One Way Voltage Block

[kevins1]

Hi,

I am working on a circuit design and have run into a problem I am unsure how to address. My design leaves me in a situation where I need to have a charge pump (LTC3204-5) outputting 5V connected to my devices voltage output line. However, this same line may in certain situations be connected a a voltage input when connected to an external device.

My design is a for a handheld transmitter that connects to various sensors for data collection and communication. The issue arises from the fact that some of the sensors have their own internal power which would be supplied to my handheld on the same line as my charge pump is outputting.

The charge pump has a 5V output and a 5.5V max on the output pin connected to the voltage line. The sensor may put out up to 12V which would surly fry my charge pump.

I have considered a diode to prevent current flow, but the charge pump chip manufacturer's FAE indicated that the voltage present would likely still cause issues despite a lack of current.

As a result of all of this, I need to design my circuit such that voltage can go out of my handheld, but that it is not able to to enter and cause damage to the internal circuitry.

I am currently thinking of putting a load switch (FDG6324L) in line which can handle voltages in excess of 12V on the output pin, but would stop the voltage from going out the input pin. I would design this switch to be constantly turned on and simply be using it as a voltage block. I am not sure if this will work (It was suggested to me by another engineer) and am in the process of getting samples to test.

Please let me know if you have a way to design this circuit simply or have any other suggestions as to what I may do about this situation. I am running out of time and ideas and need to finalize this power supply design in the next week. Any feedback would be much appreciated.

Thanks,
Kevin

 

[VEBill]

Series diode.

Then address whatever potential problems you or the OEM thinks that this approach may cause.

 

[mkeith]

The series diode should work fine. If you put a resistor to ground as a load on the charge-pump side of the diode, any reverse current that flows through the diode will have a path to ground, and there will therefore not be any problem with over-voltage at the charge pump output. If there is a load on the output already, then no resistor will be required. Since diode reverse current is very small, the resistor can be pretty big (1 MOhm).

Things to note: The diode will have a voltage drop, so sensors which attempt to draw power from your circuit will not see the full 5V output. The drop across the diode depends on the specific diode, the diode temperature, and the current flowing through the diode.

A Schottky diode will have a smaller voltage drop than a regular silicon diode.

If even a Schottky cannot be tolerated, some type of switch could possibly be used, but you would need additional circuitry to turn the switch off when the over-voltage condition is present.

 

[bradrs]

Do you only connect to a single sensor at a time? Or is there a risk of one sensor supplying 12V back into the 5V supply and damaging another sensor?

Is there a reason you couldn't remove the 12V supplied from the sensor, at the sensor?

 

[BrianG]

Hi Kevin,
You can't use the FDG6324L to isolate the charge pump from a higher voltage. If you refer to the circuit diagram on the data sheet, the P-channel FET Q2 protection diode will allow the +12V straight back in to fight your charge pump.
You may be able to provide the required isolation using an N-channel switch: this will have a protection diode the 'correct' way round but you will need some additional circuitry to sense the presence of greater than +5v from the sensor to switch it to the off state.

 

[kevins1]

Thanks everyone. I think I have decided to use a Schottky diode to prevent any reverse current. I tested this yesterday and it appears to prevent any damage to my charge pump. I appreciate all the insight thus far, it has been very informative.

-- Kevin