TIDM-BUCKBOOST-BIDIR: Gain parameters estimation on a custom board

Hello Mohan,

The C1,C2,C3, BuckModeMaxGain and other parameters are determined empirically. This means, as you already guessed, you run the code in open-loop, and vary the gain parameter (BUCK_uk_pu) manually. You can tune the parameters as you pass between buck, buck-boost, and boost regions by adjusting the gain parameter.

Both reference designs operate in Constant Voltage mode. Constant current mode is not implemented. However, I believe you can modify the code to support CC/CV. You'll need to add a CC stage where the feedback is the output current, then transition to a CV stage where the feedback is the output voltage. The controlled variable in both cases is the gain. 

Hello Martinez,

Thanks for your response. I am little confused with CC/CV modifying of code. The addition of CC/CV feedback happen only while implementing open and closed loop controls right (step 2-5)? i.e The estimation of C1,C2,C3 etc still happens with existing open loop response code i.e adjust gain such that required output voltage is reached. Based on my understanding I would like to confirm the following and please correct me in any step:

1. Open loop response:  gain parameter as reference and adjust gain such that output voltage is just above battery voltage so that current flows into the battery (12V Lead acid battery as load). In my case Vout is battery kind of fixed and varying is input. Record values and obtain gain excel sheet to estimate C1,C2,C3 etc.

2. Open-Loop Current Control: gain parameter as reference and current as feedback. Adjust gain until 10A is charging current and run sfra to get Current Control Open-Loop Frequency Response. Then tune the parameters using compensator designer to obtain CC regulation.

3. Closed loop constant Current response: Close the loop with Iout parameter as reference and check the stability margins with sfra.

4. Open Loop Voltage Control: Add CV stage which sets the reference to Iout with feedback as Output voltage. Run sfra to get Voltage Control Open-Loop Frequency Response.Then tune the parameters using compensator designer to obtain CV regulation.

5. Closed Loop Current and Voltage Control: Run sfra to estimate the stability margins.

During Open Loop response (step 1):

In my case, my output is kind of varying between 12 (empty battery) -15V (full battery) and also input PV panel in between 22V- 8V. But I am fixing output voltage and varying input voltage for my gain excel sheet as shown in the below table (Following table is just as an example to convey my thoughts). . Is this right approach ? In the provided examples input is fixed and output is varying contrary to what I am planning to do.

Also, it would be great if I can have conversation with you either via teams or any support I can get in person here in TI Freising location. Cuz there are still some open queries .

Mohan Nagella said:

he addition of CC/CV feedback happen only while implementing open and closed loop controls right (step 2-5)?

No. In open loop, you manually and directly control the gain. There is no voltage or current feedback because the system is, well, in open loop. In closed loop, the compensator gets added and there is feedback. The gain is now controlled by the compensator which takes as input the error feedback. The compensator will control the gain to minimize the error as much as possible. It doesn't know (or care) if the error is in output voltage or output current! 

I think it is easier to think of your buck-boost power stage as a block box with a gain knob, VIN input, and VOUT output. If the gain is less than 1, then it operates a buck. If it is greater than 1, then it operates as a boost. The whole point of defining different operating regions and C1, C2, etc. parameters is to ensure there is a monotonic transition between buck and boost stages, i.e. you don't want the output voltage to swing from 10 to 20V when switching from buck to boost mode. You can derive these parameters by simply fixing the input voltage, varying the gain manually, and observing the output voltage. You can probably just keep the default values in the ref design firmware and test those.

Once the parameters are fixed then your power stage (buck-boost) will behave monotonically as you vary the gain and you can do testing in closed loop. My thought was that your firmware will have two modes of operation: Mode CC, closed loop design where the feedback is current and the goal is to vary the gain to keep the current constant, and Mode CV, closed loop design where the feedback is voltage and the goal is to keep the voltage constant. In your firmware you can define when to switch from CC to CV mode. Each mode would have its own compensator, so you can independently adjust the frequency response. 

Mohan Nagella said:

Also, it would be great if I can have conversation with you either via teams or any support I can get in person here in TI Freising location. Cuz there are still some open queries

Please reach out to your local FAE. They can pull somebody from the factory team as needed to any direct support. 

Hello Maritnez,

I got your points but my queries were regarding test setup with respect to my operating points. The example provided was for a resistive load with varying output voltage. In my case it is a 12V battery. I will be controlling the gain knob to switch between operating regions (buck, boost, bk-bst)

If we talk only about it open loop, the  during the open loop response do I need to pay attention to how much current is flowing inside the battery while setting the output voltage with the gain knob? I am asking this because in TIDA-10090 in open loop response especially they adjust duty to a fixed current flowing (section 3.4.2.3 line 6.) inside the battery as it is going to be CC regulated.

Or in this case I just focus on output voltage and dont really think of the output current and just adjust gain knob until my required output voltage is reached note it down and continue the process with other voltage settings at the output.

Admittedly, battery charging is out of my area of expertise. My main point was that the C1,C2,C3, BuckModeMaxGain and other parameters should be independent of whether you are regulating current or voltage. The gain essentially gets fed into a function that tranforms that into buck and boost PWM duty cycles (highlight in red below). The feedback loop (or loops) are related to the battery charging application itself. 

Looks like your field rep is trying to set up a call to discuss this in more detail. We can make sure somebody with the right background in battery charging applications is looped in.