Linear bench power supply

[Torrentula]

Hello guys!

I'm a first year EE student and I am planning to build my own bench supply. I have built some power supply circuits with the venerable LM317 and LM337 regulators but I plan on building something with a higher current output.

My circuit isn't extraordinary, basically similar to power supply designs that other people on the forum have already done.

The basic specs I am going for are 0-20V output voltage and 0-3A output current, both a constant voltage with current limiting and constant current mode. The whole thing will be powered from a 24V laptop power adapter type of supply so I don't have to worry about mains wiring plus it already has a standard IEC input connector.

I also have an LM2576 as a tracking pre-regulator to minimize the power dissipation in the output pass transistor. The tracking control circuit is the same simple transistor based one that Dave used in one of his power supply circuit simulation videos and is supposed to keep the output voltage of the LM2576 always about 3V above the transistor source voltage.

Here is my first question: for which output voltage should I calculate the inductor and output capacitor values? As far as I understand it I should be calculating those values for 23V at 3A, right?

For the linear output section I plan on using an N-channel MOSFET something like an IRF540. I have added a capacitor in the feedback loop to slow it down to below 100Hz. Does that seem like a good value? I know that these circuits can be very unstable because the FET gate presents a capacitive load to the op amp so I also added a 10k resistor between the op amp output and the gate.

The voltage and current setting controls are very similar to what Dave has shown in his power supply videos. The PWM from the micro will go through a two-stage RC LPF and buffered by a jelly bean op amp.

Because sourcing special parts like the TS1102 isn't easy for me here in Germany I just decided to use the second half of the LT1490 as a differential amplifier with a gain of 100 for the current measurement.

Does it look good to you or is there something I am missing? I appreciate any suggestions!

EDIT: I have done some more work on the schematic. I still need to find a suitable TVS diode to protect the output from voltage transients coming in through the terminals. There still remains the question as to for which output voltage I need to calculate the inductor and output capacitor values for the tracking pre-regulator.

The new schematic can be found here: https://cdn.hackaday.io/images/8953191424640891899.png

[Cliff Matthews]

Good work sir!  The selection and prices of DC-DC convertors is mind boggling on eBay... perhaps the final output could go as high as 30v?? and... *
Since the pre-regulator can be a module, why not make all the switches, LED's and segments go modular too?

Question: Does anyone know a good supplier of key caps that could sandwich between this display and a front bezel ??
http://www.ebay.com/itm/8-Bit-LED-8-Bit-Digital-Tube-8-KeyS-TM1638-Display-module-for-AVR-Arduino-ARM-/400531985021?pt=LH_DefaultDomain_0

* As a co-incidence, while Dave was hacking a huge VFT this morning, Julian Ilett was showing off a 4 dollar Arduino compatible display/LED/switch board that I thought should fill the bill just nicely...  BTW, IMHO nice LED's are much better to see on a bench PSU 

[Cliff Matthews]

Note: I decided to provide a link to another eBay store (Alice seems recommended by some members) and not the one Julian mentioned.
I hope none takes offense, I found the Lakeyx101 store a disaster (most items not categorized) and Alice sells twice as many.

Still, I wouldn't mind finding a source or secret (apart from a lathe) for button's to pass through a front panel of a metal project box

[codeboy2k]

... There still remains the question as to for which output voltage I need to calculate the inductor and output capacitor values for the tracking pre-regulator.

Here's how you figure that out:  If you look at the datasheet, you see the choice of inductor depends on the output and input voltage only.   There's a formula there:



Lets assume your input voltage is 30V, and your output voltage will be 3 to 27V.

If you plot that, for Vin = 30V and Vout = 3V .. 27V, you see that it's a parabola:



The X axis is the output voltage, the Y axis is the Volt-seconds in microseconds. You see that there is a maximum at 15V (which makes sense, it's midway between the min and max output voltage).

So an inductor that can support a magnetic flux (Weber) of at least 144 (V.us) is needed.  If you look at their table 2, and assume that you want 3 Amps, then you get an inductor of 220 uH.  That size will support the maximum magnetic flux needed in the coil, without saturating.

For the capacitor, you will have the largest ripple when the difference between input and output voltage is the largest, i.e. at the smallest output voltage.  So size it for the smallest output voltage, say 3V. But give it a voltage rating equal to 1.5 times the largest output voltage.

 Their formula is :


So, for Vin=30V, Vout = 3V minimum, choose C = (13,300*30)/(3 * 220) = 605 uF.  A standard size is 680 uF. Choose the voltage rating to be 1.5 x 27V = 40V. A standard is 50V . So C = 680uF @ 50V. You can go higher C for less ripple, say 1000uF @ 50V if you wanted.