Issue with current sources for a GCSR laser

Thread Starter

Davo SG

Joined Jul 1, 2018
4
Currents table.png current.png Hi,

I have to do some current sources controlled with the DAC of a RedPitaya that goes from 0 to 1.8V so I can tune a GCSR laser. The specs of the laser are in the table.

Currents table.png

I started to design them using the application note of Analog Devices: http://www.analog.com/en/analog-dialogue/articles/diff-amp-heart-of-precision-current-source.html

The problem that I encountered is that the load resistance is around 1KΩ, so the current decreases significantly. I simulated the circuit in MultiSim, and I attached the simulation.
current.png

Is there any other topology for making the output current independent of the load resistance?
 
Last edited by a moderator:

Hymie

Joined Mar 30, 2018
861
Hi,

I have to do some current sources controlled with the DAC of a RedPitaya that goes from 0 to 1.8V so I can tune a GCSR laser. The specs of the laser are in the table.

View attachment 155427

I started to design them using the application note of Analog Devices: http://www.analog.com/en/analog-dialogue/articles/diff-amp-heart-of-precision-current-source.html

The problem that I encountered is that the load resistance is around 1KΩ, so the current decreases significantly. I simulated the circuit in MultiSim, and I attached the simulation.
View attachment 155428

Is there any other topology for making the output current independent of the load resistance?
It appears to me that your circuit is applying 35V across a 1k ohm resistor – with the resultant current being 35mA – Ohm’s law appears to be working well in your circuit.
 

ebeowulf17

Joined Aug 12, 2014
3,274
Hi,

I have to do some current sources controlled with the DAC of a RedPitaya that goes from 0 to 1.8V so I can tune a GCSR laser. The specs of the laser are in the table.

View attachment 155427

I started to design them using the application note of Analog Devices: http://www.analog.com/en/analog-dialogue/articles/diff-amp-heart-of-precision-current-source.html

The problem that I encountered is that the load resistance is around 1KΩ, so the current decreases significantly. I simulated the circuit in MultiSim, and I attached the simulation.
View attachment 155428

Is there any other topology for making the output current independent of the load resistance?
I know essentially nothing about lasers, but l wouldn't have thought they were much like 1k resistors!

Are you trying to provide a current source to the laser diode itself? Are you trying to send a control signal to a complete laser system which accepts external controls? Are you trying to limit total current to an otherwise self-contained laser system?

In any of these cases, I wouldn't have expected a 1k resistor to be a good simulation of the load.

Maybe if you provide a wiring diagram of what you're up to and describe what exactly this voltage controlled current source needs to be capable of, someone here can help a little more.
 

Thread Starter

Davo SG

Joined Jul 1, 2018
4
Hi,

Thank you for the responses. I will explain more about the project. GCSR laser is a widely tunable laser module whose output signal can be tuned 4THz (192-196THz). The laser has 4 sections: gain, coupler, reflector and phase. The emission at a tuned wavelength is achieved by supplying the appropiate combination of currents to the four sections. This currents will be controlled by the DAC of a Red Pitaya, that can deliver from 0 to 1.8V.

Scheme.png
Here is the pin configuration of the laser:

LASER PIN CONFIG.jpg

I measured the resistance that each section has (from input of section to chip and package ground) and is approximately 1KΩ. I tried to use difference amplifiers but the problem is that the output current is constant, but when the load resistance is getting higher, it drops.

Current vs Load.png

I've seen that for having a constant output current it has to be achieved : Io * Rload ≤ Vsat/2 . So the difference amplifier is not suitable for the application. I want to know is there another topology for a high impedance voltage controlled current source.
 

ebeowulf17

Joined Aug 12, 2014
3,274
Hi,

Thank you for the responses. I will explain more about the project. GCSR laser is a widely tunable laser module whose output signal can be tuned 4THz (192-196THz). The laser has 4 sections: gain, coupler, reflector and phase. The emission at a tuned wavelength is achieved by supplying the appropiate combination of currents to the four sections. This currents will be controlled by the DAC of a Red Pitaya, that can deliver from 0 to 1.8V.

View attachment 155453
Here is the pin configuration of the laser:

View attachment 155455

I measured the resistance that each section has (from input of section to chip and package ground) and is approximately 1KΩ. I tried to use difference amplifiers but the problem is that the output current is constant, but when the load resistance is getting higher, it drops.

View attachment 155456

I've seen that for having a constant output current it has to be achieved : Io * Rload ≤ Vsat/2 . So the difference amplifier is not suitable for the application. I want to know is there another topology for a high impedance voltage controlled current source.
Cool, thanks for the added details.

Regarding your current source capabilities, any current source circuit will be limited by its supply voltage. If you want 160mA into a 1k load, ohms law tells you:

E = I x R
(voltage = current x resistance)
voltage = 160mA x 1k
voltage = 0.16 x 1000
voltage = 160V

So, to deliver your 160mA into 1k, you need to put out 160V! That would mean slightly higher supply voltage, because you lose a little voltage across a sense resistor, and a little more across the junctions of your output transistor.

Personally, I'm a bit skeptical about that 1k input impedance. Measuring any aspect of an IC can be wildly misleading because of the non linear behavior of diodes and transistors. Just to illustrate the point, here's an amplifier input stage from an Analog app note:
EDC2102A-FD85-4E1E-88B8-8BE8301365E6.jpeg
If you tried to measure resistance from Vin to ground on this circuit with a typical multimeter, you'd probably simply read the resistance of R2, because the voltage a meter uses for testing is typically too low to overcome the forward voltage of Q1's base-emitter junction, so test current only flows through R2.

