I'm an hobbyist without experience in electronics, but I like to understand how the things works. Just now I'm playing with a PWM solar charger circuit.

After some research I produced a schematic, but after the first breadboard test it doesn't work.

My goal is to make a working NMOS PWM solar charger (sure I can buy it for a few dollars, but I'm trying to learn how it works.)

This is my circuit:


  • The gate driver is working fine, at GPV_DRV I have a perfect PWM in the range of 0-12V (so, enough to turn on the NMOS)
  • MOSFET Q6 is working fine, I can switch on\off the load on the line, or dimmer a led light, etc.
  • I can turn ON Q2 and Q8, the two back to back MOSFET with common drain (to protect the panel from reverse current and to connect\disconnect the solar panel from the battery)

The problem is, when I put GPV_DRV to GND, the panel don't disconnect from the battery, Q2 and Q8 don't turn OFF.

I read a lot to understand how MOSFETs and specifically NMOD works, but I can't figure out the problem. Can you help me to understand what is wrong with my schematic?

I tested this circuit with a 40W panel (erogating 2,4A at the time of test) with a VOC > 20V and a 12V lead acid battery with a voltage of 12,6V.

When I connect the panel and the battery, the voltage on BATT+ becomes 13,1V (with the battery absorbing 2,4A,) but if I connect GPV_DRV to GND nothing happens. The panel do not disconnect.

The MOSFETs used for the test: IRFB7545Pbf from Infineon.

So, I found another post as reference: PWM solar charge controller MOSFET

It's the same configuration that I'm trying to replicate.

I did a simplification forgetting for now of the first MOSFET (Q2), trying to make working at least the second one.

As per previous comment I checked the Gate Voltage, it seems to be good:

Yellow: Gate Pin
Blue: Batt+ (12,9v when disconnected)
Purple: Supply (14,8v when disconnected)
In all screenshot, battery are absorbing 2,5A no matter of Gate Status:

Pin on OpAmp Input Set to GND:

enter image description here

Pin on OpAmp Input to PWM 50% duty @ 100Hz:

enter image description here

Pin on OpAmp Input to 3v3

enter image description here

Now I'm confused and I can't understand why it's not working (I used a new just unpacked MOSFET, just for avoid bad part issues (I buy it on Mouser so not fake for sure).

  • \$\begingroup\$ have you tried to swap D and S of Q2,8? \$\endgroup\$ Aug 28, 2021 at 17:14
  • \$\begingroup\$ I tried to do that, and the result is that when I put common Gate to GND Q8 become very hot (maybe it conduct over body-diode?). But the result is the same, no change in current absorption, I have 2,6A absorbed with Gate @ 12V and the same with Gate on GND. Thank you for the suggestion, I can't figure why don't work. \$\endgroup\$
    – CamelSat
    Aug 29, 2021 at 12:59
  • \$\begingroup\$ So you have actually seen some promo drawing without a schematics and you made a circuit believing it should work? \$\endgroup\$ Aug 30, 2021 at 14:39
  • 1
    \$\begingroup\$ No, I produced the schematics in this post with my very limited knowledge in electronics. After some research in this forum I met that post with that promo drawings with a similar pseudo schematic. But like I can't say that my schematic work I can't say that the other schematic in that post is working. \$\endgroup\$
    – CamelSat
    Aug 31, 2021 at 14:00

1 Answer 1


LM2904 is not a gate driver, rather an opamp. You can't control the back to back N-MOSFET switch in a such way. You would need a gate driver with a charge pump. By the way, they should have a common source, not drain.

The only solution is to replace the Q2 with a diode.


N-MOSFET transistor has in-built intrinsic diode. As you may see the M1 will still conduct even if the transistor is OFF. That's why the promo schematics shows two back to back transistors, a workaround to eliminate the diode conduction.


simulate this circuit – Schematic created using CircuitLab

For controlling the N-MOSFET you have to change Vgs gate-to-source voltage. One way is to connect both transistors with common source.


simulate this circuit

As you might see the S potential has to be different from GND, it has to be floating. So you would need a isolated dc/dc converter or a gate driver IC with a charge pump.

The second setup is a common drain as from promo. You can see that with your opamp circuit you can control just the lower transistor M1, for the upper M2 you do need an isolated dc/dc converter or charge pump IC gate driver.


simulate this circuit


There is no simple remedy. Further doing PWM without an inductor (buck converter) makes no sense. Why would you like a PWM?

  • \$\begingroup\$ The Diode is the simplest solution but there is also the problem of the voltage drop (and maybe some heat to dissipate?). But even with diode the switch is not working so.. I have to fix it for first. \$\endgroup\$
    – CamelSat
    Aug 29, 2021 at 13:02
  • \$\begingroup\$ Yes, I used the opamp as gate driver, I'm wrong? As Mosfet datasheet I read that for fully turn on the mosfet I need at least 8-9V, but I have only 3V3 output on MCU. So my idea was to use an opamp supplied with VBATT to amplify a 3V3 PWM signal. It works, but maybe it don't work as Gate Driver? \$\endgroup\$
    – CamelSat
    Aug 29, 2021 at 13:05
  • \$\begingroup\$ The OPAMP idea is perfectly fine. You don't need a gate driver. The only thing that matters is the Gate-Source voltage. When you are trying to turn it off, measure the Vgs voltage on both MOSFETs. If non-zero, then you're not turning off the device. Check the wiring again and the pinout of the device. If one of the MOSFETs is getting hot, that's an indication that the body diode is conducting. \$\endgroup\$
    – Randy Nuss
    Aug 29, 2021 at 13:35
  • \$\begingroup\$ Thanks, I checked it with an oscilloscope, and the voltage seems to be good, it go to 12v to 0v. I'm investigating for other mistake now, but wiring seems ok also \$\endgroup\$
    – CamelSat
    Aug 29, 2021 at 15:31
  • 1
    \$\begingroup\$ Hello Marko, your reply is exhaustive and now I understand the problem. I need PWM to do Battery Charging. I will switch to P-Channel Mosfet. With the common ground between Solar Panel and Battery I think it will be easier to make thing working. I'm just simulating some schematics. Thank you for the explanation. \$\endgroup\$
    – CamelSat
    Sep 4, 2021 at 12:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.