The right digital potentiometer to control a Buck and a Boost Converters

[shamooooot]

Hello guys

I have two circuits (a buck and a boost) based on LM2673 powered from 20V power supply, and I would like to control them using a uC preferably through i2c.

The output ranges of these circuits should be adjustable with one being ~0 - 18V while the other -18V - ~0.

I am thinking of digital potentiometers, but I am not sure how to pick the one with correct parameters.

For example how about the MCP45HVX1:
 
https://ww1.microchip.com/downloads/aemDocuments/documents/MSLD/ProductDocuments/DataSheets/MCP45HVX1-Data-Sheet-DS20005304.pdf

[shamooooot]

Seems like MCP45HV is a good fit, but I learned that I need at least +18V and -18V on the analog voltage supplies to achieve the required range.

I will have to invert my 20V source into -20V. The simplest way seems to be a charge pump, but the IC's with this voltage range looks expensive.

[ledtester]

Doing a search for:

control a buck converter with a digital potentiometer

on electronics.stackexchange.com comes up with discussions like:

https://electronics.stackexchange.com/questions/630076/buck-converter-adjustable-output-voltage-using-digital-potentiometer

https://electronics.stackexchange.com/questions/92995/can-i-use-digital-pots-for-my-feedback-resistors-to-a-buck-boost-regulator

https://electronics.stackexchange.com/questions/472588/how-to-replace-a-potentiometer-with-a-digital-solution-in-a-buck-feedback-loop

...

Common critiques are that the digipot may not operate in the desired voltage range and will also introduce capacitance which will affect stability.

A frequently suggested alternative is to use a DAC to inject current into the feedback pin. The DAC could be implemented with PWM from a microcontroller.

[shamooooot]

Doing a search for:

control a buck converter with a digital potentiometer

on electronics.stackexchange.com comes up with discussions like:

https://electronics.stackexchange.com/questions/630076/buck-converter-adjustable-output-voltage-using-digital-potentiometer

https://electronics.stackexchange.com/questions/92995/can-i-use-digital-pots-for-my-feedback-resistors-to-a-buck-boost-regulator

https://electronics.stackexchange.com/questions/472588/how-to-replace-a-potentiometer-with-a-digital-solution-in-a-buck-feedback-loop

...

Common critiques are that the digipot may not operate in the desired voltage range and will also introduce capacitance which will affect stability.

A frequently suggested alternative is to use a DAC to inject current into the feedback pin. The DAC could be implemented with PWM from a microcontroller.

Thank you very much.

I have updated the schematic using DS4432 https://www.analog.com/media/en/technical-documentation/data-sheets/DS4432.pdf.

My calculations:
Vfb= Vout x R0B/(R0A+R0B)

R0A= 1.5 x R0B
Vfb= 1.21V
Vout= 18V

R0A= 13.88 x R0B
Assuming R0B= 10K, R0A= 138.8K

[mtwieg]

Injecting current into the feedback network via a DAC is indeed the better approach. But there are some issues with your schematic:

First of all, your buckboost converter won't work as shown in your schematic. See AN-1157 for an example of using the LM2673 in a buckboost circuit. This approach means that the FB will usually be negative (with respect to your GND), meaning you can't connect it directly to your DAC (unless that DAC is also reverenced to Vout, or some other interface circuitry is used).

Also it looks like your equations show the nominal resistor values to get Vout=18V, but nothing to do with the adjustment range.

[shamooooot]

Injecting current into the feedback network via a DAC is indeed the better approach. But there are some issues with your schematic:

First of all, your buckboost converter won't work as shown in your schematic. See AN-1157 for an example of using the LM2673 in a buckboost circuit. This approach means that the FB will usually be negative (with respect to your GND), meaning you can't connect it directly to your DAC (unless that DAC is also reverenced to Vout, or some other interface circuitry is used).

Also it looks like your equations show the nominal resistor values to get Vout=18V, but nothing to do with the adjustment range.

Thank you for the valuable input..

I can't seem to find how it can be done with the Buck-boost!, Would feeding the DAC output into an Op-amp shift the voltage?, but then I again need to invert the supply voltage to power the op-amp.

Regarding the equations:

for Vfb= 1.21 and Vout=18:
R0B=0.072 x R0A
100uA = IR0B - IR0A

When I solved the equations on Page 9 of the datasheet I got: R0A=155.5Ω and R0B=11.2Ω

[PGPG]

I am thinking of digital potentiometers,

I suppose they have poor total accuracy while more precise relative accuracy. So I think application should be designed the way to use them as potentiometers and not regulated resistors.

[shamooooot]

How about this?

[PCB.Wiz]

I am thinking of digital potentiometers, but I am not sure how to pick the one with correct parameters.

You should search for best tolerance, and lowest ppm/°C. A 20% pot is not much use for setting voltage.

Some have EEPROMs or proms to store a calibrated set point, as a default power-up value, you can then modify.

[shamooooot]

I am thinking of digital potentiometers, but I am not sure how to pick the one with correct parameters.

You should search for best tolerance, and lowest ppm/°C. A 20% pot is not much use for setting voltage.

Some have EEPROMs or proms to store a calibrated set point, as a default power-up value, you can then modify.

Thank you. Yes as suggested above I switched to DAC..

[mtwieg]

Thank you for the valuable input..

I can't seem to find how it can be done with the Buck-boost!, Would feeding the DAC output into an Op-amp shift the voltage?, but then I again need to invert the supply voltage to power the op-amp.

I don't think there's any simple approach to level-shifting a bidirectional current source.

Do you need the positive and negative outputs to have independently set voltages, or is it ok if they "track" each other (meaning they always have equal magnitude but opposite sign). If the latter, I think there are some tricks you can use to simplify things....

Regarding the equations:

for Vfb= 1.21 and Vout=18:
R0B=0.072 x R0A
100uA = IR0B - IR0A

When I solved the equations on Page 9 of the datasheet I got: R0A=155.5Ω and R0B=11.2Ω

The equations in the datasheet assume you have a nominal Vout (2V in their example), and want to adjust that by some percentage (+/-20% in their example). In your case, the middle of your Vout range is 9V, so you get R0A = 6.438*R0B.

To get a +/-9V adjustment range with a DAC current range of +/-100uA, you need R0A = 9V/100uA, so R0A = 90K.
Then you calculate R0B = 90K/6.438, so R0B = 13.98K.

[shamooooot]

Thank you for the valuable input..

I can't seem to find how it can be done with the Buck-boost!, Would feeding the DAC output into an Op-amp shift the voltage?, but then I again need to invert the supply voltage to power the op-amp.

I don't think there's any simple approach to level-shifting a bidirectional current source.

Do you need the positive and negative outputs to have independently set voltages, or is it ok if they "track" each other (meaning they always have equal magnitude but opposite sign). If the latter, I think there are some tricks you can use to simplify things....

Regarding the equations:

for Vfb= 1.21 and Vout=18:
R0B=0.072 x R0A
100uA = IR0B - IR0A

When I solved the equations on Page 9 of the datasheet I got: R0A=155.5Ω and R0B=11.2Ω

The equations in the datasheet assume you have a nominal Vout (2V in their example), and want to adjust that by some percentage (+/-20% in their example). In your case, the middle of your Vout range is 9V, so you get R0A = 6.438*R0B.

To get a +/-9V adjustment range with a DAC current range of +/-100uA, you need R0A = 9V/100uA, so R0A = 90K.
Then you calculate R0B = 90K/6.438, so R0B = 13.98K.

Thank you very much for helping me to understand it..

I actually want to control each rail separately, and what I understood is that my circuit won't work for the negative voltage..