Help with building a current supply

[neomoses]

I'm looking to build an adjustable current power supply for anodizing. I've got everything to make a 12V, 10A, unregulated power supply (Transformer, Bridge Rectifier, 27,000 uF Capacitor to smooth out ripple).

I now need some help with the closed loop current control. Basically I'm wanting to use a pot to set a desired current (somewhere between 0-9A). Anyone know how to do this?

Bryan

 

[Heydave]

I have a page with some basic curent source information:

http://www.inventgineering.com/current_supplies.htm

The page is meant only as a guide. For your application, you will need a current source with some high power output transistor such as a darlington since a typical op amp does not have the power to drive the base of a single PNP for 9 amps. You can create a PNP darlington by cascading two PNP transistors. The reference resistor should be about 0.1 ohm for the current levels you are wanting. The zener voltage reference and potnetiometer establishes a voltage across the reference resistor. With 0.1 ohm, 1 volt will source 10 amps. The output transistor will need a heat sink!

 

[lcsjk]

Look at the first stage of that circuit. It is an op-amp driving a PNP transistor, and is a current source. Using that circuit topology, change the PNP to a P-channel mosfet, and add a gate source resistor of about 2k. Change the 1 ohm resistor to a 0.05 or .01 ohm resistor. It will be a power resistor (I squared times R). The voltage across the .01 ohm resistor will be the same as the voltage from the + supply to the pot wiper.

Be sure to put the mosfet on a large heatsink and plan on using a fan. Keep in mind that the bridge rectifier will be dissipating at least 10amps at .8 volts for each half cycle. Also note that the voltage droop from the 27000 uF cap being discharged at 10 amps for a half cycle is only 3 mV or 6mV based on your configuration.
This means that the capacitor must be recharged for a very short interval near the peak of the rectified sine wave and this current will be far greater than 10 amps, even though it is only for a 3-5 milliseconds. The current is limited only by the transformer winding resistance. The bridge must be able to handle this repetitive surge current, and also the power of more than 10amps times .8 volts.

 

[Heydave]

One thing to be concerned with is stability. If the output does not respond fast enough, then the op amp outputs will saturate trying to get the transistor to follow them. The result will be oscillation. Large transistors are typically slower and have higher base (gate) capacitance that will exaggerate the problem. The op amp to drive a power fet may need dampening. IGBT power transistors might be a little better but not without concern. An IGBT is a combination of FET and bipolar transistor much like a darlington. Do not invest in high speed op amps but look at devices with external compensation available. One way to dampen the op amp and isolate capacitance laod is to put a resistor in series with output and put a small capacitance from output to neg input. The feedback from the transistor should also have a series resistance for the feedback capacitor to be effective.

 

[cbarn24050]

Hi, a simple constant current circuit with a large power transistor would do. 2 transistors a resistor and a pot, only snag is the heat dissipation in the power transistor. Or you could use a switching regulator, a lot more effort but much more efficient.

 

[hacksaw]

just put a variac on the primary. won't be perfect (current or voltage), but you can adjust quite easily for one-of type situations. you can provide some degree of current stabilization with a dropping resistor(or lamp).