How to (appropriately) drive an ignition coil and/or a flyback?

[Speakerguy]

This is for HV hobby stuff. I am well beyond college My senior project there was a DSP based loudspeaker system. I am/was much more an analog signals and DSP type guy with a side interest in embedded systems.

Anyway, here is the first attempt at a schematic:

**broken link removed**

+12 to +35 VIN from bench supply

VIN regulated to +5V and an isolated +5V.

Isolated +5V feeds PIC and logic circuitry. PIC communicates via RS232 to PC. PIC outputs 4 signals.

4 signals go through optoisolators, which in turn control the gate of a p-type MOSFET. This mosfet controls base current to the drive transistor, a BU941ZT (NPN darlington specifically meant for ignition coil systems - "very rugged" according to data sheet).

Craploads of transient voltage suppressor diodes (both uni and bi directional) as well as standard zeners all over the place. Several are probably unecessary, but I'll let you guys advise me. Trying to keep anything from frying Most all the parts I could make to be DIP's are (so I could socket and replace them easily).

Thanks!

 

[mneary]

Where does the arc current flow? To "ground"?

 

[Speakerguy]

Generally yes

 

[mneary]

Hmm, that could be a problem. When the transistor turns "off", the only path to "ground" is through +VIN or a TVS. Is it possible to re-orient the circuit so the coil can be ground-referenced?

If this isn't possible, I would suggest a lot of low impedance bypassing from the coil +ve to ground as close as possible to the coil terminals. Also a very fast TVS rated to keep the coil +ve from going too high. And maybe a series reactance (lossy ferrite bead) between +VIN and the coil +ve.

General principle is "know where the current is flowing". And take it where you want it to go by both decreasing the resistance to desired paths, and increasing the resistance to unwanted paths.

 

[chemelec]

This is for HV hobby stuff. I am well beyond college My senior project there was a DSP based loudspeaker system. I am/was much more an analog signals and DSP type guy with a side interest in embedded systems.

Anyway, here is the first attempt at a schematic:

**broken link removed**

+12 to +35 VIN from bench supply

VIN regulated to +5V and an isolated +5V.

Isolated +5V feeds PIC and logic circuitry. PIC communicates via RS232 to PC. PIC outputs 4 signals.

4 signals go through optoisolators, which in turn control the gate of a p-type MOSFET. This mosfet controls base current to the drive transistor, a BU941ZT (NPN darlington specifically meant for ignition coil systems - "very rugged" according to data sheet).

Craploads of transient voltage suppressor diodes (both uni and bi directional) as well as standard zeners all over the place. Several are probably unecessary, but I'll let you guys advise me. Trying to keep anything from frying Most all the parts I could make to be DIP's are (so I could socket and replace them easily).

Thanks!

Cute, But what An OverKill Circuit.

That simple 555 ciruit of mine can drive any number of Mosfets to drive an equil number of Coils.

And its relilable.
Its been running continuous, outside on my electric fence for over a month now.

 

[Leftyretro]

Cute, But what An OverKill Circuit.

That simple 555 ciruit of mine can drive any number of Mosfets to drive an equil number of Coils.

And its relilable.
Its been running continuous, outside on my electric fence for over a month now.

I was very impressed with your simple 555 circuit and have saved it for possible future application. However the OP's circuit does have the capability to vary the drive frequency and duty cycle via software which for a general purpose HV generator could be very useful in bench experiments and applications requiring 'tuning' of such parameters. Not that the 555 circuit couldn't be adapted by changing the Rs and Cs, but I think both circuits have value. Yours for simple, low cost, finished HV applications and the other for general experminational purposes.

Lefty

 

[Speakerguy]

mneary,

I want the current to have a heckuva time finding where to go That's the di/dt spike I want. If I didn't have to worry about blowing the NPN darlington, I wouldn't even have the TVS in there. Once it clamps @250V and allows current to flow, my HV on the secondary goes away. The darlington itself clamps at 350V but I want to make sure it doesn't go poof.

