the best design for a <15A linear regulator for solar power supply

[PinkNoiser]

Hi all. I've got a tk2050-based board from sure electronics (4x100W at 4ohm), and i want to power it with solar cells. It needs 30Vdc and about 500W of power (so, let's say 15Amps).
I'll use photovoltaic cells as power source: 12V modules in series/parallel. This will give me a 35/40V unregulated bias voltage, and i have to control it in the best way.
There's some stuff I already have:
-10x lm317 in T case
-5x lm338 in T case
-2x TIP142 (NPN darlington pair)
-2x 2N3055 in K package (quite old, don't even know if they still work)
-other smaller seemiconductors like TIP31, TIP32
-4x 22000uF 50V electrolytic capacitors, (and many smaller others)

Yes, i've looked for some schematics. lm338's datasheet suggests some circuits but they use a (discontinued) op-amp to drive 3xlm338 in parallel. But this shouldn't give me 15A due to high voltages and T pakage, and i wouldn't like to use and op-amp. I've looked at some schematics with a lm317 and a power transistor like tip142, but they don't go over 10A. I would like to parallel 2 of this big darlington but i don't know what's the right thing to do.

And before someone asks, yes, I still don't have 500W solar cells, at the moment I've got around 300W (stable and continuous), but I could get more power soon, and eventually use the regulator with a switching power supply. So I would design this for full-power use.

 

[palstanturhin]

I think you need to go SMPS if you want to take all available energy out from your panels.
I know this is not the answer you are looking for.

But with linear regulation you will waste most of the energy to heat and losses.

 

[DCPreamp]

Agreed about the losses with a linear supply and the recommendation for a SMPS.

However, don't get scared away by a switching supply. You can run a pretty simple Buck-type switching supply to regulate and drop the 35-40V down to a regulated 30V. You can probably even use the bipolar devices and filter caps you have.

Look at http://www.ti.com/lit/ds/symlink/tl2575hv-adj.pdf This may do just what you need and you will only need an appropriate inductor and switching diode. Plus, you can likely order a free sample from TI/National.

With this type SMPS, you can achieve 90% efficiency or better, have good regulation, and keep the circuit fairly simple.

 

[ericclemens]

I think you need to go SMPS if you want to take all available energy out from your panels.
I know this is not the answer you are looking for.

But with linear regulation you will waste most of the energy to heat and losses.

Compared with SMPS,is there any available solution to conduct motive power battery supply from your panels?Thought it will be an interesting and meaningful project.

 

[dangus]

I think this isn't the best application for solar cells. If you're listening to music, not continuous sine waves or pink noise, the average power consumption will be something like 1/10 of the peaks. If you had some storage (battery or capacitor bank) that could supply the peak current, you could get away with a much smaller solar array.