# basic constant current source

#### dlatch

Joined May 15, 2016
85
In the circuit below, what is the purpose of R1? The circuit works with a wide variety of resistors at that point. If current is limited by the transistor why is it needed? And how do we find it's more optimum value

I'm want 1mA with a load varying from 100 ohms to about 14K ohms.

thanks all

#### Attachments

• 91.3 KB Views: 34

#### dl324

Joined Mar 30, 2015
12,677
Circuit in question for the convenience of others:

R1 sets the current. For larger currents, a single diode may not be sufficient.

Won't work for a 14kΩ load.

#### crutschow

Joined Mar 14, 2008
26,732
In the circuit below, what is the purpose of R1? The circuit works with a wide variety of resistors at that point. If current is limited by the transistor why is it needed? And how do we find it's more optimum value
That circuit will not work as shown.
It needs two diodes in series between the supply and the transistor base.
One diode cancels the transistor base-emitter voltage.
The other is then the voltage reference for determining the constant current.

The resistor is what determines the constant current value.
The transistor is what regulates the current to the value established by the resistor.

The current value is approximately one diode voltage drop divided by the resistance (0.6 / R1).
When the voltage across R1 reaches ≈0.6V, the diodes starts to conduct.
This prevents the transistor base current from increasing any further so the transistor current becomes fixed at that point.

Note that you obviously can't go to a 14KΩ load with a 13.8V supply, since it requires 14V to force 1mA through a 14kΩ resistance.

Below is the LTspice simulation of the two diode circuit.
As you can see, you need a supply voltage of about 15.5V to generate 1mA into 14kΩ, to allow for about a 1.5V intrinsic circuit operating voltage.

If you want tighter regulation of the current with load change then shown, than you would need a more complex circuit, such as one with an op amp to sense the voltage across R1.

#### Bordodynov

Joined May 20, 2015
2,801
See

#### dlatch

Joined May 15, 2016
85
Very nice gentleman. Suppose I want the current to reflect changes in vcc? Just a pot on the base? I've tried that but it seems fussy. Get me in the ballpark with around 13.5 vcc and apx 1mA.

#### WBahn

Joined Mar 31, 2012
26,398
How do you want changes in Vcc to be reflected in the current? We can't tell you how to achieve a certain behavior unless you adequately describe the behavior you are trying to achieve.

The primary goal of your starting topology is to make the current NOT depend on Vcc -- the goal is to make it CONSTANT. So if you don't want it to be constant, then you probably need a very different topology.

#### crutschow

Joined Mar 14, 2008
26,732
It's confusing when you say you want a constant current that's not constant.

#### dlatch

Joined May 15, 2016
85
It's confusing when you say you want a constant current that's not constant.
This is the hall wiper slotcar controller I have posted previously. As built, the hall wiper current is not just constant but regulated (LM 234). So it achieves Max too soon on a lesser voltage track and does not go high enough on a higher voltage track. So I want the current source to be constant as the resistor ladder sweeps. But the overall maximum voltage signal needs to reflect the given track voltage more or less.

So it's constant through a changing load at a given voltage but not regulated.

#### crutschow

Joined Mar 14, 2008
26,732
Then you need to state exactly what currents you want for what voltages.

#### dlatch

Joined May 15, 2016
85
Apologies for not having a proper image file of the schematic and the schematic is not the latest version of the circuit but, it's all I have right now.

The supplied Voltages can range from 12 to 14 volts. My resistor string is apx 10K.

The wiper voltage feeds the op amp; that creates an output voltage that can match the supplied voltage. "Ideally" I want that output voltage (supply match) to land at the end of the wiper. With regulated current, it comes at the same point regardless of the supplied voltage.

I built it with a 12 volt supply. Took it to a 14 volt track and found it only went up to 12 volts output. OK....just learning. (turned up the current and it was fine) However, now, on a 12V track it reaches 12 volts too soon on the wiper and is perceived as too quick

So I guess I want the wiper voltage to be linearly related to the supply voltage.

I could make the current easily adjustable so it could be adjusted per track since having it fully regulated has some benefits but I am experimenting.

#### Attachments

• 318.8 KB Views: 11

#### dlatch

Joined May 15, 2016
85
Your explanation of the dual diode ccs is perfect. But how do we understand a single diode ccs or a ccs with no diode?

I want the current constant as the load changes but not regulated independent of the supply voltage.

Why:

This CCS feeds a varying resistor producing varying voltage. The varying voltage needs to be higher if the supply is higher and vice versa. The current is constant so the voltage is linear as the resistor changes.

I guess I need a voltage law analysis of the loop around a single diode set up to understand how to optimize the component values. I want the voltage to return as close to the supply as possible while maintaining linearity.

#### WBahn

Joined Mar 31, 2012
26,398
You need to describe HOW you want the current to be a function of the supply voltage. What is the relationship you want? What you are asking for is probably quite doable with a pretty simple circuit, but before you can solve a problem, you have to be able to clearly describe exactly what the problem is that you are trying to solve.

What do you mean "and vice versa" -- you want the supply voltage to somehow go up if the varying voltage goes up?

#### crutschow

Joined Mar 14, 2008
26,732
It sound like you want an adjustable current that is constant with load but varies with the supply voltage, is that correct?

#### dl324

Joined Mar 30, 2015
12,677
This CCS feeds a varying resistor producing varying voltage. The varying voltage needs to be higher if the supply is higher and vice versa.
Sounds like you just need to replace the diode(s) with a resistor. Just make the divider stiff enough to not be affected by varying emitter current.

#### dlatch

Joined May 15, 2016
85
Thanks crutshow, WBahn and dl234 I thought I lost you with too many words at post 10.

What I am asking for is perhaps so basic it's hard to grasp why anyone would want this. The information I find on CCS is ALL about regulation.

I have experimented with only a voltage divider on the base side and it seems to saturate too easily. Should I persist, or do I need an emitter resistor and how would you calculate it?

Bahn....HOW do I want the current to relate to the supply? I guess linear: If the supply is higher by 1 volt say up from 12 then I would want the current to be 1 1/12th mA? I assume an unregulated set up would be roughly linear.

I understand basic BJT...where collector current is base current times hfe. Is that really all there is to it?

#### WBahn

Joined Mar 31, 2012
26,398
Bahn....HOW do I want the current to relate to the supply? I guess linear: If the supply is higher by 1 volt say up from 12 then I would want the current to be 1 1/12th mA? I assume an unregulated set up would be roughly linear.
So, reading between the lines, it sounds like you want the current to be 1/12 mA/V. Is that correct?

Is there a reason you can't use an opamp?

#### crutschow

Joined Mar 14, 2008
26,732
Is this sort of what you want?
The current is constant for a change in load resistance but varies with the power supply (shown for power supply voltages of 12V and 14v).
The current can be adjusted by pot U2.

#### ebeowulf17

Joined Aug 12, 2014
3,275
Is this sort of what you want?
The current is constant for a change in load resistance but varies with the power supply (shown for power supply voltages of 12V and 14v).
The current can be adjusted by pot U2.

View attachment 130382
I've been trying, with little success, to follow this thread, and this is the first thing that's really made sense to me. I like the circuit - it's the first time I've seen this basic setup.

I know this is a tangent off topic, but is something like this at the heart of current loop (4-20mA analog signals and such) sensor outputs?

#### crutschow

Joined Mar 14, 2008
26,732
but is something like this at the heart of current loop (4-20mA analog signals and such) sensor outputs?
Yes, they use some type of a constant-current circuit to generate the 4-20mA signal.