MOSFET current limiter

Thread Starter

Clearlycountry

Joined Jan 17, 2016
2
I have a very basic current limit circuit configured with an IRF4905 and a fixed -VGS, intended to limit charge current into a 12VDC battery. The idea was to use the least possible parts to accomplish the task, so not PWM, no Rsense/bipolar feedback, etc. By allowing the MOSFET to transition from saturation and drop through ohmic as the difference between supply and battery voltage equalize, current would be max initially, then drop to some minimum value.

Simulations in LTSpice show expected results, but on the bench the current tends to climb above the limit set by VGS. If an additional current load is placed in parallel with the battery to test current limiting, the current through the MOSFET will also increase. I've confirmed there is no drift in the VGS voltage, no parts are overheating, etc. Can anyone offer suggestions on what might be going on? Thanks.
 

ronv

Joined Nov 12, 2008
3,770
I have a very basic current limit circuit configured with an IRF4905 and a fixed -VGS, intended to limit charge current into a 12VDC battery. The idea was to use the least possible parts to accomplish the task, so not PWM, no Rsense/bipolar feedback, etc. By allowing the MOSFET to transition from saturation and drop through ohmic as the difference between supply and battery voltage equalize, current would be max initially, then drop to some minimum value.

Simulations in LTSpice show expected results, but on the bench the current tends to climb above the limit set by VGS. If an additional current load is placed in parallel with the battery to test current limiting, the current through the MOSFET will also increase. I've confirmed there is no drift in the VGS voltage, no parts are overheating, etc. Can anyone offer suggestions on what might be going on? Thanks.
Can you post a schematic and/or you simulation file?
 

crutschow

Joined Mar 14, 2008
25,432
The reason such simple circuits are not often used is because of their deficiencies, such as you found.
The Vgs threshold voltage is quite temperature sensitive.
Typically it's -2mV to -4mV per degree C change.
Thus as the MOSFET heats up due to the current, the current limit value will increase for a given Vgs.
The simulation does not include self-heating effects so that's why you didn't see the problem there.

You will likely need to go to a more complex circuit.
A simple one uses two transistor and two resistors as below.
T1 is a power transistor, such as a 2N3055.
T2 is small transistor, such as a 2N2222
The limit is approximately 0.7V / Rsense.

upload_2017-2-28_10-9-47.png
 

Thread Starter

Clearlycountry

Joined Jan 17, 2016
2
The reason such simple circuits are not often used is because of their deficiencies, such as you found.
The Vgs threshold voltage is quite temperature sensitive.
Typically it's -2mV to -4mV per degree C change.
Thus as the MOSFET heats up due to the current, the current limit value will increase for a given Vgs.
The simulation does not include self-heating effects so that's why you didn't see the problem there.

You will likely need to go to a more complex circuit.
A simple one uses two transistor and two resistors as below.
T1 is a power transistor, such as a 2N3055.
T2 is small transistor, such as a 2N2222
The limit is approximately 0.7V / Rsense.

View attachment 121505
Thank you, I really appreciate you quick response to my dilemma. In answer to your reply and the one prior, without the frills, the circuit is really nothing more than a zener across a divider network to establish a constant Vgs, sourcing current to a 12V battery at various states of charge from 12-12.8V, supply is 14V.

Paying the price for simplicity... The reason I went without any feedback was strictly due to space restrictions and no real requirement for an accurate current limit. +/- 500mA is more than adequate for this design. Perhaps before giving up on this circuit completely, the next option is to test a different MOSFET with better output characteristics. The IRF4905 is pretty wild to the upside wrt increasing Vgs, and something more like the IRF9Z34 has much flatter Vgs/Id response. The 4905 has a great rds(on). If that doesn't do it, then unless anyone has a suggestion, it's feedback time I guess...
 

ian field

Joined Oct 27, 2012
6,539
The reason such simple circuits are not often used is because of their deficiencies, such as you found.
The Vgs threshold voltage is quite temperature sensitive.
Typically it's -2mV to -4mV per degree C change.
Thus as the MOSFET heats up due to the current, the current limit value will increase for a given Vgs.
The simulation does not include self-heating effects so that's why you didn't see the problem there.

You will likely need to go to a more complex circuit.
A simple one uses two transistor and two resistors as below.
T1 is a power transistor, such as a 2N3055.
T2 is small transistor, such as a 2N2222
The limit is approximately 0.7V / Rsense.

View attachment 121505
The circuit you posted would serve the TS with a MOSFET in T1 position, but the MOSFET will not be able to conduct below its VGSthr unless you supply R1 with a higher auxiliary supply voltage.

I was going to suggest that circuit - but you already did.
 
Top