Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Switchable constant current source

Status
Not open for further replies.

atferrari

Well-Known Member
Most Helpful Member

Attachments

  • Switchable MOSFET_Current_Regulator_web.png
    Switchable MOSFET_Current_Regulator_web.png
    5.4 KB · Views: 1,953
The gate really is not on well at 5V G-S. Then the source is at +0.6V so the gate could not get above 4.4V. I think you need a "logic level" MOSFET.
 
Problem is to get one of these locally...:(
 
Here is how I would do it. Note that the emitter follower is a constant current source and switch in one. This will work provided the forward voltage drop of your Green LED is 3.2V or less. You likely have a 2n3904 or an equivalent...

DF53a.jpg
 
Before I buy components, given the circuit (attached) could I go away if Vcc is just 5V, replacing the MOSFET with a 2N7000?
You can just use another NPN BJT transistor in place of the MOSFET; then you don't have to worry so much about the Vcc. R1 will have to be decreased appropriately.
 
Here is how I would do it. Note that the emitter follower is a constant current source and switch in one. This will work provided the forward voltage drop of your Green LED is 3.2V or less. You likely have a 2n3904 or an equivalent...

View attachment 83642
Mike, I like it......but.
The input voltage from the drive source needs to be very consistent.
A "0" is anything less than 0.6 volts. GOOD.
A "1" might be 3.3V or 5V or 4.5 today and 3.8 the next day. This will cause the current to change. If the signal comes form the week pullup of a micro that might be a problem. If it is from a CMOS gate living on a regulated 5 volt supply then it will work well.
ALSO
All these ideas depend on the base-emitter voltage to remain constant. It changes with temperature.

I agree the 2N3904 or 2N2222A etc will solve the "gate turn on voltage" problem.
 
Mike, I like it......but.
... If it is from a CMOS gate living on a regulated 5 volt supply then it will work well.
ALSO
All these ideas depend on the base-emitter voltage to remain constant. It changes with temperature....
Atferrari implied the logic levels are indeed 0 and 5V in his first post. I was banking on it being a Cmos output like a Pic pad. Lightly loaded (base current is only a few hundred uA, that will put out close to 5V.

The Vbe will not vary enough to make much of a difference compared to the Ve of 1.6V. Here is the LED current at -20 to +40degC in steps of 20deg. -20 is green....

DF53b.jpg
 
Last edited:
How did you calculate 15mA to 20mA at the output in your circuit with the high power Mosfet?
The base-emitter voltage of the 2N3904 at a low current is 0.6V at room temperature. Then the constant current at the output is 0.6V/0.82 ohms= 732mA and a little LED will blow up!
 
Hola AG

Sure this is for higher current. I did not calculated actual values for 15 to 20 mA yet.

My concern is to define a valid MOSFET (for my smaller current) that I could obtain locally.
 
Atferrari, are you saying that you want a LED current of much higher than the 15mA you specified in your original post????
 
and constant current is around 15 to 20 mA

I was thinking of implementing that topology but had not calculated actual values for 15 to 20 mA yet.
 
I was thinking of implementing that topology but had not calculated actual values for 15 to 20 mA yet.
What is the Green LED used for? Is it just an indicator? If so, why do you need to PWM it?
 
You could buy a few thousand 2N7000 Mosfets and test each one. You might be lucky like NorthGuy and find one that works with a gate voltage very low.
 
Hi,

There has been some talk of using only bipolars rather than a MOSFET and that might be fine too, but if you still want to use a MOSFET then that 1M resistor is probably far too large, if that schematic is still valid. 1 Megohm is too large to charge up the gate of the MOSFET if the frequency of the PWM is higher, but even with lower frequency PWM i dont think you want that high of a gate pullup because it will take too long to turn the MOSFET on. Maybe at the lower level current it will be ok, but it would probably be better to decrease that resistance significantly. If this is battery operated you may want to think about this more carefully but if not a lower value would be better in most cases. Remember that the equivalent gate capacitance has to charge up before the MOSFET will actually start to turn on, so whatever value resistance is in series with that capacitance determines how fast the MOSFET can turn on, and this is always something to consider with at least some care.
 
Como Estas.
I went to Spanish night school to learn Spanish. The first evening everybody was talking in Spanish except me. They already knew Spanish. They went there to meet Spanish-speaking people. I never went there again.
 
You could buy a few thousand 2N7000 Mosfets and test each one. You might be lucky like NorthGuy and find one that works with a gate voltage very low.

My data sheet says max. value for Vgs(th) is 3V, and typical is 2.1V, so all of them are guaranteed to work with 3.3V. They may not give you the full current, but I'm yet to find one that couldn't fully light a LED.

Atferrari has 4.5V. According to the graph from the datasheet, this should be Ok for any allowed current.
 
My data sheet says max. value for Vgs(th) is 3V, and typical is 2.1V, so all of them are guaranteed to work with 3.3V. They may not give you the full current, but I'm yet to find one that couldn't fully light a LED.
The threshold voltage is when it is almost turned off with a current of only 0.25mA. A VERY dim LED.
You need a gate voltage of 10V to guarantee it will turn on with a current of 500mA and some will have a 2.5V drop.

According to the graph from the datasheet, this should be Ok for any allowed current.
The graph is for a "typical' device. You cannot buy a typical one, you get whatever they have. Maybe they had a bad yield.
 
The threshold voltage is when it is almost turned off with a current of only 0.25mA. A VERY dim LED.
You need a gate voltage of 10V to guarantee it will turn on with a current of 500mA and some will have a 2.5V drop.

The graph is for a "typical' device. You cannot buy a typical one, you get whatever they have. Maybe they had a bad yield.

The graph is certainly for a "typical" device, but that's the only graph that we have. So, we have to assume that the "worst" device will not be too far off from it. The graph certainly does not guarantee anything, but it lets you make reasonable assumptions. I have a datasheet from Fairchild and from STMicroelectroncs, they both show typical current for 4.5V Vgs at about 750mA. That's definitely not far off from 500mA, so you wouldn't use it for 500mA with Vgs of 4.5V, but 100mA is very reasonable, and 15-20mA which we need here is absolutely acceptable.

I agree about the the 2.5V drop. It wouldn't be a good idea to use them for 500mA anyway. However, at 20mA, even 10 Ohm Rds will produce only 0.15-0.2V drop.

May be I don't understand something. With R2 at 0.82 Ohm, looks the circuit is aimed around 750mA continuous current (certainly too much for 2N7000). But atferrary says it's 15-20mA?
 
Como Estas.
I went to Spanish night school to learn Spanish. The first evening everybody was talking in Spanish except me. They already knew Spanish. They went there to meet Spanish-speaking people. I never went there again.

Muy bien. Gracias.

Not fair from their side but I suspect that you did not try harder because your lady already speaks it, AG! ;) :p
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top