PSU
OK, I’m happy with the PSU performance now, and I’m ready to put it into a case.
Again this is my take on a popular project. I’ll try to outline some changes I’ve made, and why.
First a few general specs:
1: 0 – 25 V 0.04 – 3 A. (10 mV and 1 mA meter resolution).
2: Regulation @ 3 A <0.01%.
3: Stability @ 1 A after 1 hr <0.01%.
4: Ripple @ 3 A < 100 mV.
5: Push ON /OFF toggle switch that defaults to OFF, with over volts cut-off.
References are from the original circuit.
The voltage and current range were dictated by the transformer I had (24 V @ 6 A ). This gives a NO LOAD voltage of 33 V. Adding the -1.3 V to this gives a voltage range of 34.3 V. As the IC’s are rated at 44 V I could dispense with the zener diode D13.
Reading through the various posts it’s apparent that a lot of problems are caused by trying to use a transformer with different specifications! Unless you are prepared and able to make circuit changes, stick as close as possible to the original specified transformer.
The heart of a good PSU is the voltage reference performance. After doing a number of tests of voltage and temperature stability on the original, built around D8 and U1, I changed it to a constant current FET circuit. The component count is slightly less and the performance is better. As the reference is 6 V not 11 V the gain of the OP Amp U2 was increased from 3 to 4 (approx).
I changed the current sense resistor from 0.47 R to 0.1 R @ 1%. The gain of U3 was also increased to account for this. I made the resistor from 10 x 1 R 0.6 W 1% in parallel. This gives a 0.1 R 6 W 1% resistor. The advantages of this is reduced voltage drop and reduced heat dissipation. (0.1R @ 3A = 300 mV - 900 mW, 0.47R @ 3A = 1.41 V – 4.23 W ). It also neatly drops 100 mV per 1 Amp of current. i.e. 300 mV at the 3 A max. As I used a 4 ½ digit 200 mV LED display (19999 max) I put a 10:1 3 resistor divider on the input of the meter and set the decimal point to 19.999 this gives a me a 1 mA to 3 A reading with a meter resolution of 1 mA ( 0.001 - 3.000 A ). I’m not pretending the precision is as good as the resolution, but it’s within a few mV. To further improve performance I separated the sense outputs and joined them directly to the output terminals, I could have brought them out to the front panel but it meets my requirements as it stands. I also used a 4 ½ digit 200 V LED display (19999 max) for the voltage display. This time the decimal point is placed at 199.99 giving a meter reading of 10 mV to 25 V with a meter resolution of 10 mV. (00.01 – 25.00 V ). As I wanted a push ON/OFF DC switch, I made a “stand alone” toggle circuit. (See previous post for details). This is not directly incorporated into the PSU and can be omitted. Some DC switch should be fitted however and turned ON after the AC power up and OFF before the AC power down. This is important as there are some voltage spike issues! The toggle switch defaults to OFF and incorporates an over voltage (pass transistor fail) shutdown function. I will post some pictures of the built PSU when it’s finished. (I’m still waiting for the push switch with built in LED).
I’ve inc. a few pictures including the final circuits. Again this is my own take on the “old” design. Why I made the changes may be of interest to someone.
