Solar charger - TL494 and what transistor?

[Artlav]

Good evening.

I want to charge a supercapacitor bank from a solar panel.
The panel gives max power of about 0.5A at 32V, so the charger should track it's input voltage.
The output should go down to zero and up to about 30V as the bank charges, with no need to explicitly limit output A or V.

To that effect i made this little circuit:
Q1 is TIP32C.


Problem is, i'm not sure what transistor to pick, or how to drive it.

It works as shown, but R2 gets hot.
If i make it a darlington, and put a 10k resistor, then the transistor gets hot.
I can't use a FET, since i can't figure out how to (simply enough) drive it within +-20Vgs limit.

Should i use a transistor with higher gain, and higher R?
Some sort of darlington arrangement?
Is there a simple way to use a FET in there?
Maybe there are other design flaws?

[T3sl4co1l]

See the example here:
http://www.fairchildsemi.com/ds/KA/KA7500C.pdf

Same chip, different number.

Note the B-E resistor which improves BJT turn-off, and yes, the series resistor (an even lower value in this example) will get hot.

To drive a MOSFET, try a bootstrap type driver IC, or build a level shift circuit.  You can use N or P channel types here.

TL494 is quite old and not the best choice for this application.

Tim

[mariush]

Why not just use existing ICs like LT3652 ?

Does everything you need except only up to 18v output:

Description

The LT®3652 is a complete monolithic step-down battery charger that operates over a 4.95V to 32V input voltage range. The LT3652 provides a constant-current/constant-voltage charge characteristic, with maximum charge current externally programmable up to 2A. The charger employs a 3.3V float voltage feedback reference, so any desired battery float voltage up to 14.4V can be programmed with a resistor divider.

The LT3652 employs an input voltage regulation loop, which reduces charge current if the input voltage falls below a programmed level, set with a resistor divider. When the LT3652 is powered by a solar panel, the input regulation loop is used to maintain the panel at peak output power.

The LT3652 can be configured to terminate charging when charge current falls below 1/10 of the programmed maximum (C/10). Once charging is terminated, the LT3652 enters a low-current (85?A) standby mode. An auto-recharge feature starts a new charging cycle if the battery voltage falls 2.5% below the programmed float voltage. The LT3652 also contains a programmable safety timer, used to terminate charging after a desired time is reached. This allows top-off charging at currents less than C/10.

Can be adapted to charge banks of capacitors but alternatively, you can just pick another ic that does supercap charging: http://www.linear.com/parametric/Supercapacitor_Chargers

[Artlav]

Why not just use existing ICs like LT3652

TL494 is available in any store for spare change, LT-anything would have to be shipped across continents.

Same chip, different number.
Note the B-E resistor which improves BJT turn-off, and yes, the series resistor (an even lower value in this example) will get hot.

Yes, that is what i used as a reference.

The series resistor dissipates 1W of heat, and does this constantly during all the time the capacitors are sitting there charged full.
My main question was - how to avoid that extra heat dissipation, aka how to improve efficiency?

Meanwhile, added the B-E resistor, with no noticeable effect on heat.
Also added output current limiting, since the thing should actually survive something close to indefinite short circuit.