However, in a real world application of this amp, where the input voltage is kept above Q1's Vbe threshold, the input impedance of the amp would be dependant primarily on Rb1.

I'm oversimplifying quite a bit here, and I'm not suggesting that your laser IC's input stage looks exactly like the circuit above - I'm just trying to demonstrate how misleading multimeter readings on ICs (or components in circuit with other components) can be.

It may well be the case that your IC input has a 1k pull down resistor internally, but that the signal in normal use passes through a diode or BJT junction with relatively low input impedance beyond that junction...

In other words, you may not need 160V to get 160mA, but we can't know without a more definitive understanding of the IC's input characteristics. A multimeter reading isn't trustworthy here. What does the datasheet say about the inputs? Can you link to the full datasheet? If not, maybe upload all pages related to input electrical characteristics.

Cheers!
 

MrChips

Joined Oct 2, 2009
21,911
Ohm's Law still applies.

I = V / R

or

V = I x R

R = V / I

Since I is constant, Rmax = Vsupply / I

Once the load resistance exceeds Rmax, the supply voltage Vsupply is too low to maintain the desired constant current. You need to increase the supply voltage.
 

ebeowulf17

Joined Aug 12, 2014
3,274
Ohm's Law still applies.

I = V / R

or

V = I x R

R = V / I

Since I is constant, Rmax = Vsupply / I

Once the load resistance exceeds Rmax, the supply voltage Vsupply is too low to maintain the desired constant current. You need to increase the supply voltage.
I'm not sure we know yet what the load resistance is. I could be wrong, but I don't trust that multimeter measurement.
 

Thread Starter

Davo SG

Joined Jul 1, 2018
4
Hi,

I attached the datasheet of the laser.

I can't find any more data about this laser because as you can see is kind of old and the company that produced it doesn't exist anymore.

Maybe you're right, the reading of the multimeter could be wrong because the laser is a semiconductor, but it is weird that still doesn't receive the programmed current.
 

Attachments

Hymie

Joined Mar 30, 2018
861
Hi,

I attached the datasheet of the laser.

I can't find any more data about this laser because as you can see is kind of old and the company that produced it doesn't exist anymore.

Maybe you're right, the reading of the multimeter could be wrong because the laser is a semiconductor, but it is weird that still doesn't receive the programmed current.
It would have been good if they had supplied a typical application (on the data sheet), showing voltage/current connections at each of the device’s pins.

I suggest you post your complete circuit, with the connections to this device and let people critique it.
 

Thread Starter

Davo SG

Joined Jul 1, 2018
4
Here is the full schematic of the circuit.

At the left I have a voltage regulator with soft start, next are the current sources, then the laser, and at the right I have a PID controller so I can control the temperature of the device.
 

Attachments

ebeowulf17

Joined Aug 12, 2014
3,274
Hi,

I attached the datasheet of the laser.

I can't find any more data about this laser because as you can see is kind of old and the company that produced it doesn't exist anymore.

Maybe you're right, the reading of the multimeter could be wrong because the laser is a semiconductor, but it is weird that still doesn't receive the programmed current.
I misunderstood your original post and didn't realize you had already built and tested this - I thought it was simulation only so far.

Did the measured values match the sim values fairly well (around 35V output and 35mA?)

Have you tried lower settings? If you set the DAC output for a target current of 20 or 30mA, does the circuit deliver that? If so, at least you know your current source is working. Assuming you can get lower current values set, you could measure the output voltages for each current setting and see if the 1k laser input impedance stays consistent (20V = 20mA, 30V = 30mA, etc.)

If that all works and you do have a mostly stable 1k input impedance, then you just need a much higher supply voltage. You can't beat ohms law!

If you're not able to test the current source circuit effectively with the laser chip, you could back up a step and just test the current source into a lower value resistor. If you tested with a resistor around 170-180 ohms, then 30V would be more than enough voltage to deliver your 160mA current. If you find that the current source works fine into these lower value resistors, then you know the source is fundamentally ok but can't deliver enough voltage for the laser. Bear in mind that the resistor in this scenario will need to be BIG, as it will be dissipating around 5W!
 

ebeowulf17

Joined Aug 12, 2014
3,274
For what it's worth, I've simulated an approximation of your current source circuit, and it works perfectly. The sim is rough, because I don't have the right part models, so I just made a rough substitute with an op amp and 4 resistors. Still, signal voltage to output current tracks perfectly as long as load resistance vs. supply voltage doesn't become a limiting factor.

The simulation shows an input signal ramping from 0-1.8V, with current ramping from 0-150mA (not sure why I thought it was 160 before, but your values are chosen correctly for 150.) The sim runs 5 times with different load resistor values. The output behaves perfectly for load resistances of 50, 100, or 200 ohms, but falls short on 600R and 1k loads.

laser-current-source_01.png

***EDIT: I did one more test, playing with supply voltage. In simulation, if I turn the supply voltage up to 155V, then the desired output current can be delivered even into a 1k load, but in real life the op amp would not be at all happy at 155V! You would need a different version of the current source circuit in order to handle a higher supply voltage.
 

Attachments

Top