33
Joined
Feb 26, 2012
Messages
33
Actually, I'm totally wrong. Not sure what I've been smoking....

OK, the output from RB3 switched between 0 - 5V. good.

The other side of the 1k resistor switches between 3.7 - 7.2V

The gate on the IGBT is 4.8 - 5.5V
 
Top