LTC7803 in LTspice (Disable briefly)

[Flyback]

Hi,
We want to disable our Buck (LTC7803) then re-enable it, and have it come back on within 1ms......the attached LTspice shows that pulling down the TRACK pin is the best way of achieving this.
Would you agree this is the best way?......Alternatively pulling the RUN pin to between 0.7V and 1.2V is the other way, but results in massive overshoot when re-enabled...would you say this is correct, as often they dont bother modelling the RUN/TRACK pins correctly in LTspice?

 

[crutschow]

as often they dont bother modelling the RUN/TRACK pins correctly in LTspice?

Why to you think that?

I would use which ever inhibit input works best in simulation.

 

[Tony Stewart]

The LTC7803 regulates the VFB voltage to the lesser of 0.8V or the voltage on the TRACK/SS pin. An internal 12.5μA pull-up current source is connected to this pin.

A capacitor to ground at this pin sets the ramp time to the final regulated output voltage. The ramp time is equal to 0.65ms for every 10nF of capacitance.

---------------
You still have excess ringing on step loads due to missing phase lead compensation.

 

[Flyback]

Why to you think that?

Thanks, if you run that sim, you can see that having RUN voltage of 1V doesnt stop it from switching...but datasheet says it should.....LTspice sometimes doesnt bother doing all features...eg, famously, at least in the past, the SYNC pins were never modelled....so you'd put a SYNC signal in...and it'd do nothing.

 

[Tony Stewart]

Make the design work without the SYNC 1st, the SS does work in LTSpice but only if all other inputs are in the nominal range

 

[ChrisP58]

Build it and test.

 

[eTech]

Thanks, if you run that sim, you can see that having RUN voltage of 1V doesnt stop it from switching...but datasheet says it should.....LTspice sometimes doesnt bother doing all features...eg, famously, at least in the past, the SYNC pins were never modelled....so you'd put a SYNC signal in...and it'd do nothing.

Actually, the "RUN" pin does work.
But the simulation of that function is rather poor. If the voltage is changed to 1v, the switching stops but only after some delay, during which, the controller struggles a little trying to maintain a constant output voltage.

The "TRACK/SS" pin shouldn't be used to shutdown the output. That's not it purpose.
Its used to cacade (track) the startup of multiple controllers or, to provide slow startup of a single controller.

 

[simonbramble]

We don't model the SYNC pin on dc/dc converters. If the pin is shared with another function, then it will still appear on the symbol, but (hopefully) the SYNC name will be removed.

I dont see any huge surge of current. There are switching spikes around 4.6ms into the simulation, but this is due to the slow reverse recovery time of the body diode in the low side FET. The high side FET switches on while the low side FET's body diode is still conducting, creating switching spikes. Put a Schottky diode across the low side FET and this will improve things

These spikes are higher than earlier in the simulation because the ITH pin is high, so the overall current into the circuit is higher (the ITH pin controls the peak inductor current). Lower the capacitance on the ITH pin to allow it to discharge quicker. You would think the SS pin would stop this, but the SS pin is low trying to pull the ITH pin low, but the large amount of capacitance on the ITH pin is keeping it up and not throttling back the peak inductor current

Increase both feedback resistors by a factor of 10, then you can add a humane amount of feed forward capacitance across the top feedback resistor and this will improve the transient response - it adds phase lead/phase margin and pulls you away from the point of instability (or 'D' feedback if you are into PID control)

Increase the softstart capacitance. You can still have a healthy amount of capacitance (10nF) on this pin and achieve your startup time

Use a bipolar across the SS pin (like you are doing) as the leakage current of an NFET will swamp the SS current at high temperatures

Or... keep the dc/dc converter running and gate its output using a series FET