Suitability of TI (formally National) LM25576 for a New Design

[Chips & Chips]

I am doing a design for a supply that will feature switchable output Voltages in the range of 2 to 14 Volts. This switching would be accomplished with power on and load connected. Load current from 0 to 1 or 2 Amps. My preliminary plan is to design the feedback divider for the lowest desired Voltage and mechanically switch additional resistors in parallel with the low side resistor to change to higher output values. The input to the regulator will probably be filtered 18VDC but this can change. The specs are fluid at this point.

I am looking at the LM25576 from Texas Instruments for this design. Is this a good choice or is there a better device that I should look at?

One point of concern is the statement in the LM25576 data sheet that there is a minimum load current for CCM (continuous conduction mode). I am not sure what "continuous current mode" actually means and what would happen if the load current is lower than this minimum value. Would it go out of regulation? Would the ripple increase? Or what?

Any help would be appreciated.

 

[schmitt trigger]

CCM means that the load current is higher than half the inductor's ripple current; such that the current "valley" never reaches zero.

The transfer function changes and it may lose regulation. It all depends on the controller..... specially if it is a synchronous solution, because the bottom Fet will start discharging the output capacitor

There are a couple of ways to get around this....one is to increase the inductor's value and/or the switching frequency. Another is to put a "bleeder" load to have a minimum current all the time.

 

[ZekeR]

Not sure if the LM25576 will work for you; you can request a demoboard and try it.

If it doesn't work, I have lots of confidence in this chip: the LT8610. It has pretty much the same specs, but is newer, has more features, and has higher performance. Also, its datasheet actually shows waveforms and data during light load operation. For example, to see what the inductor current looks like when the chip is not in CCM, look at page 7, middle right waveform. The chip has been very thoughtfully engineered.

 

[schmitt trigger]

Indeed, most of LT's devices appear to be well designed and thought out, and offer additional features and performance benefits as compared to their direct competitors.

My only gripe with LT is that all of their devices are only available in the smallest SMT packages. I understand their rationale, they want to focus on high performance markets. But this makes it difficult for the casual hobbyst to experiment with one.

 

[ZekeR]

My only gripe with LT is that all of their devices are only available in the smallest SMT packages. I understand their rationale, they want to focus on high performance markets. But this makes it difficult for the casual hobbyst to experiment with one.

Indeed, this is a problem of all high-performance switching regulators these days. It's virtually impossible for a switching regulator to get good performance in a hobbyist package like DIP for various reasons (thermal, electrical, reliability...), so nobody even tries anymore. The big packages are relegated to the low-frequency/high voltage stuff. We just have to cope by buying better soldering equipment and practicing our soldering skills.

 

[Chips & Chips]

Some good help here. Thanks to all.

OK, I am coming to understand a bit about CCM. I still have a bit of trouble translating the current in the inductor to the Voltage and current at the output, after filtering. But I am reading all I can about switchers.

You mention synchronous below. That throws another monkey wrench into the works. I assume that this refers to synchronizing the oscillator frequency. But, if it is a single regulator, then what would it be synchronized with? And why would it be synchronized? Even though I did plan to have two or three regulators for several output Voltages, why would I want or need to synchronize them?

Or does synchronous mean something else here?

And yes, I am beginning to see that I am going to have to breadboard two or more circuits to pick a good one.

Paul A.

CCM means that the load current is higher than half the inductor's ripple current; such that the current "valley" never reaches zero.

The transfer function changes and it may lose regulation. It all depends on the controller..... specially if it is a synchronous solution, because the bottom Fet will start discharging the output capacitor

There are a couple of ways to get around this....one is to increase the inductor's value and/or the switching frequency. Another is to put a "bleeder" load to have a minimum current all the time.

 

[schmitt trigger]

Synchronous is a short hand for Synchronous Rectification. Meaning that the diode is replaced with a Mosfet. Since the Mosfet will not commute naturally like a diode, one must synchronize its gate drive to be the opposite of the main Mosfet. Doing this will result in a substantial efficiency improvement for low voltage, high current supplies.

Now, having said that, the Mosfets have an intrinsic "body diode" and will conduct even without gate drive. But this diode usually has lots of losses, so it is not wise to allow it. Some vendors specifically design Mosfets with low-loss diodes in the same device, like the "Fetky" family from IR.