Re: simpler circuits - trust me, I've been there and done it. Both 555 as well as PIC based solutions, simple low side switches, half-bridge FET systems, H-bridges. Problem is it's very easy to make one circuit that works well with the coil you have, but I have access to LOTS of old coils and such and a circuit that works for one coil will get toasted by another. It might be an ignition coil, a flyback xfrmr, the resonant transformer I ripped out of an electronic neon sign transformer, etc.... I might be tinkering with drive voltage, charge times, frequencies, etc - all of which are a real pain to do with a 555 alone since you have to swap out resistors/caps/pots to change your range much. With this setup I could even include something like a Dallas Semiconductor DS1020 programmable delay interfaced to the PIC. It, the signal from the PIC, and an OR gate would get me down to nanosecond timing resolution if I wanted it, all instantly controllable from the PC. I've done that with one previous board that had a PIC on it (no PC, programming was with a DIP switch and pullup resistors) but the DS1020's really like to blow up on me even on the isolated side. Way too sensitive of a chip. Only device on the isolated side of the old ckt's that I've ever lost.

The nice thing about work is, when I need high voltage there, I just go tell the boss I need the Matsusada. 10kV @ 130mA. Serious stuff

 

[multichnl]

use for ignition coils

just out of curiosity, what are you using these ignition coils for?

Interference generator.

Underground wire locater. Make sure to disconnect both known input and output of the wire in test and unplug all computer equipment in the area etc. Use a portable am radio tuned at the low end for a receiver.

 

[audioguru]

I saw a 5.1V zener diode connected to the output of a 5V supply.
A smoke generator?

The 5V zener diode might be 5% low at 4.75V.
The LM317's voltage might be 5% high at 5.25V.
The result is smoke from the zener.

 

[Speakerguy]

I replaced the 5.1V regular zeners with 6.0V unidirectional TVS zeners (faster response time to transients) and the 317 with a TL780-05CKC from TI that specs a +/-50mV output tolerance (4.95-5.05 guaranteed). I was surprised to find a 78XX series regulator that did 1%. Figure why bother with the 317 and 1% metal films and all the extra parts.

Anyway, any other problems in there? I caught one last night that I would have found doing the layout but it was good to catch it earlier anyway. Had the pinouts on the darlington incorrect It's fixed in the newer schematic.

Oh, and I LOVE having work tools to do this stuff with. Not sure if I'm going to mill it out for free on the PCB mill we have here or if I'll just do a barebones order at advanced ckts. If I go the latter route I can make boards available to any who want them cheaply.

 

[chemelec]

I replaced the 5.1V regular zeners with 6.0V unidirectional TVS zeners (faster response time to transients) and the 317 with a TL780-05CKC from TI that specs a +/-50mV output tolerance (4.95-5.05 guaranteed). I was surprised to find a 78XX series regulator that did 1%. Figure why bother with the 317 and 1% metal films and all the extra parts.

Anyway, any other problems in there? I caught one last night that I would have found doing the layout but it was good to catch it earlier anyway. Had the pinouts on the darlington incorrect It's fixed in the newer schematic.

Oh, and I LOVE having work tools to do this stuff with. Not sure if I'm going to mill it out for free on the PCB mill we have here or if I'll just do a barebones order at advanced ckts. If I go the latter route I can make boards available to any who want them cheaply.

If your trying to eliminate Voltage Spikes, Movistors would be far more useful for this application.
Zeners are Far too Slow in Response times to do much good.

 

[Speakerguy]

The TVS zeners I'm using have a <1ns response time (they are the fastest I could find), are MOV's better? I was worried about reliability with MOV's.

 

[mneary]

I think you only want the current on the *high* side of the coil to have a hard time finding ground. Once that secondary current finds "ground", you should make it easy to find its way from ground back to the low side of coil.

It should serve no purpose to be generating excess noise on the low side of the coil.