LT1242 Current mode control help

I am trying to implement current mode control in my buck converter using the LT1242 device. The reason I am using this rather than a buck regulator/synchronous buck device is because my input voltage is too high for the ratings typically. 

Anyway, I am having a few issues with the design. In the data sheet, https://www.analog.com/media/en/technical-documentation/data-sheets/lt1241.pdf, there are some examples of circuits using this device. I have changed slightly the flyback converter example slightly, I will attach a schematic to this post. 

My first questions is whether it is possible to implement both soft-start and slope compensation at the same time. My buck converter will spend most of the time above 50% duty cycle so therefore I will need slope compensation to ensure stability - but I would also like a soft start time to allow my output to reach the desired setpoint. It looks as though they are exclusive, because unfortunately the two functions use the same input pins on the chip - is there any other way I could implement both these functions within the same chip, or not?

Secondly, in the slope compensation where we use an NPN BJT that connects to the Isense pin through a resistor - how do we calculate this resistor value? There doesn't seem to be any equations for selecting the desired R.

When selecting RT/CT - is this purely a case of looking at the graphs and selecting both an RT and a CT that will give the right operating frequency? I wish to operate the LT1242 at around 500kHz, so have chosen 20kOhm and 100pF from the data sheet, but how do I know whether these values are optimized? 

I also have an issue with my Isense voltage sawtooth that is passed in to the current amplifier and compared with the error amp output. It is not symmetrical, and at times does not look what I expect. Maybe this is due to incorrect values of RT/CT or something, but I am sure someone more knowledgable than myself can input on this. 

The reason I have a buck converter and this chip is to regulate the DC bus voltage to a fixed frequency resonant half bridge converter. Therefore, the output voltage of the device is not fixed any will vary. How do I go about deciding on a maximum peak current etc in this case? I understand if we have a fixed output voltage our max peak switch current is fixed - however, if the output voltage varies, how can we choose a fixed value for max switch current for the regulator?

Finally, in all the designs the current sense must be taken from the switch node to ground. Is there a way to sense current through the buck output inductor instead, or elsewhere where the current sense voltage will not be referenced to GND?

  • 0
    •  Analog Employees 
    on Nov 15, 2019 1:33 AM over 1 year ago

    My first questions is whether it is possible to implement both soft-start and slope compensation at the same time. My buck converter will spend most of the time above 50% duty cycle so therefore I will need slope compensation to ensure stability - but I would also like a soft start time to allow my output to reach the desired setpoint. It looks as though they are exclusive, because unfortunately the two functions use the same input pins on the chip - is there any other way I could implement both these functions within the same chip, or not?

    Comment: Per page 13 of the LT1241 Data Sheet, soft start uses the COMP and VREF pins and per Data Sheet page 15, slope compensation uses the RT/CT, VREF and ISENSE pins. I believe both can be implemented together.

    Secondly, in the slope compensation where we use an NPN BJT that connects to the Isense pin through a resistor - how do we calculate this resistor value? There doesn't seem to be any equations for selecting the desired R.

    Comment: With slope compensation, one is trying to add a ramp across the resistor in series with the ISENSE pin. You can adjust resistor values in LTspice and measure the slope directly and/or solve the equations for the voltages and two resistors. This simple slope compensation circuit is not ideal and affects the current limit. There are many, many newer ICs that include slope compensation and soft start pins. Without knowing you application (input voltage range, output voltage and current) it is difficult to recommend an appropriate IC.

    When selecting RT/CT - is this purely a case of looking at the graphs and selecting both an RT and a CT that will give the right operating frequency? I wish to operate the LT1242 at around 500kHz, so have chosen 20kOhm and 100pF from the data sheet, but how do I know whether these values are optimized?

    Comment: There are equations in the Data Sheet for oscillator frequency or you can play in LTspice. Smaller resistors reduce maximum duty. Smaller capacitors increase frequency.

    I also have an issue with my Isense voltage sawtooth that is passed in to the current amplifier and compared with the error amp output. It is not symmetrical, and at times does not look what I expect. Maybe this is due to incorrect values of RT/CT or something, but I am sure someone more knowledgable than myself can input on this.

    Comment: I would first verify it is real by limiting the maximum timestep on the .TRAN statement.

    The reason I have a buck converter and this chip is to regulate the DC bus voltage to a fixed frequency resonant half bridge converter. Therefore, the output voltage of the device is not fixed any will vary. How do I go about deciding on a maximum peak current etc in this case? I understand if we have a fixed output voltage our max peak switch current is fixed - however, if the output voltage varies, how can we choose a fixed value for max switch current for the regulator?

    Comment: Isense sets the peak inductor current (reduced by slope compensation). The peak inductor current is determined by max load current and half the inductor ripple current. See Application Note 19 (AN19) on Analog's website.

    Finally, in all the designs the current sense must be taken from the switch node to ground. Is there a way to sense current through the buck output inductor instead, or elsewhere where the current sense voltage will not be referenced to GND?

    Comment: If the load is an isolated supply, you could consider a buck with a low side switch and output referenced to Vin). High-side current sense is not an easy task depending on your input or output voltage. High-side sense amplifiers exist but often have limited bandwidths. Current sense transformer's work well (on main switch) if duty cycle is not required to go too high.

  • Hello Kurk,

    First of all thank you kindly for your answer, it is very insightful and will help me. With the peak inductor current however, I think maybe my point was missed. I understand when you have a buck converter that has a fixed output voltage, the peak inductor current will be fixed. However, with my design, the output voltage of the buck varies, to regulate the DC bus voltage to the fixed duty and frequency resonant bridge. In this case, how would one define the maximum peak inductor current to regulate in this fashion? 

    My buck converter has an input voltage of 243-297VDC, 270VDC nominal. This is why I've found it difficult to find an IC that I can use for this purpose. The output voltage and current, again, will be variable depending on the load on the secondary side of the transformer. Could you recommend any techniques for this or as you say, a more appropriate IC for this purpose?

    Thank you!

    Joel

  • 0
    •  Analog Employees 
    on Nov 22, 2019 2:30 AM over 1 year ago in reply to extraeng