I have been racking my brain so hard on how to control a PMOS with my 5v logic. I need the PMOS to control a voltage range from 8-60 volts, the pmos can handle the voltage and hast a fully on at Vgs = -5v.
the hard part is its -5v meaning the gate voltage must always be atleast 5volts below the input voltage.
I tried making a voltage divider but only works ar higher input voltages the lower you go the difference will be lowered too ?
Any solution to this?
Hi,
You would be best to look for a HIGH SIDE MOSFET drive IC.
If you google
HIGH SIDE MOSFET drive IC.
There are number that you can use, the fact that your supply is so high and variable would suggest a ready purposed IC to do the job.
What is the application?
Tom... 
I am planning on making a power delivery circuit, with half of driven by N - channels and the other half by P channel, so that if the device needs a reference to ground it is available.The usual loads are inductive loads, But for the power of the digital electronics will be connected here, If i dont want to decide to power my digital electronics i can atleast use the P channels back to the Input voltage lines
TomGeorge:
Hi,
You would be best to look for a HIGH SIDE MOSFET drive IC.
If you google
HIGH SIDE MOSFET drive IC.
I did try to search some but what i have found need atleast 10 other components.
I was able to to make a 12-60v using This schematics. My PMOS gate turns on at 3.3v and maxes out at 20v i found this values to be okay to have allowance.
TomGeorge:
Hi,
Ops diagram.
Try putting a 5V zener between the gate and source of the P-MOSFET.
Tom...
Yes there are actually protections in place, i just went ahead and simplified the circuit to make it a bit less cluttered looking.
I would have hoped to push it a little bit lower, Is there any way to like subtract a constant value voltage to any voltage (assuming it does not pass to zero) , like at 10v it will subtract 5v and at the same time at 60v it will still subtract 5v? I dont mind this circuit to a bit complicated since i only need to make 1 and mux it all to the p channels
TomGeorge:
. . .
Try putting a 5V zener between the gate and source of the P-MOSFET.
. . .
Exactly. This is the most flexible solution if you need to accommodate a wide range of voltages.
Attempting to design a potential divider that always delivers an in-spec Vgs can be very difficult to impossible.
6v6gt:
Attempting to design a potential divider that always delivers an in-spec Vgs can be very difficult to impossible.
Diagram #4 on this page delivers a constant Vgs.
Maximum voltage across R2 is 1000/330 times the voltage across R1,
meaning 1000/330 * 4.3 = 13volt on a 5volt Arduino.
Zener D2 is only added for extra security.
Are you sure your (unspecified) P-channel fet is happy with 5volt gate drive.
10-12volt is more common for full saturation.
A common voltage divider, with high resistor values, makes the fet switch slow.
Could be very dangerous for the fet if high currents are involved.
Leo..
PS
Can't of course use a 2N3904 for 70volt.
Can't use a mosfet there either.
TomGeorge:
Try putting a 5V zener between the gate and source of the P-MOSFET.
I was also thinking of a zener but more like a 10-12V one. This as a higher gate voltage lowers the on resistance of the MOSFET.
John41234:
My PMOS gate turns on at 3.3v
It may turn on at that voltage... but likely no more than a little bit. However you may be right, but as you didn't give the actual part number we can't check the data sheet.
Wawa:
Diagram #4 on this page delivers a constant Vgs.Maximum voltage across R2 is 1000/330 times the voltage across R1,
meaning 1000/330 * 4.3 = 13volt on a 5volt Arduino.
Zener D2 is only added for extra security.Are you sure your (unspecified) P-channel fet is happy with 5volt gate drive.
10-12volt is more common for full saturation.A common voltage divider, with high resistor values, makes the fet switch slow.
Could be very dangerous for the fet if high currents are involved.
Leo..PS
Can't of course use a 2N3904 for 70volt.
Can't use a mosfet there either.
This seem to be a unique solution using the zener as a constant voltage drop. yes I am using a logic level specific trench PMOS FDS8935, It even pretty happy with 3.3v.
I am not very good with transistors, but what i seem to be looking for is a npn with a Vceo of 70 ish volts, with a beta of 100 am i right?
but what i dont understand is how exactly are we tricking the zener??
wvmarle:
However you may be right, but as you didn't give the actual part number we can't check the data sheet.
The actual PMOS is the FDS8935 the NMOS on my schematic is just some genereic NMOS.
John41234:
This seem to be a unique solution using the zener as a constant voltage drop. yes I am using a logic level specific trench PMOS FDS8935, It even pretty happy with 3.3v.
The datasheet doesn't agree with that assessment. It really wants more.
]I am not very good with transistors, but what i seem to be looking for is a npn with a Vceo of 70 ish volts
That should be AT LEAST 70V. More is fine. The beta doesn't really matter in this application.
The zener goes between source and gate to limit the voltage seen by the gate. As suggested before, dedicated gate driver may be in order here indeed.
John41234:
This seem to be a unique solution using the zener as a constant voltage drop.
But the zener is not responsible for that.
If there is a constant voltage across the 330 ohm resistor (5volt pin - 0.7volt BE junction = ~4.3volt),
then there flows a constant current through that 330 ohm resistor.
That current comes from the collector (mostly), so the same current flows through that 1k resistor.
Using Ohms law, we can calculate that the drop across the 1k resistor is about 3x of the drop of the 330 ohm resistor.
The zener is only there as backup.
Leo..
wvmarle:
As suggested before, dedicated gate driver may be in order here indeed.
With a dedicated driver i have to add so many complementary components not ideal for my application when i have multiple channels of this( im space limited) , But not only that its cheaper overall.
I am thinking of abandoning the idea of using my building my own PMOS switch, i think i will just use a All - in - one package IC. It would seem that there are available High side switches.
Im currently looking at this boy it even has two channels it would seem. Infineons data are rather confusing though
It has two channels (as is clearly indicated), but I though you wanted to use up to 70V DC, so why are you picking a part that has an operating voltage of no more than 58V?
wvmarle:
It has two channels (as is clearly indicated), but I though you wanted to use up to 70V DC, so why are you picking a part that has an operating voltage of no more than 58V?
Apologies, i made a typo on the first post, i meant 60v, i willing to take 58 volts too but i will take 60 if i ever find another one.
Do you think this way bette than creating my own ? It even already has protections agains inductive loads so i really dont place flybacks
In general the IC is the way to go. Less parts, less that can go wrong, less to worry about.
60V on a 58V rated part will most likely work, but it's not a good idea. The data sheet says in big letters "suitable for 42V", and at another point I see a VIN rating of 42V. This 58V appears to be the absolute maximum rating for the part.
Note that the over voltage protection will kick in at around 63V, but maybe already at 58.5V. That is a major issue if you want to run it at 60V.
