In a previous post about the Test Bench Rack I received helpful comments from John Loffink, who had also refurbished the same Pittsburgh Modular power supply board. He educated me about Low DropOut voltage regulators.
LDO Regulators
An LDO regulator is a linear regulator that can operate at a very low potential difference between the input and output voltage, the Drop Out Voltage. The common 7812 and 7912 TO-220 package regulators we’ve used for decades have typical specifications of:
LM7812 Voltage In: 14.5V to 27V; Drop Out Voltage 2V
LM7912 Voltage In: -14.5V to -30V; Drop Out Voltage 1V
The LDO regulators that I used here:
LM1085 Voltage In: 13.5V to 25V; Drop Out Voltage 1.3V typical
LM2290 Voltage In: -13V to -26V; Drop Out Voltage 0.6V typical
I don’t know exactly why the minimum input voltage minus the 12 volt output spec doesn’t match the Drop Out Voltage spec. It’s a vagary of data sheets maybe John L. can explain.
The schematic shows this simplest of analog power supplies. The input comes from an AC wall wart, specified at 12VAC, 2A. Half-wave rectification charges up a 2000uf capacitance, and the voltages at the input to the regulators measure 18 volts with zero load. As load is added, however, this voltage will drop. My understanding is that if the load is high enough that the Drop Out Voltage goes lower than the spec (i.e. the input voltage drops too low), then 60Hz ripple will start to appear on the output.
Thus using LDO regulators increases the maximum current the regulator can deliver without exceeding acceptable ripple. You want to minimize the ripple. I didn’t take the trouble to measure this. I’m happy that ripple will at least be less likely a problem with the LDO regulators. And after all, this supply will only be used in a test rack where the load will remain low anyway.
Parts and Assembly
John kindly sent me a list of the parts he used. However the positive regulator he listed, Microchip MIC2940A-12WT, was not available. So I went searching on Mouser.com for a +12V LDO regulator in a TO-220 package. I found a T.I. LM1085IT-12/NOPB 3A LDO Positive Regulator. The negative regulator John listed was available, T.I. LM2990T-12/NOPB LDO Negative Regulator.
I also ordered the heat sinks John listed, Mouser 532-577102B00, but decided I didn’t need them at the current draws I expect.
I didn’t look at the data sheets before buying. Now, hmm, the LM1085 has a different pin out from the LM7812, which is normally In, Gnd, Out.
So it took some trace cutting and kludging (notice the red and black wires soldered to the regulator in the top picture) to fit it onto the board.
The LM2990 negative regulator has the same pin out as an LM7912 fortunately.
Another improvement made was to fortify the current-carrying traces, which were tiny. Now there’s 22 gauge wire throughout the power carrying part of the circuit. In addition to the inputs and outputs of the regulators I beefed up the wiring from the AC input to the half-wave rectifier diodes. The only part of this board design that seems to make sense is the ground plane.
I added 10uf 25V caps on the regulator inputs and 33uf 25V caps (gold) to the outputs. Nothing special about the 33uf caps, except I had them around.
Conclusion
Did I really need LDO regulators here? Maybe not, but fortifying the current carrying traces was definitely needed and so I decided to do the regulator replacements. Thanks, John, for all your helpful suggestions.
