Op Amp Constant AC Current Source

Not my area of expertise either, but here...

The basic bipolar current source circuit is the Howland current pump. Also look up the "improved Howland current pump", which is better with low impedances like you're working with.

For the current and frequency range you describe, I'd bee looking at audio power amplifier circuits. Things that would help narrow it down:

• How stable does the 600 mA need to be? ±5% or ±0.05%? What about with changes in impedance?
• Do you need only 600 mA, or do you need adjustable up to 600 mA?
• Is there a significant reactive component to the load, or is it mostly resistive?
• Whats are the distortion specs? Will a square wave do, or do you need very limited overtones?
• Is your load grounded at one end, or can a bridged amplifier be used?
• How many channels do you want? Just one, or do you want multiple outputs?
• What are the specs on residual DC component?

If you can tolerate the accuracy loss, a transformer to increase the impedance and decrease the current would increase efficiency.

I'd be thinking of a class-D small-speaker driver. In particular, given your frequency range, a single-channel subwoofer driver would be a likely candidate. And for your supply voltages and low impedance, look at automotive-targeted products.

Another option I'd consider, if you can tolerate the low accuracy, would be a low-power current waveform generator with an output made up of current mirror stages. There's a neat trick where you have an op-amp drive a grounded resistor with a voltage waveform, and you mirror its draws on the rails to generate an output current.

Sorry I haven't got any canned answers, but hopefully this brainstorming helped a bit.

James,

Here is a modified Howland voltage controlled current source using the OPA548 power op amp. The output current is -Vin/R5. I've shown diodes as a load to demonstrate that the output current is accurately proportional to input voltage regardless of the obviously nonlinear load voltage.

The OPA548 is a bit marginal on positive output voltage swing on +/-6V supplies. You can see this as the output limits at about +1.9V and 1.4A. This is close to the max condition of your AC waveform at a 2 ohm load and this simulation is with a nominal device, not the worst case. If you could get an additional volt on the power supplies, you'd be in good shape. Alternatively, you could gain some margin by making R5 a smaller value. If you make it too small, input offset voltage may create an unacceptable output current offset.

Regards, Bruce.

Thanks Guys.

I will try the OPA548 circuit and see how it performs.

Hi Bruce, 

Can I ask a question here ? If yes, I would like to ask if this would work equally well with inductive loads or not ? 

I am using a transformer to step up the current. In simulations, it shows that it would work, cause the transformer is ideal. Do you think in real life situation, this circuit would work too ? 

Thanks for a reply. 

AQ

AQ,

A Howland current source (or pump, as it is often called), is a voltage controlled current and would control the current in the load without regard to its complex impedance (inductive, capacitive, etc.), within is voltage compliance range. The current it controls is whatever is connected to its output, which includes the transformer. The accuracy of the stepped-up current to the secondary depends solely on the fidelity of the transformer. That may be hard to predict.

Hope this helps,

Regards, Bruce.

Thanks Bruce for the reply. 

Yes I understand what you are saying is true. But apart from that, resistive load regulation is not good as well. Can you look at the attached two pictures. The difference in current would be amplified on the secondary side :) 

Is there a better way to regulate with such high impedances ?? Or am I missing something here ?

AQ,

If you change R2 to 5001 ohms I think you will see better performance. You have R7 set to such a low value that a tiny voltage must be developed on this resistor to create the small output currents you are simulating. This makes resistor ratios extremely critical in order to achieve high output impedance. If you are using the OPA548 to create amps of output current, perhaps the change of a hundred microamps is not critical. I'm not sure of your actual needs. If you need very small currents, I suggest making R7a much higher value. R2 should be made equal to R6+R7.

You may want to read this Bob Pease application note on the Howland current source. It provides an intuitive look at the issues of the Howland (improved Howland, in your case). It will help you optimize for your application. If you have further questions, it would be wise to start a new thread with more background information on your requirements.

www.ti.com/lit/an/snoa474/snoa474.pdf

Regards, Bruce.

Thanks Bruce for the reply. 

I actually did create a new thread in 'Audio Amplifier's' section but did not get any replies. Changing R2 to a better value did help a bit. Later on, by reading Bob's paper, I think improved Howland pump would not give me 100% accuracy at all load levels. I would try to see if their other topologies, which I can use to realize the design. Thanks for the help again :)

AQ