Post Go back to editing

LTC4418

Category: Hardware
Product Number: LTC4418

Hello,

I would like to Parallel the P-ch FETs in LTC4418 application for sharing the load/power dissipation. My concern is about removing back-to-back FET connection. Can it be removed and connected 3 P ch FETS in parallel? What is the purpose of connected back-to-back fets? To create a bidirectional switch? Can it be removed? How many total FEts can be paralleled? See attached LTspice model for clear understanding. If this is not the right way to connect them in parallel, could someone please show me how it needs to be done? 

4418 Parallel.asc

  • Sarunas,

    You can parallel MOSFETs in this circuits. The MOSFETs are fully enhanced and they will share the current when conducting.

    You have to use back-to-back MOSFETs in the application, because of the uncontrolled path through the body diodes in the MOSFET. In your spice model current can flow through the body diodes from the input to the output even if the gate voltage, Vgs, is zero. This is a high dissipation condition, because the forward voltage drop will the same as a silicon diode. It also means the MOSFET switches can not turn-off if the input voltage is higher than the OV setting.

    Start by looking at the demo board DC2707B.

    Regards,

    JayBee

  • Hello Jay,

    Thanks for a reply. Much appreciated. Both supplies are the batteries by the way operating 20-30VDC. Please take a look at this new LTspice model that i connected back-to-back in parallel? THe problem I see, that the secondary FET's dissipating significant more power when connecting them like this way. Any advice? See the LTspice simulation and screenshots. Or if you want to parallel the FETs for dissipation do you need to have multiple LTC4418? Please advise? 

     Back-To-Back.asc

  • I just looked at the original LTC4418 model in LTspice and the second FET has the same Power Dissipation spike. Why is that? Would that cause a damage to FET? 

  • Another question came up. Driving a capacitive load of around 300uF during a switchoff the second FET see the negative current due to capacitor discharge. All the voltages are positive and none of them goes negative. I'm assuming that current gets discharged through FET body diode. is that expected in this application? The current is at Source side, only can be catched at that point 

     

  • Try this configuration. I have adjusted the soft-start components and changed the MOSFETs to 40V MOSFETs. I would not use 30V MOSFETs if you have a 30V battery.

    You enter 'm=2' on the in the value2 line in the component attribute editor.

    CTRL-RIGHT CLICK on the part to bring up the Component attribute editor. It is a nice way of putting parts in parallel on the schematic without clutter.

    Do these changes and evaluate the model.

    JayBee

  • Hi Jay,

    many thanks for your input. You amazing with your help. I have only last question left. If you place more than 300uF capacitance on a load and you measure the current at a second FET drain (see photo attached) you will get a negative current there a spike of 0.3ms. I think it's because the output capacitors discharge through the FET body diode? Would that cause any problem ? Changed the FETs to 40V as well. Attaching a model of LTspice as well. 

    LTC4418 RATTLER.asc

  • If I zoom on that event, I see:

    What is happening is Q2 and Q3 are full enhanced they are on, and they should be on.

    The battery voltage V1 is lower than the output voltage, so current will flow from the output to battery, V1, The current is charging the battery. 

    In practice the battery should not go from 28V to 0V in 1ms. If the time is increased to 10ms, the load resistor, Rload, discharges the output capacitors faster than V1 is falling and there is no reverse current.

    If the battery was disconnected by a switch or a fuse there would be no current flowing in the reverse direction.

    JayBee

  • Thanks Jay, you are right. 

    Last question and we can close this: Would ceramic capacitors be suitable for input/output filtering rather than aluminum electrolytic ones? 

    Have a great day. 

  • Hi,

    You can use ceramic input and output capacitors. You need to be aware of the transients when 'hot-plugging'. Batteries are always 'hot-plugged'.

    Have a look at Linear Technology AN88.

    https://www.analog.com/media/en/technical-documentation/application-notes/an88f.pdf

    This will explain how to dampen or clamp the transients.

    JayBee