Ignition Coil Driver problems / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Ignition Coil Driver problems

Move Thread

LAN_403

Gsu

Mon Jun 25 2007, 07:53PM

Registered Member #858 Joined: Mon Jun 25 2007, 07:23PM
Location: Holland
Posts: 3

Hi there,
I'm using this circuit to power an Ignition coil:
I'v blown up over 20 LM393's and 4 MOSFETS so far

Where:
A 555 timer and an LM393 comparator (IC2) give me a nice square wave with PWM at the gate of an IRF840 MOSFET (500V 8A).
Of course this circuit needs proper back-emf protection when driving an inductive load like an Ignition coil.
I'm not sure but I think the voltage on the 'load-' easily rises over 1kV, if I hook up my multimeter (which only can take 1kV) it reads about 1 to 1.2kV when doing single discharges.

A snubber should damp the rate of rise right ? well they dont seem to matter anything for me. I'v tried snubbers (.1uF, 10nF, 8nF with or without a resistor in series (1-10kOmh)) and varistors over the load. an the capacitance of the snubber cap be too high ?
It just popped in my mind that I'v been using 1kV caps, they'r all still working so I thought the voltage didn't rise over 1kV (I only jÃºst measured the 1.2kV spikes with my multimeter) since I thought the caps would break when driven too far or are MKS caps that resilient ?

With the varistor (60V one) over the load I stopped blowing up LM's but after about 2min the FETs started to die on me.

The circuit often holds up for quite a while when there is a discharge, if the spargap is too large and there is no discharge I instandly blow up an LM393.

I'v seen the ZVS circuit beeing extremely popular here (still have to look into it) but even tho you might want to recommend it to me, and if it actually performs better I will build that afterwards, but I am at a point here where I absolute H A V E to get this circuit properly protected, working, and able to withstand switching without a discharge.

My goal is to drive a jacobs ladder with 2 or 4 coils in antiparallel alltho I might be better off using a flyback when I decide I want the current of multiple coils in series

Dr. Dark Current

Mon Jun 25 2007, 08:24PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

If you want a simple driver, use just a 555 with its output driving a MOSFET (IRF640 or similar.) This circuit will not likely fail until you push it into hundred Watts. Most mosfets today are avalanche rated which means that when their D-S voltage rating is exceeded, they will just clamp it like a zener diode and don't get damaged (of course unless they overheat). If you need PWM you can use such driver IC as SG3525 (very good) or the well-known (but not so good IMO) TL494.

Gsu

Mon Jun 25 2007, 08:43PM

Registered Member #858 Joined: Mon Jun 25 2007, 07:23PM
Location: Holland
Posts: 3

Well then what could be killing my LM393's

And could I have blown my FETs by supplying 60V thru the varistor, thru R3, into their Gate ? Overheating is not it, I have had them attached to a PC heatsink and I have never let them heat up to a temperature that you cannot comfortably touch.

sparky

Mon Jun 25 2007, 10:37PM

Registered Member #530 Joined: Sat Feb 17 2007, 07:56AM
Location: Victoria BC, Canada
Posts: 178

No, the easiest way I know is using a 10uF 330V cap in series with a triac all purpose 600 w dimmer switch.... you get 60 KV off a single ignition coil and its plenty powerful! Damn near indestructible too..

Sulaiman

Tue Jun 26 2007, 09:13AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Here are a few points that may help;

Due to the 555 you are switching the transistor at a fixed frequency set by VR1
Due to the 393 you are switching the transistor on for a fixed time each cycle set by VR2
So the overall result is that you are trying to pass constant POWER through the ignition coil.
SO - When there is no load (an arc is a load) where does all the power go?

Due to the (economical) construction of an ignition coil,
not all of the magnetic flux generated by the primary is coupled to the secondary.
This results in significant primary 'leakage inductance'
So even with a load significant energy is returned to the primary drive circuit each cycle.
This is why really good snubbing or clamping is required.
(e.g. I use 10x 1.5KE350CP in parallel on a heatsink across my ignition coil)
(I don't like relying on the avalanche capability of the transistor, just a personal thing)

There can be quite a significant current FROM the gate of a mosfet or igbt in a circuit like this,
due to drain-gate capacitance and the fast voltage change after transistor switch-off. (ringing voltage)
I doubt that the 393 can handle it so it 'dies'
Use the 393 to drive a transistor which connects to R3/gate (you'll have to exchange + and - inputs of the 393).

For this kind of circuit I usually put a diode from 0V to gate and another diode from gate to V+
to 'catch' any 'ringing' currents from the gate.

MKP and MKT capacitors ARE quite tough, they are commonly tested at 2.5 x rated voltage !
I've not come across MKS capacitors, I do have some 18nF 2500V KS capacitors
I didn't know styrene could be metal coated.
(M means 'metalized', MM means double-sided metalised, no M means foil)

I use just a 555 directly driving either mosfet or igbt (with diodes and TVSs) for this kind of circuit, easy/cheap.

Hope this all helps rather than confuse

thedatastream

Tue Jun 26 2007, 09:00PM

Registered Member #505 Joined: Sun Nov 19 2006, 06:42PM
Location: Yorkshire!
Posts: 329

Perhaps, as Sulaiman suggested, an anti-parallel diode across the load would help to prevent damage to the MOSFET.

If I understand your circuit correctly, you are comparing the voltage ramp across C1 in the 555 circuit against a reference formed by VR2 to form a PWM circuit. However, the drive formed by IC2 and R3 probably isn't man enough to turn on the MOSFET fast enough.

You can use another 555 as a buffer by connecting pins 2 and 6 to the output of the comparator IC2, pins 4 and 8 to Vcc, 1 to ground, pin 7 and 5 unconnected and pin 3 to the MOSFET via a small resistor (say 2R2). This may invert the signal, but I cant be bothered to figure out which ;)

If this fails, you could try an RC snubber across the MOSFET.
You should also use a nice large electrolytic decoupling capacitor across the rails next to the ignition coil
You should aslo keep the loop are for the primary current small (supply +ve > load > MOSFET > supply -ve) to prevent ringing due to stray inductance.

HTH
James

Steve Conner

Wed Jun 27 2007, 10:02AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Try a better MOSFET driver than the LM393, too. It has an open-collector output, and you could be getting kickbacks into it.

Gsu

Wed Jun 27 2007, 01:38PM

Registered Member #858 Joined: Mon Jun 25 2007, 07:23PM
Location: Holland
Posts: 3

Thanks, I think these comments will prove great help

Sulaiman wrote ...

Use the 393 to drive a transistor (you'll have to exchange + and - inputs of the 393)

Good you mentioned + and -, I tried without exchanging them

sparky

Thu Jun 28 2007, 07:51AM

Registered Member #530 Joined: Sat Feb 17 2007, 07:56AM
Location: Victoria BC, Canada
Posts: 178

Why don't you try the dimmer switch with a 10uF 330V rated motor run capacitor... $10 for 60kV @ 150W .... maybe more? Why not!!!

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob