CC/CV Regulated Bench supply

           I needed a bench supply with CC/CV mode so I looked for online designs. All I can find were LM723 based and have fixed current limit set. I needed adjustable so I kept on looking.

           And then browsing my library of PDF datasheets, I stumbled upon an application circuit from the LM317 datasheet.

Some parts were already hard to find and I didn't have a 250K 10 turn pot in my parts bin so the values were adjusted according to parts availability.

I doubt the transformer I used could handle 5A so max output current was lowered to 2.5A.

I also wanted separate LED's for CC and CV mode and a clever little circuit with an optoisolator and a transistor was used to make it work.

Here's the power supply.

Fan sucks air through the heatsink fins through the area under the blue 10 turn pot and blows hot air out through a hole in the back.

Panel meters are digital and requires an isolated 9V supply. I made a capacitor based isolator but has trouble displaying voltages when there is noise on the AC line. I'll have to redo that and I'll take more pics of the internals when I get to messing with it.

Here's another with labels and while testing a car amp....

An output switch is necessary for when adjusting the output voltage before connecting the load. Also, during shutdown, the output fluctuates and sometimes overshoots. The switch allows me to be able to connect sensitive loads only when the supply has stabilized.


11 Oct 2011

After using for a while and noticing the meters go crazy when there is noise in the AC line, I decided to make a high frequency switchmode PSU for the meters. The circuit concept was taken from here. I also added the voltage divider and current shunt in the same PCB. I also used a relay to switch the output on/off instead of wiring the output to the front panel switch.

This is the complete circuit.

The circuit is powered off the main transformer's extra 16-0-16V secondary (in this case, D1 and D3 is not used - use only when secondary doesn't have center tap). 'Vin' is from the Regulated PSU circuit output, 'Vout' goes to the front panel output terminal.

A nice feature in using a relay to switch the output is realized when charging a battery. When the power fails, The power to the circuit collapses in a short time (due to small value of C8) so the relay opens up. Once the relay opens up, voltage from the battery connected to the output terminals won't flow back into the Regulator board basically preventing it from discharging again and possibly damaging the regulator circuit.

Another is during turn off when the output switch is left in the 'ON' position. Relay cuts out the output before the output of the regulator board spikes up preventing damage to the connected circuit.


       12 Jan 2014:

       With difficult loads (some switchmode supplies) the supply has instability switching in between CC and CV modes (it works perfectly fine with most loads) so I decided to investigate. I also took this opportunity to recalibrate the meters as they seem to have drifted in time.

       My original assumption was that the tantalum output capacitor may have popped due to the various torture it has endured since I built it back in 2009. A quick inspection says otherwise. The cap is fine. Analyzing the circuit further, the capacitor across the op amp (C5 in the schematic above) still remained at 75pF even when I changed resistor values altering the loop stability of the op amp circuit. I changed it to 2n2 (since I'm using a 10K multiturn pot and I had to adjust other part values as well) to maintain a 15kHz RC time constant between C5 and the parallel value of R2 and R5. By checking the problematic waveform (yellow trace - PS output, blue trace - op amp output), I also added a new 33uF cap (should be bipolar) between the op amp output and LM317 adj pin which has fixed the problem. This is caused by having a heavy load operating near the set current limit along with high load capacitance (input C of the tested DC-DC converter)resulting to oscillation and a rectifying effect of the op amp output to adj pin since they are coupled only by diodes. Here's the waveform after the fix.

       Here's the improved and updated schematic for the Bench PS as used and I also included a schematic and PCB layout of the meter power supply which provides isolated power for the two digital panel meters, includes the current shunt and voltage divider for the A and V meters, and finally a relay switch circuit that is used to turn the output on or off. I used a relay so that when charging a battery and the power is cut off, the relay de-energizes and disconnects the load from the supply. For now, enjoy the pics I took when I had it apart.

Uncovered. I already forgot how this looked since it's been years since I last took this apart.

Transformer, probably from a cheap DVD player.
The 3-0-3V tap is used for the -7V supply of the op amp (CT not used).
16-0-16 is used to power the meters and relay board.
0-12.5V powers the fan via a 317 regulator that is temp controlled
and finally, 0-16.8V tap is for the main high current rail.

Back of the meters.

Regulator board.

After the changes. I decided to twist all the long wires to improve shielding and noise rejection.

Detail of the relay/meter PS board.

Page created and copyright R.Quan © 06 Feb 2011.