New Type of Coil Driver? (Hall Effect Inside)

[mark.f]

So, this is just a little thing I thought up for driving an ignition coil efficiently and at "maximum power" (an optimum frequency).

Rather than doing the convoluted time constant calculations for your coil/capacitor and coming up with an optimum frequency for driving your ignition coil at, why not simply measure the magnetic field around the coil and switch current through the primary accordingly?

Here's the circuit I came up with.

buzzer2.png (9.08 KiB) Viewed 3949 times

Basically, it uses the strength of the magnetic field around the coil to judge inductor "charge" and switch current accordingly. The two comparators and potentiometers are used to "trim" the sensor values.

Components (as far as the hall effect sensor is concerned) are tentative, as I don't really know any "recommended" linear hall-effect sensors (which I was looking to hopefully get some recommendations on). The opto-isolator (between the MOSFET and the output) would probably be homemade unless I can find a reasonable discrete one.

The pair of NAND gates forms a "combinational" logic block with the following truth table. The second gate is used to "latch" the first even after the second input goes low.

table.png (3.83 KiB) Viewed 3949 times

This circuit was initially going to use a pair of reed switches with a coil to "latch" the second one during magnetic field decay, but the fiddley-ness drove me to create the solid-state implementation.

It would require some measurement and adjustment, but nothing that can't be done quickly with an oscilloscope or even an Arduino.

Any comments/suggestions?

[mobile chernobyl]

Just briefly looking over your concept (So what I'm about to say may have already been considered and may be completely bunk given your scenario)

- have you considered using a variation of the ZVS (Zero Voltage Switching) Style drivers used for flyback coils? They work in a semi similar manner where they take field effect into consideration when switching the polarity, and may be a simpler alternative to your design criteria, avoiding the use of logic gates and other things like opAmps.

http://wiki.4hv.org/index.php/Flyback_t ... er_Drivers

[mark.f]

Indeed, I did consider a driver which measured voltage across the inductor (not the same thing, and I am glad you posted that link as it is a lot simpler than the design I had scrawled down).

Most of the parts for the circuit I posted are "jellybean" (except for the hall sensor) enough to justify it, I will however reconsider after looking at the page you posted.

Thanks for the feedback MC.