The completed power supply with two adjustable 0 to 20 volt floating supplies @ 1.5 amps each.
I recently updated my test bench power supply with new circuit boards, a robust recycled power transformer, fine voltage control adjustments, output switches with indicator leds, floating outputs, earth grounded jacks, and a power cord jack.
The inspiration for my power supply updates came from an article in the March 2007 issue of Nuts & Volts magazine titled "A Test Bench Power Supply". I used my existing power supply case so I had to be somewhat creative with the front control panel layout because of the existing holes in the case.
On the back of the power supply I have two LM317K regulators in TO-3 packages mounted on a heat sink recycled from an old 1960s electronic organ.
This power supply uses the core voltage regulation circuit (with modifications) and some of the other features that were used in the original magazine article.
Each supply has separate coarse and fine voltage adjustment potentiometers and are adjustable down to a minimum of about zero volts through the use of LM385BZ voltage reference ics.
The power supplies are electrically isolated so you have flexibility in how they are configured with reference to each other and earth ground.
I have each of the two power supply circuit boards mounted to the back of the case. The start-up delay / shutdown relay circuit board mounted to the bottom of the case. I am using the large transformer to provide the primary output power for both power supplies and the negative reference voltage of one of the supplies. I also have a small transformer mounted to the back of the enclosure to provide the negative reference voltage for the second supply.
After finishing the updated power supply and using it a few times I discovered the voltage reference ic would get fried if a power supply output was accidentally overloaded.
I fixed that problem and some other issues by adding a few components and a relay delay circuit. The relay delay circuit allows time for the power supply to stabilize before the outputs become active and more importantly shuts them down as soon as the line power (120VAC) is removed.
To the right is the modified power supply circuit with components added to resolve the previously cited issues. The added components are in red with the exception of the C5 capacitor that I changed from 100uf to 470uf.
I added the C7 capacitor in parallel with the voltage reference ic to prevent the C6 capacitor from being discharged through the voltage reference ic if the output is abruptly overloaded.
I also added the D4 Schottky diode in parallel with the voltage reference ic that prevents the ic from being forward biased more than .3 volts and prevents the 470uf and 100uf capacitors from being subjected to a reverse polarity greater than .3 volts during line power shutdown.
I added the R4 2.2K resistor to speed the discharge of the C1 4700uf main filter capacitor. This was more of an earlier effort to reduce output voltage rise problem during line power shutdown but also helps to reduce the time that the voltage reference capacitors are reverse biased.
The power supply has two resistors that have to be calculated. I used
a value of 1.5K for resistor R3 to provide a current flow of 10ma
through the voltage reference ic. The effective current flow through
the reference ic will actually about 5ma because of an opposing
current through the voltage adjust circuit.
The R2 resistor value(s) will be chosen to to set the maximum voltage output for the power supply.
My actual values are 11K and 13K.
Updated power supply board with components added to protect voltage reference ic and reduce voltage overshoot but I have since changed the regulator ics and mounted them on the back of the enclosure.
To resolve the output voltage swing problem that occurs when the line power is shutdown, I added a relay delay circuit.
I was going through my electronics parts and discovered that I had a Potter & Brumfied KHU-17D11-12 relay with a 12VDC coil and four sets of contacts rated at 28VDC, 3A that would be a perfect fit for this project.
Here is the completed relay delay circuit board ready to install. The delay time shortened quite a bit on the final grid board build as compared to the breadboard build but not so much that I had to adjust the timing.
For the relay circuit I used a DF06 bridge rectifier package fed from one of the power transformer secondary windings to power the circuit. Substituted a 2N2222A transistor in place of the BC549 specified and used an old discontinued Radio Shack 276-1020 because it was the only SCR I had in my parts inventory.
I built the time delay with relay circuit on a breadboard trying different parts until I got the delay time right. I also changed the specified 1N4001 D1 diode to a 11DQ06 Schottky diode to improve timing circuit's capacitor discharge. This tightened the delay timing repeatability.
I used a 680K resistor for R1 and a 47uf capacitor for the C1 timing components. The delay time is about 2.6 seconds for the initial power on and 1.6 seconds for repeated startups.
I didn't want to put too much time into the relay board project so I built it on a Radio Shack 276-149A circuit board.
I marked the locations on the circuit board where the relay contacts would pass through and cut slots in the board with my Dremel rotary tool than removed the copper lands around the board slots with a small X-acto knife.
I mounted the relay and then positioned the rest of the parts around the board bending the leads into position. After I had all the parts positioned I soldered them into place.