TPS563257: Variable output voltage with MCU control

Jens Schönfeld

Part Number: TPS563257
Other Parts Discussed in Thread: DRV8220

Tool/software:

Hi,

I'm thinking of using the TPS563257 for a design that needs variable voltage output between 0.6 and 9.5V from 12V input. I'd add resistors in parallel R5 (fig. 8-1 in the datasheet) and make them "switchable" by controlling NPN transistors to GND. Just calculating the resistor values with the equation from 8.2.2.2 of the datasheet gives me the desired results. However, I'd like to know if I may run into more issues, as I can't adapt the inductance (would have to use 4.7µH), and I may trigger protection circuits of the chip if I demand a too-steep voltage change.

I specifically looked for a converter with high switching frequency, so it'll react faster. Ultimately, I'd like to generate a 100Hz sine wave for a load of up to 1A.

thanks!

3 months ago

0 Lucia Gao 3 months ago

TI__Mastermind 19310 points

Jens Schönfeld

Thanks for reaching out. Since you'll need to follow the same inductor and output capacitor, pls make sure to double check the loop response performance (load transient) in order to avoid any unstable of the application circuit. We have PSpice model for you to do the circuit simulation.\

BTW, could you pls kindly share me your application/end equipment? I am just wondering where you need several Vout between 0.6 to 9.5V with one Buck?

BRs

Lucia

0 Jens Schönfeld 3 months ago in reply to Lucia Gao

Prodigy 20 points

I want to replace a transformer-based power supply that delivers 9VAC with a rated load of 1A. The goal is to increase system efficiency. The standby power of the transformer is low hanging fruit, so that's replaced with a modern 12V DC PSU.

Since the 9VAC is largely unregulated on the original/legacy part (only determined by the number of windings), some voltage fluctuation is OK. The target device does not actually need precise frequency, so I'm thinking of increasing the frequency to 100-150Hz, which will reduce heat in the target device (I already verified this successfully with a completely linear circuit and function generator).

A total of eight different pull-down resistors on the FB voltage divider let me choose 256 different output voltages at about a step of 0.035V, which is sufficiently-linear according to the Vout calculation given with equation (3) of chapter 8.2.2.2.

So I'd essentially abuse the TPS563257 as an AC converter by constantly changing the voltage divider to the FB pin. This has several advantages for me:

1) it's cheaper than a PWM circuit with, say, a DRV8220, as the device is cheaper

2) the higher switching frequency lets me use a smaller inductor (DRV8220 is spec'd to support only 100kHz PWM).

3) under-voltage and over-current protection is already taken care of

My biggest concern is to trigger one of the protections with a 40-50kHz update rate of the FB divider. My gut feeling tells me that I should add a small capacitor to the FB voltage, so the step changes will be a little softer.