PWM Mosfet voltage working chart?

You should get some 0.1µF capacitors. One should go across A0 near the board. I would but one at both the input an output of the DC/DC.

Also keep in mind you should limit the Gate to Source voltage to 12 to 15 Volts (best device is a Zener). If you put the PChannel on 28 volts and pull the gate to ground.... "pop goes the MosFet"

thanks for the advices

allready was calculating a voltage divider instead of zeners to keep the Vgs between -12 and 0 (28.7V) for the P-channel Gate

Resistors will work fine, however there will more apt to go linear if the input voltage is low and you try to turn it on.
I understand this is a condition you expect to never be in, however good protection is always helpful on a hobby design where the added few cents doesn't matter.

Yo mean to put a ceiling on the Gate of the P-channel to avoid go linear?

For what I've been reading/searching when "linear" the mosfet Resistance increases a lot and thus the power disipated and lots of "not so happy" things can happen with to much heat, right???

so whats the pourpose of PWM on a POWER Mosfet? always Run in saturation mode or OFF???


With a zener the supply could drop all the way to 10 volts and the Mosfet will be/ can be fully on.
If you use a divider the gate voltage will drop into the linear region at a much higher input voltage.

I realize the Arduino will not be turning the device on unless the voltage is above some high level, however if there is an error or an unexpected condition it is best to keep the mosfet on instead of partially on .

with a mosfet driver is easy to make the PWM go from the Arduino to the BJT and then to the MOSFET Gate, but here , when switching the HIGH side (28.7V) with a MOSFET it would be recommended not to go beyond saturation.

Does this mean in this particular situation , that the PWM way of working only gets used to turn on (saturation) or off and not the values in which it will be linear/high resistance/hot/problems???

(and by this I'm asking, should so I program a way to make a cicle of PWM 100% and then PWM 0% , to control the floating part of charging batterys?)

Not sure what you mean here.

This is the circuit I would expect. Only I would put a 12V zener across R4

by "beyond saturation" I meant going into linear , as by all I've read and understood from everybody's advice , the resistance would go higher

so I should (in this specific situation) avoid using a Power Mosfet as a variable resistor and just a "smart" switch with a charging logic as to how much time charge and not charge

I would say Yes. In any such design, using a device in the linear mode (as opposed to full on or full off) will convert a lot of energy into heat.

We haven't mentioned the charging current (except your reference to 2A). I built a small UPS for a 5V device. I recall the max charging current for the 18650 is 1A. I don't know what your Solar Panel can provide. You might look at this page.

solar panels deliver 35V 2Ah (soon more to come to reach 8Ah) which will charge a 7S44P Battery Bank

so 44 x 18650 in parallel will receive nicely 8 amps of charge (180mA each)

I was giving more time looking for schematics and ways to connect all...

figure 1 is to switch the high side from solar panels to the battery

Using 5.1K or 6.8K would get the max Voltage of Solar panels (47V without load) under the 20V Vgs of the P-Channel Mosfet , but when charging voltage drops to the battery voltages which goes from 21 to 29,4V (this translates from 8,5-12V as Vgs in the GATE.


I remember you told about the zener to control the Gate voltage which controls the peak (47V translates to 21,2V which the Zener would block) and the operating value of the Battery would go from 9,5 - 13,2V (battery from 21 to 29,4V)

If the zener way is the way to go, is it better to have a 10V zener or just a 12V or even 15V zener?

how much wattage should I calculate for the zener?

Zener voltage doesn't matter too much should be between 10 and 15. Below 10 you might not be fully on (close but not as much as it could be) and you want to stay way below 20. I though 12V would be rather common.

Zener Power:

From your dwg I assume you want the Mosfet full on between 21 and 29 v so we will calculate for between 20 and 30V

So at 20V we want the 10V zener to just be conducting. That makes the 8.2k --> 10k
at 30V the power is:

I = (30 -10)/10k
I = 20/10k = 2 ma

Pz = V * i = 10 * 2ma = 20 milli watts

P10k = I^2 * R = 2ma * 2ma * 10k = 40 milliwatts.

Not much power in either.


I made some calculations ,

with 10K + 10K voltage divider the Vgs of Mosfet Gate would go from 10,5- 14,7 (21v-29,4V of the battery) and this would mean I could have the 15V Zener just to stop the unloaded Solar Panel 45V (23,5V after voltage divider) to kill the Mosfet when it connects/switches.

having the work-place till 15V on the GATE on a 20V maximum Vgs Mosfet would be ok, right?

Yes it will be fine.

1 Like

The last schematic won't work.
The whole design is poor and probably dangerous. What about cell balancing? How do you plan to do the CV charging part?

Note that a resistor divider to set the gate voltage isn't robust - due to the
internal capacitances the drain can back-drive the gate during a switching
transient. Basically use a zener between gate and source to protect the
gate oxide, then you won't get any risk of over-voltage on the gate.
MOSFETs are extremely robust if the gate is protected, and pretty vulnerable
if not.

If using fast PWM its normally a requirement that the gate is driven hard,
such as with a MOSFET gate-driver chip, reducing switching losses by keeping
switching events to a few 100ns in duration.

With a 10k resistor pull up on the gate, switching could take as long as 20µs or so,
whereas proper gate drive switches most MOSFETs in under 200ns - gates
are very capacitive.

Did you mean the last as the One proponed by 79galinakorczak????

  • i allready have two ADS1115 for measuring each of the Seven battery bank and was planing to put i each one a mosfet/bjt switch with a resistor to balance each bank

Each balancing group would be trigger with optoswitches from digital pins from Arduino

As for CV (constant voltage?) i was considering an alternative way , as avoiding the linear part of the mosfet and also the “not so happy 18650 with the traditional float charging after bulk “

You mean to put just a 10 or 12v zener from Gate to Source and a 10k resistor from gate to ground (with a BJT as switch) ????

(Will do P and I calculations)

Prove it

You cannot turn off the transistor.