Machinist

I’m looking at the FAN3224, a gate driver capable of several amps at 5V to run my MOSFETs. I’ll probably play with direct driving them with the ATtiny at first just to familliarize myself.

The original circuit is running some sort of bridge configuration. I’ve used a 1.6ohm load for years, with wattage set to 14.5 or 14.8. However, the orignal is capable 25w in the lowest output version. I’m not sure what resistance range it’s capable of pushing that 25w.

And, yes, the goal is to build a variable DC supply. I’m finding it really interesting. I’m looking at the synchronus inverting buckboost. I think my final version would use half the components of the original.

If I understand this stuff correctly, (real big if), with the convertor pushing negative voltage, I sidestep high side drive problems. My load is dumb resistance so it doesn’t matter how noisy or negative its power is.

So the ATtiny has an ADC capabale of running in windowed mode. A lot of the peripherals are able to run independently of the CPU, you can link them up with onboard programable logic called CLC. So, using logic I can tie the ADC to my PWM at pretty fast speeds. It also has a selection of internal voltage references. I’m planning to use some sort of voltage divider for feedback to the ADC, using the logic to vary my duty cycle when the voltage goes out of bounds.

Before I get that far out in the weeds, I need to play with mosfets, pwm, an inductor, and my new scope.

Plan is for the ATtiny to actually mostly be asleep on the battery voltage. The 5V driver IC has an enable pin and goes into bypass mode when it’s not enabled. Only when it wakes up will it turn on the 5V.

I want the 5V for the 4.5V gates on my MOSFETs, the OLED display (IIRC, it will also run on 3.3V), and the 20mhz for PWM and ADC.

DC/DC convertor. I’m hoping I can make a synchronus inverting buckboost supply in the neighborhood of 25w/6V max.

I’ll be reading up on push-pull. Thank you!

Main power source is a single cell lithium, so 3.7ish volts. Supply for the ATtiny is a 5V driver IC (I’ve already got it working).

The convertor I want to build will be inverting as the load is basically a pure constant resistance. (14.5w 1.6 ohm typical). I think that means I sidestep all the problems with a voltage higher than my control.

I’m hoping to use the window mode of the ADC and the programmable logic (CLC) to actually control the convertor mostly independent of the CPU.

I’m bumping against the edges of my ignorance here. I haven’t done any calculations yet. I’ve been reading the DC/DC Book of Knowledge by Steve Roberts.

I’ve got a working 5V switch mode supply using a driver ic for the ATtiny. Plan is to use that supply and the ATtiny to variably control a larger inductor/MOSFETs. This will power a resistive load around 25w max, 14.5w typical, 1.6ohm so roughly 4V/3A typical.

The data sheet gives an absolute maximum of 40mA for output on a pin with 5-10mA being ideal. My understanding of the chart above is that the voltage will also sag below the 4.5V gate actiivation. I’m assuming trying to switch the MOSFETs with just the ATtiny either leads to a blown output or an overheated MOSFET.

I guess I need to figure out my target frequency and work backwards from there. I’ve got a working reference circut, but all the ICs and MOSFETs have been laser obufuscated. Inductor is 1.5uH and 27A.

So I’m assuming I need a higher frequency than what I can push with the ATtiny but I don’t know that.

Sorry for the novel. Writing this down helps me work the problem. Do you see anyrhing wrong in my assumptions/reasoning?

I was thinking bjt or opamp to drive the mosfets, but I’m not sure.

And yeah, I’ll be scoping it as I start testing. I’m just still wrapping my head around the synchronus inverting buckboost I think I want to use.

I’m hoping to end up with a pcb that has the features of an obsolete ecig that my spouse and I prefer. I’ve probably purchased a dozen of them over the years. The switches always fail.

Given the politcal climate towards ecigs (What about the children???), my particular harm reduction method continues to become more difficult to source. So, I’m just going to future proof things and build my own. I doubt the design will be worth any money so I’ll probably open source it if it works.

So pretty serious about durablility, it would be a tool for daily use. I plan for switches to be on a daughter board so I can just change them out.

I’ll be checking out the video this evening.

Thanks!