In the schematic of the BoostXL-3PhGaN from Texas Instruments, in the current sensing section, there is this output filter that produces the VA signal that will be sent to the microcontroller and you can plot it with oscilloscope etc.


In addition to understanding why it is necessary, I took the opportunity to do some hand and simulated calculations.

Could you kindly check whether or what I did wrong in these calculations?
sorry, C is 33nF !!!!


I would like the calculated data (especially the cutoff frequency and Gdc) to match those simulated on LTSpice

SOLUTION, ERROR FOUND:

 

hi k89,
Your handwritten equation you have the Cap as 333m and the LTS its 33n ???
E

 

hi k89,
Your handwritten equation you have the Cap as 333m and the LTS its 33n ???
E
View attachment 345397

Sorry! It's 0.033u

 

The cutoff frequency will be around Fc = 0.16/(R56||R59 * C1) = 1.2kHz.
And the DC gain is R59/(R56 + R59)

 

hi k89,
Your handwritten equation you have the Cap as 333m and the LTS its 33n ???
E
View attachment 345397

The cutoff frequency will be around Fc = 0.16/(R56||R59 * C1) = 1.2kHz.
And the DC gain is R59/(R56 + R59)

Sorry, I saw now the error in the calculations (just posted updated)

 

In the schematic of the BoostXL-3PhGaN from Texas Instruments, in the current sensing section, there is this output filter that produces the VA signal that will be sent to the microcontroller and you can plot it with oscilloscope etc.
View attachment 345392

In addition to understanding why it is necessary, I took the opportunity to do some hand and simulated calculations.

Could you kindly check whether or what I did wrong in these calculations?
sorry, C is 33nF !!!!
View attachment 345390

I would like the calculated data (especially the cutoff frequency and Gdc) to match those simulated on LTSpice

SOLUTION, ERROR FOUND:
View attachment 345441

Hi,

I could not read your handwriting very easily but I did get a similar result of 1191.1Hz. I usually use 16 digits of precision throughout the calculations unless it calls for better.

 

Hi,

I could not read your handwriting very easily but I did get a similar result of 1191.1Hz. I usually use 16 digits of precision throughout the calculations unless it calls for better.

Can I ask you why there is that filter at 1kHz?
Consider that Va (filtered) will be sent to a connector that will be connected to another board containing the uController.
What do you think is the reason?

 

hi k89,
The filter is to reduce the switching noise, refer to this PDF
E

 

Can I ask you why there is that filter at 1kHz?
Consider that Va (filtered) will be sent to a connector that will be connected to another board containing the uController.
What do you think is the reason?

Hi,

As Eric pointed out, it helps to reduce noise due to any switching or irregularities in the armature.
It also averages the output which is not always possible to do in a microcontroller using digital techniques, and attenuates the signal so the microcontroller can handle it.

In many cases we want to regulate the average performance not the instantaneous. Here it looks like it may be part of a speed controller?

 

Hi,

As Eric pointed out, it helps to reduce noise due to any switching or irregularities in the armature.
It also averages the output which is not always possible to do in a microcontroller using digital techniques, and attenuates the signal so the microcontroller can handle it.

In many cases we want to regulate the average performance not the instantaneous. Here it looks like it may be part of a speed controller?

I think it is definitely strong of a speed controller since it is the current that is elected by the inverter and definitely used within the algorithm