LM324 equiv with output very close to GND?

I have a current feedback loop which controls a FET driver which outputs 0-2A based on a 3.3V DAC with 255 step resolution = 13mV per step.



Initially I used a LMV324 but have measured a GND output of 124mV, which makes current control at the low end problematic at a minimum output of around 60mA. According to LTSpice the standard LM324 has an output of around 40mV, which would require a minimum DAC resolution of 4/255 and simulates to an output of about 20mA.

I just wanted to check if there's a reasonably priced alternative to the LM324 which can meet or exceed this ground output voltage?

Vcc is 4.5v-5.25v, +input won't exceed 50mV. Feedback loop has to be kept under 1MHz for the Fet driver at around 0.5MHz.

The high gain (~80) of the left op amp means that any op amp input offset is amplified by that amount.
You need a very low offset op amp to keep the output offset below 13mV.

It you need a 500kHz loop bandwidth, then the op amp must also have a gain-bandwidth-product of at least 50MHz.

So the gain isn't realistic, I can reduce that with a higher sense value resistor, or live with it.

The simulation is working fine with the 324 which obviously doesn't support anything like 50MHz. It's a hysterestic mode buck and LT has it measured at 55KHz with a 22uH inductor so it looks like I was way off with the 500KHz.

That means I need a op-amp with a 5MHz bandwidth, assuming I don't change the gain? I'd expect the simulation to fail as the 324 is only rated a little above 1MHz?

Some of the newer output rail to rail opamps can swing extremely close to ground, but never to 0.0 volts

There are additional tricks, like pulldown resistors, which will reduce the output a few additional millivolts.

In this instance, your best friend are the manufacturer's web based parametric selectors.

I thought about a pulldown but that will affect the DAC input somewhat. At a guess the proper way to do it would be to generate a negative rail, but I can live with it as is.