Precharge schematic critique (24-60V)

[skyfire360]

Hey all!

Brand new to electrical engineering as of a week or two ago, so I'd love to get some feedback/critiques on my probably-going-to-release-the-magic-smoke schematic. I'm looking to design a soft-start precharge circuit for starting up circuits with substantial input capacitance (>=1.5J @ 24V). The main requirements are:

• 24-60V compatible
• Constant current pre-charge
• Enable/disable via low-current switch
• Protection against inrush if accidentally connected with switch enabled
• No sharp current/voltage deltas during precharge or transition
• LED indicator to show when capacitors are charged
• Connection order doesn't matter

This is the circuit I've come up with (Falstad link):



Note that the switch would control the 12v regulator, but I can't find one in Falstad so I had to simulate one.

The theory of operation is that once the battery and downstream load is connected and the 12v regulator is enabled via the switch, the precharge power MOSFET gate begins to charge. Once it hits the threshold and maximizes in the non-linear region (limiting current, in this case to ~3.5A), current begins to flow through the downstream capacitors to a 2Ohm power resistor. As it charges, an op-amp measures the voltage on the negative side of the downstream circuit. Once the difference in voltage falls below the voltage on the non-inverting amp input, the op amp turns on, delivering 12v to the final stage. The transistor is switched on, causing the gate of the final power MOSFET to begin charging. Once the gate threshold is hit, it turns on, bypassing the power resistor and enabling full current to flow.

AFAICT it works, at least in simulation. The current slowly increases up to a constant level, and voltage across the capacitor begins charging linearly...



...the current slows down as the capacitor fills up...



...and the final transition to the power MOSFET is a smooth, small "bump" in current as the last bit of capacitance charges through the power MOSFET:



Charge times:

• 1mF@24V: 19ms
• 10mF@24V: 140ms
• 1mF@50V: 28ms
• 10mF@50V: 240ms

Does this seem reasonable? Any obvious mistakes/concerns? The only point I don't think I have solved is #7 above. If the switch is off, the battery/load connection order doesn't matter. If the switch is accidentally left on and the load is connected first, followed by the battery, everything works. However, if the switch is on and the battery is connected first, the op-amp thinks the voltage across the capacitor is the full voltage and happily charges the cap in the lower right. That's a problem, since when the load is eventually connected the cap keeps the power MOSFET's gate high and allows hundreds of amps to go through before the cap can drain and the remaining circuity catches up. I figure I can be careful to always connect the battery last, but if possible I'd love to protect against my own absent-mindedness!

[Zero999]

Why not simply use a current limiting resistor and a timer relay to bypass it, after a certain length of time has elapsed?

[Siwastaja]

Why not simply use a current limiting resistor and a timer relay to bypass it, after a certain length of time has elapsed?

This needs extra monitoring against the case that there is, for any reason, excess load preventing proper precharge.

Timer is seldom a good idea, because it's equally easy (if not easier) to implement a voltage comparator instead.

In any case, the resistor needs to be protected against overheating, for example, by using a PTC (polyfuse) in series, thermally coupled to the resistor. If there is short or excess load, the PTC trips, precharging halts, and the voltage comparator never enables the relay.

[Zero999]

It depends on what you're doing. A timer relay is much easier than a comparator, as the whole thing can be purchased as a module for minimal cost. I've seen plenty of audio amplifiers with a timer relay + resistor soft start and they work quite well. A fuse or PTC resistor provides short circuit protection, whether the timer relay activates or not.

In this case the load is motor, which will could be more of an issue, than charging the capacitors. The initial current surge due to the motor will be smaller, than the capacitor, since it'll have a higher resistance, but the surge will last for longer.

It's also questionable if such a large capacitor is necessary for a motor powered by a battery?

[Siwastaja]

Oh, if it's actually a motor directly connected, such a precharge circuit won't work. The motor is almost like a short circuit when the RPM and back-EMF is zero at the start (stall condition). This startup is where the highest currents happen, over 10x the nominal motor current is well possible. So any precharge circuit won't ever succeed rising the output voltage at all, and any timer relay would just snap on fully, bypassing the precharge completely and doing a hard power-on cycle. A voltage comparator based would just never activate, so these two have different error modes.

I don't know about the simplicity of time delay. Sure, time delay modules exist. Won't voltage comparator modules exist as well? Look at Ebay, I'm sure something pops up.

And yeah, such a direct motor connection doesn't need capacitors.

But if it really is a motor, I strongly suggest building an actual motor controller, which would be a MOSFET, freewheeling diode, DC link caps, a small microcontroller, and a current sense resistor and amplifier. It's a bit more complicated, but might be the right thing to do. This would actively limit the motor current and protect the motor as well. But funnily enough, such a motor control circuit might now need the precharge circuit we are discussing, if it's large enough to have significant DC link capacitance.