Can you help me with this circuit. I want to realize that when the power supply is energized, the external load coil can be activated for about 0.2s before it is disconnected;
The problem encountered now is:
22V-24V when the power supply is on, is that the circuit is working fine.
25-30V when supplying power, the PMOS is broken as soon as it is energized, and measurements show that the DS is energized.
If I change the external load from 5 ohm to 20 ohm, the circuit works fine even at 22V-30V.
It looks like it's too much current.
I tried to add 100nf capacitor between GS but no effect.
Can you help me analyze how to deal with this, really appreciate it.
220uF and 1Meg would give an RC constant of around 0.2 seconds. You could start with that.
However, if you allow the cap to charge gradually (as caps do), and apply this voltage directly to the PMOS's gate, you'll drive the MOSFET into its linear region much of the time. I.e. it'll dissipate energy. Since it's only 200ms it may survive if you pick a big TO220 part - not the SOP8 SMD part you've got now. So not very surprising it goes into smoke-generation mode.
Better use something like a Schmitt-trigger to apply a clean signal to the MOSFET so it switches hard & fast.
Or, since this is an Arduino forum anyway, use a second small-signal NMOS, use that to drive the PMOS, and drive the small signal NMOS from an Arduino pin. If your project involves a microcontroller anyway, might as well use it, too.
There are roughly 1 million ways to skin this cat and everyone will have their own preference, but I think this is reasonably intuitive/easy to understand.
I did not verify if the BSS123 will match your criteria, and the PMOS is really a pretty random one I picked in LTSpice.
You can dimension the resistors according to your preferences. You could even use NPN BJT's if you prefer those over NMOS.
Thank you so much for such an enthusiastic and quick reply. Thank you.
The PMOS had to be used because the circuit has other power supplies that need to be grounded to this one.
I have a question, Vgs should be <= +- 20V, but when the power supply is turned on, where the capacitor is charging, Vgs is greater than 20V between them. Why doesn't it burn out at this time?
Also I like to use SOP8 SMD for small size. Maybe this one is not suitable;
I've re-updated the diagram below, can you help me see if this is ok?Thank you again. You're too kind.
It will burn out because the Vgs goes above 20V.
Trying to do timing with Rs and Cs on a MOSFET gate is quite problematic.
I think you need to use a simple timing circuit maybe using a 393 comparator or a schmitt trigger gate
That's fine in principle, as long as you don't drive it into its linear region. Your new version suffers from the same problem, so any SMD part will still fry - or at least heat up momentarily.
Like @jim-p says, use a comparator, a Schmitt trigger - or even a 555 comes to mind. The alternate version with the small-signal NMOS's will also work because only the first small-signal MOSFET is driven into its linear region and since it doesn't have to pass much current, it'll do just fine.
Oh yeah:
You've switched to an IRF9540. That's a TO220 part, which has a much bigger thermal mass. Apparently it survives the momentary abuse. Until of course the day it won't any longer...
PPS:
That, too. Check the specs on your PMOS:
You start it up with gate at 0 and source at VIN. If that's 30V...oops!
I see several problems. By turning on the MOSFET that slow leaves it in the linear reason where it will dissipate a lot of power. You validated that when you stated
The voltage drop across the MOSFET times the current is watts it needs to get rid of. Also mentioned earlier you need to protect the gate/source from exceeding its max voltage. I found this schematic on line but it also has some problems. R4 should be on the 3.3V side of R5, so it does not divide the voltage. R4 is to keep the device off when reset etc is active. when reaching about 12V. You could put put a zener in the collect lead to prevent the gate/source from being over voltaged.
My load current (5Ω load) is too high and the MOSFET burns out. As soon as my load current is lowered (20Ω), the MOSFET works fine with a 22V-30V supply; this looks like it has something to do with Cgs. So I don't really understand. I'll try to drive the gate first with a small current NMOS as you said. Let's see if I can improve it, I'd like to do it with a very lean circuit. As soon as I hear something, I'll reply back.
If timing with Rs and Cs on the MOSFET gates is problematic, why does it work properly when the external load is only 20 ohms? I would like to use a simple circuit for this;
R1,R2 ,R3 have made the voltage of Vgs less than 20V but still burned;
I shorted R3 and Vgs is greater than 20V, yet the MOSFET works fine with an EXTERNAL load of 20Ω. thanks ，I'm still going to try the solution you mentioned though;
The load determines how much current the MOSFET needs to pass. Because you're driving the MOSFET into its Ohmic region, dissipation is the square of the current. Compare the cyan peak to the magenta peak; cyan is dissipation when using a 5R load, magenta when using a 20R load.
And of course, also ensure that you don't violate the MOSFETs ratings as pointed out by others.
I think you have already shown that it's poblematic. It works for some loads not others, it works for some voltages not others. Later you will find it might work for some MOSFETs and not others, even if they have the same part number.
hi rsmls , also thanks jremington and jim-p gilishultz and other friends.
It's really a big family that makes me feel warm and fuzzy. You are all so friendly and welcoming ,
I tried your method(rsmls) today. It tested down between 22-30V without any problems. Thank you very much. This could be the excessive power consumption caused by the slow rise in voltage due to the parallel delay capacitor on the G terminal in my circuit. below is your diagram
my test validation :
I'm wondering again if there is an easier way to achieve this high current delayed disconnect. Or can I make the delay time not fluctuate due to external voltage fluctuations and be able to discharge quickly when disconnected.
