Laboratory Power Supply with Switch Mode Pre-Regulator [130234-I]
Commercially available laboratory power supplies often have disadvantages regarding the user interface, the performance or the price.In this project a laboratory power supply module with the following commonly used output characteristics is developed:
Commercially available laboratory power supplies often have disadvantages regarding the user interface, the performance or the price.
In this project a laboratory power supply module with the following commonly used output characteristics is developed:
- Uout = 0 … 30 V
- Iout = 0 … 2 A
The described power supply module is equipped with a switch mode pre-regulator. The pre-regulator ensures high efficiency and low heat generation in any load point. In order to achieve a good output voltage quality and control characteristics a linear post-regulation is also implemented.
The Module itself provides galvanically isolation in the switch mode pre-regulator. Consequently, several modules can be connected to the same power source in parallel. The input voltage for each module is 12 V, thus a simple commercial power supply, like a computer power supply, can be used for mains connection of several modules.
The following features are already implemented and tested:
- Switch mode pre-regulator
- Linear post regulator
- Output relay for save disconnect
- High accuracy of the voltage control loop
- Adjustable voltage and current presets when the output is inactive (no short circuit required for current limit adjustment)
- Minimized overshoot of the output voltage at recovery from short circuit
These features shall be implemented or will be tested in the future
- Galvanically isolated RS232 connection via RS485 for remote control (not tested jet, but implemented)
- Control panel with 7 segment LED display and rotary encoder
- Evaluation of a PCB integrated transformer for simple buildup of the module
Overall, this project focuses on usability for all tasks in everyday laboratory life. Consequently, the behavior of the module is optimized and tested in order not to destroy the connected circuitry like it sometimes happens with cheap power supplies from dubious manufacturers.
Update 07.05.2013
New measurements of short circuit current limiting and voltage recovery were done.
Added file: 2013_05_07_Measurements.zip
Test conditions:
Current limit 350 mA, Output voltage 5V, short circuit was done with IRFP3206 MOSFET, several component values differ from the values given in the schematics, due to optimizations.
Screenshots show:
Channel 1: Linear regulator output
Channel 2: Switch mode preregulator output
Channel 3: Output current
Test results:
Current overshoot at short circuit: 4.2 A, used output capacitor 470 nF (different from schematic)
Current overshoot duration < 10 µs
Voltage recovery overshoot < 10 mV
Voltage recovery time < 40 µs
Under the tested conditions the power supply behaves nearly perfect. Especially the voltage recovery with nearly no overshoot is as desired.
The switch mode preregulator control needs further improvements to reduce the occurring oscillations in the preregulated voltage.
Discussion (6 comments)
Barbara Perkins 4 years ago
James Karter 4 years ago
am_gd1991 li ming 4 years ago
Arne.Hinz 11 years ago
Last Friday final tests of the prototype were finished. The measurements include the transition from open circuit to short circuit and vice versa, and from 90% load to 10% load. Further, the static accuracy of voltage and current under open circuit, short circuit, and 90% load conditions were tested.
The dynamic tests were done by switching the loads with a 2SC2837 npn transistor.
Results:
static accuracy without load:
- voltage deviance: < 1% for voltages bigger than 0.5V; below this line deviance increases
- current deviance: > 4% due to the failure of the INA193; can be fixed by a greater shunt resistor
voltage accuracy at 90% load:
- deviance < 2% for voltages higher than 5V; below this line deviance is increasing
- ripple: either voltage nor current ripple could be discovered
dynamic test results:
transition from open circuit to short circuit and vice versa:
- current overshoot: duration < 250µs
amplitude < 10A
- voltage recovery: duration < 7 ms
overshoot: < 200 mV
transition from 90% to 10% load:
- overshoot: < 200 mV
This results show some disappointing setbacks according to the dynamic characteristics, as the current overshoot amplitude and duration has gone up and a voltage overshoot takes place. The reason for this is that at the moment when the load is suddenly decreased the current does not decrease instantly and instead flows into the output capacitor increasing the output voltage very quickly. Thus, as the controller has got a limited bandwidth, it can not react as fast as the voltage is rising and an overshoot occurs.
I haven´t found a way yet to speed up the controller to minimize this overshoot to an acceptable level and I also haven´t any proper idea to slow down the rising of the output voltage without getting into trouble with oscillations except with a huge output capacity, which causes the worsening of the current overshoot.
If you have a clue how to solve this problem in a smarter way, do not hesitate to post a comment or write me an e-mail!
Another point is, that it turns out that the PCB transformer is not capable of providing 60W and just delivers 30W. Thus, without an improved transformer the specifications of the power supply must be corrected to 30V 1A for voltages above 15V and 15V 2A for voltages below.
However, the next step now is to redesign the pcb.
Arne.Hinz 11 years ago
Moving to a output current of 2 A the circuit first worked fine but suddenly starts to oszillate, also affecting the before tested working points at 100mA (see attached file).
It seems like the origin of the disturbance can be located in the switch mode preregulator as the other regulator haven´t been changed. Additionally, the elcos at the output of the preregulator heat up to about 65 °C when operating at 2A output current, which isn´t that suitable. Thus taking into account the time spent on setting and testing at this operation point, it wouldn´t be suprising if the elcos are damaged.
Another point is the size of the inductor of the preregulator, as there already have been some minor oszillations in the previous test at 100 mA, which have faded away quite fast. This could be caused by a too big current ripple, which could also be the reason for the elcos to heat up.
However, a further issue showed up. At output voltages below 5 V the deviation between the set voltage and the actual output voltage rise constantly. The output voltage doesn´t even fall below 2.5 V.
This behaviour may be caused by leakage currents in the amplifying stage thus a current of just about 0.3 µA is necessary to set the output to 2.5 V. But this is rather a suspicion and need further investigations.
Arne.Hinz 11 years ago
Promising measurements of the new prototype were made.
Test conditions:
output current 100mA, output voltage set between 5V and 30V, short circuit was done with a 2SC2837 transistor.
Screenshots show:
Channel 1: output current
Channel 2: output voltage
Test results:
This results show an improvement in comparision with the previous power supply version as the voltage is stable and not overshooting over a wide range. Further, the current overshoot doesn´t exceed the amplitude of the former supply version although the output capacity is significantly bigger.
However, this improvements are paid with a setback in the dynamics of the system; especially the voltage recovery time increased.