5V or Nothing - the 5V DC-DC Buck

[ACharnley]

Hi All,

In my quest to develop the greatest bicycle dynamo to USB converter going, I've hit a final snag. When the user is at a standstill and moves off the voltage generated is less than that required for the DC-DC, beginning at 0v (naturally) and increasing. Most USB devices can handle this but some have their fancy electronics lock out, so that when the 5V is attained they wont be charging.

Thus what I need is to turn the voltage on at 5V or not at all, which sounds like a comparator with a mosfet, however the Vin to the comparator would also be this changing voltage. My only other solution is the old mechanical relay, a small 5V pcb type. Any other ideas?

(note, that voltage dropout is critical).

Andrew

 

[zahwi]

Or a zener diode at the gate/base.

 

[dr pepper]

Depends on the topology, for instance if you were using a proprietary smps controller ic there is often a input voltage lockout, you set this to a voltage that the ic is guaranteed to generate 5v from, so your either making 5v or shutdown.
Because a pushhog might be travelling at a speed where the circuit could be switching on/off rapidly you could put a capacitor somewhere to slow this down, as some phones use more power this way due to the backlight allways being on.

 

[ACharnley]

The problem is when there isn't 5V some devices lock out requiring a hard reset before they'll reinit. Usually this is 'smart' powerbanks. The DC-DC chip I'm using at the moment doesn't have a lockout (seems most dont).

Unsure how a zener with a mosfet would do it given I need 4.5v to turn the mosfet on. I did try using a mosfet hoping the Vgs would be more instant but it gradually turns on (unlike a comparator), so the problem is the same.

 

[zahwi]

Of course you can't have the switch on voltage equal to the switch on voltage. With a zener and mosfet you can use the output 5V as a feedback to create hysteresis.
You are right, if both voltages need to be accurate then you do need a comparator or an op amp and a ref zener.

 

[MikeMl]

You have a chicken and egg problem. You need to know that the output of the smps is 5V before you turn on 5V to the USB. Seems like you need some hysteresis... Do you put the switch behind the SMPS or ahead of it? If ahead, what voltage do you sense. If behind the voltage you sense cannot be the USB voltage...

Here is a stable circuit that turns on at ~5.25V and turns off at 5.02V using as few part as I can dream up... Using a modern PMOS, the voltage drop is minimal even when switching 1/2A.

 

[ronsimpson]

Some PWM have a pin that measures input voltage. If the PWM is set to start at 10 volts (not 5.3V) this might keep the output at zero until there is enough power to get going.

Most "boost up" PWM just pass input to output until they start. Another type can block in to out until it starts. This way if the input is too low the out will be zero.

 

[Flyback]

given I need 4.5v to turn the mosfet on.

…there are plenty of mosfets with vgs(th) less than that…try “logic level” mosfets.
You are saying tat you need to switch voltage in to the USB device only when the voltage has reached 5V……but does your USB device have significant input capacitance?, because if it does, and you suddenly switch the 5V into it, then the input voltage will immediately fall to below 5V anyway, because the usb’s input capacitance will rapidly draw inrush current from whatever capacitor you have at the output of the dynamo.