i2c controlled current limited boost switching regulator

[dosensuppe]

hello. as the title says I'm trying to build a boost converter with current limiting that is digitally adjustable with I2C or SPI. It should be able to provide up to 500mA, take 5V in and output up to 10V.
So far I haven't found any switching ICs that have current limiting integrated and such would have to implement that seperately with an op amp feeding into the fb pin of the switching IC. https://www.instructables.com/Adding-a-Current-Limit-Feature-to-a-BuckBoost-Conv/

If you know any switching ICs that are directly controllable with I2C/SPI I'd be interested to hear, since DACs >10bit are a significant price factor and I'd like at least 10mV accuracy.
Right now I'm eying the TLV61048 which is just a simple boost converter without CC control or I2C, and the MCP47FVB2X 12bit dual DAC to digitally adjust the regulator and CC limiting circuit.

[PCB.Wiz]

Maybe the NCT3933U from NUVOTON ?
Intended for power supply setting/margin and trim i2c 3 DAC channels

[dosensuppe]

Not what I'm looking for, I'm afraid. That one is only available in quantities of 3k and it also only has 7bit which would result in 30mV steps at 5V adjustment range 5/(2^7)

[dosensuppe]

I guess what I'm looking for doesn't exist in integrated form.
if I were to find an I2C adjustable boost converter I'd still have to get a single DAC for the current limiting.
But maybe there is a better way to do this.

https://www.digikey.de/de/products/detail/microchip-technology/MCP47FVB12A0-E-ST/5033373
The 10bit of this dual channel DAC might be enough after all, that's already 10mV accuracy. Still expensive at 1€/pcs, but acceptable.

[dosensuppe]

how can I improve this current limiting circuit?
the comparator always swings around when limiting the current. there must be a cleaner way to implement this.

[mtwieg]

The comparator effectively has infinite gain, so the current feedback loop is bound to be unstable.

The cleanest way to implement both average voltage and average current feedback is to give each their own error amplifier with its own compensation network, then feed the lesser of the two to the PWM modulator. This isn't really practical to hack onto an existing PWM controller like the LT1307B.

[mtwieg]

Attached is an example of what I described above. The current sense feedback has its own error amplifier and compensation network, which can override the main VC node.

Though looking for a more suitable PWM controller with an externally controllable current limit would avoid having to add all this external circuitry.

[dosensuppe]

indeed. this seems to work fairly well. thanks a lot for the effort.

Though looking for a more suitable PWM controller with an externally controllable current limit would avoid having to add all this external circuitry.

I couldn't find any that would meet my requirements of delivering >=500mA at 12v at an affordable price.
I found the Silergy SY21222ABC. combining that with a cheap rail to rail op amp seems to be the most cost effective method to me.

[dosensuppe]

the sy21222 doesn't have an error amplifier compensation pin. Is there any way to get this to work regardless?

[mtwieg]

the sy21222 doesn't have an error amplifier compensation pin. Is there any way to get this to work regardless?

You could try having the current feedback override the FB pin, but in a more linear way, using an ideal diode. See attached. Note how V5 offsets Isns to get the desired 100mA limit.

Whether this ends up being stable with the SY21222 is impossible to say, but it's worth a try.

[Phil1977]

Of course it depends on the application, but sometimes it´s worth considering driving the regulator switch transistor directly with the MCU.

I once did this for charging the caps of a xenon flash - so regulation speed was of no concern, the MCU measured the slowly rising cap voltage and set its PWM-output as a simple PI-regulator.

[MariuszD]

Doesn't it bother you that the current limiter won't work for output voltages below 5V?