linear mode 10A MOSFET current source

[shyon]

Hi,

I have seen many P-mosfet based linear current sources but these are for small current more over the main dissipating components is mosfet not resistors.

But now
I am looking for p-mosfet based 10A linear current source circuit in which major dissipation component would be resistor.

regards

 

[(*steve*)]

You might want to consider why resistors are not used in this application.

You could investigate a switchmode solution to your problem.

 

[KrisBlueNZ]

I have seen many P-mosfet based linear current sources but these are for small current more over the main dissipating components is mosfet not resistors.
But now I am looking for p-mosfet based 10A linear current source circuit in which major dissipation component would be resistor.

Hi and welcome to the forums.

Your post is very short on information. For a start, how about answering these questions.

1. WHY do you want the main dissipating component to be a resistor?

2. Do you just want to reduce the power dissipation in the MOSFET?

3. Why do you want to use a P-channel MOSFET instead of an N-channel one?

4. What is the input voltage to the current source?

5. What is the output voltage range that the current source needs to be able to operate over?

6. What other constraints are present? Is heat dissipation a concern? Size? Cost?

7. What is the PURPOSE of the project? What are you going to be feeding this current into?

 

[shyon]

Thanks kirisBueNZ

1- To reduce the power dissipation in the MOSFET in this way i want to use low power rating MOSFET
2-yes
3-because i want to use load on lower end means load will be directly connected to ground
4- that would be 24volt lead acid battery
5-10 to 20V
6-heat dissipation is not a major concern because i want to use this current regulator to drive high current flashing LED, that led will flash for 200ms per second

 

[(*steve*)]

i want to use this current regulator to drive high current flashing LED, that led will flash for 200ms per second

You want a switchmode LED driver that can be enabled/disabled by a logic input or somesuch.

 

[KrisBlueNZ]

Can you post the part number for the LED? And preferably a link to a data sheet for it?

 

[shyon]

i bought these LED from scrap market so i do't have their part #

and do't want to use switch mode current source because it is expensive and complex.
I just want simple generic type 10A current source, which could also be used on any other application, like, to drive motor (just an example) etc

 

[KrisBlueNZ]

OK, if you don't have a part number, at least tell us the specifications. Also, a photo of it, and any markings on it, would be very helpful.

A "simple generic 10A current source" doesn't really exist. At that level of current, there are extra factors you need to consider in the design - heat dissipation, and other losses.

I could design a linear current source that will provide 10A from a 24VDC source with an output compliance range of 0V~22V but do you really want fan-forced cooling and very expensive components?

I think Steve is right; you may need to use a switching converter.

You really need to tell us a LOT more about your application.

 

[shyon]

Fine,

I want to flash 120W super bright LED for 200ms and that will flash once in a second. So, i think, i do't need continuous 10A current source, means if a linear regulator will provide current for 200ms then it would be fine.

2nd my compliance range is from 10V to 20V

My LED is like the figure i attached. Please note it is not the picture of my LED, i just google it, since it is similar to my LED so i attached it.

As Steve suggested switching converter. In my opinion, a linear regulator can full fill because its on time will be only 200ms, moreover a large heat sink will dissipate the accumulated heat.

If you help me in designing the 10A source, i will be grateful to you.

Note: i do't know how to attach file thats why i am pasting here link

 

[(*steve*)]

It's a poor idea, but here's a circuit that will work.

The mosfet could be one of these.

Be aware that the average dissipation will be around 40W at 200ms on per second.

You would need a 2degC/W heatsink at this duty cycle.

The 0,06 ohm resistor would dissipate 6W peak. I would recommend you use 16 or 17 1 ohm 1/2W resistors in parallel. Add or remove resistors to tune the current (more resistors = higher current)

The 10V zener ensures that you don't kill the mosfet with too hgh an input voltage.

The input (the left hand side of the 1K resistor should be taken to 10V to turn the constant current source on. Grounded it will turn it off.

Leave it open and the LED will stay lit for some time, the mosfet will overheat, go short circuit, the LED will get very bright and the resistors will smoke. It will be a race to see whether the resistors burn up before the LED dies. A 100k resistor in parallel with the zener will allow the mosfet to turn gracefully off in this situation.

 

[KrisBlueNZ]

Steve's design should do what you want. But your statement that the circuit needs to operate with an output voltage range of 10~20V makes me wonder whether you know anything about this LED at all.

You know that if you feed 10A into an LED whose forward voltage is 20V, it will dissipate 240W, not 120W? And you know what will happen to it then, don't you?

You can convert Steve's design to use a P-channel MOSFET if you want, by turning the whole circuit upside-down and reversing the direction of all the arrows in the circuit. The drive for the MOSFET gate needs to be relative to its source, which would now be at +24V. The P-channel MOSFET needs a gate drive signal that is equal to the +24V rail (which is the MOSFET's source) to turn the MOSFET OFF, and about 12V lower than the +24V rail to turn the MOSFET ON.

I definitely recommend adding a resistor across the zener diode to give the circuit a defined behaviour when the input is not connected to anything.

Power dissipation in the MOSFET will be significant. If it's dropping 14V at 10A for 20% of the time, its average dissipation will be 28W. That equates to a fair amount of heatsinking.