I am designing a PoE PD with integrated flyback converter controller based on LT4295. I understand most of it but the ITHB Pin (current threshold pin for primary MOSFET) gives me headache. I do not fully understand how to design its external components.

In the datasheet, page 6, it states that the voltage of this pin corresponds to the peak voltage of primary MOSFET. So far so good, this means the max. peak current of the FET can be set via this pin. On page 15 it can be seen, that there is an RC network connected to the pin. The capacitor get charged up via the feedback pin, which sets the output voltage of the converter. The ITHB voltage is than sent through a linear amplifier to get VSENSE. VSENSE is then compared with the current sensed by a sense resistor and if the current sensed is greater than VSENSE, the FET will turn off.

This is what I don't understand: How are the RC components chosen to set the threshold voltage?

Any help on this topic would be appreciated.

  • 1
    \$\begingroup\$ The voltage at pin ITHN indirectly sets the peak current via the amplifier whose gain is given by the slope of Figure 1. The voltage ITHN is either coming from the OTA in a primary-side-regulated flyback converter (and in this case the components page 15 at pin ITHN form a type 2 compensator) or directly driven by the optocoupler in a active-clamp forward and all the compensation in this case takes place in the secondary side. The inversion is probably there to let you drive the pin via the opto emitter rather than the collector for a reason that I ignore though. \$\endgroup\$ Commented May 2, 2020 at 20:34
  • \$\begingroup\$ This is also what I understood. But what's the design criteria for the external components (RC network)? \$\endgroup\$ Commented May 2, 2020 at 22:35
  • \$\begingroup\$ You mean to design a type 2 compensator with an OTA when used in a primary-side-regulated flyback converter? I can write an answer if it helps. \$\endgroup\$ Commented May 3, 2020 at 9:33
  • \$\begingroup\$ Correct. This would be great and much appreciated. I read quite a lot on type 2 compensator but it's still hard to figure out how to calculate the needed components \$\endgroup\$ Commented May 3, 2020 at 10:55

1 Answer 1


The extra \$RC\$ elements connected to the ITHN are forming a type 2 compensator together with the internal operational transconductance amplifier or OTA. This configuration is adequate to implement a primary-side-regulated flyback converter for instance where there is a need for some regulation means in the primary side.

Before you tackle the calculations of the type 2, you absolutely need the control-to-output transfer function (TF) or the power stage ac response of your converter. That means if you stimulate the converter from the ITHN pin, how does the stimulus propagate through the control section and the power stage to generate the response observable across the output load? The best is to use a program like SIMPLIS to study this response. Examples on how to do that are available in the seminar I taught at APEC in 2018. Without this transfer function, you cannot attempt to stabilize the converter. The typical information you need from this TF are the gain (or attenuation) at the selected crossover frequency \$f_c\$ and the phase lag at this point.

Now for type 2 compensator, look at the below circuit I excerpted from my APEC seminar taught in 2010. You can see the output of the OTA connecting to the \$RC\$ elements. This OTA is affected by a transconductance \$g_m\$ and that is a design parameter as you can see in the formulas.

enter image description here

When you have the control-to-output transfer function you read the magnitude curve at the selected crossover which is 5 kHz for instance. Assume you read -15 dB. It means the type 2 compensator should be tailored to offer a 15-dB gain at 5 kHz to force a 0-dB loop gain at 5 kHz. 15 dB means a gain of 5.6 and this is parameter \$G_{fc}\$ in the formulas. Then, you have to position the zero \$f_z\$ and the pole \$f_p\$ based on the phase boost you need. Assume you read a phase of -70° at 5 kHz from your control-to-output Bode plot. This is \$Arg(H(f_c))\$. Then the phase boost generated by the type 2 converter for a 70° phase margin (PM) for instance is calculated as follows: boost= PM - \$Arg(H(f_c))\$ - 90°. In this case, the necessary boost is 70-(-70)-90 = 50°. From this value, you determine how to position the pole and the zero (all is in the seminar). If everything goes well, you obtain this transfer function for the type 2 section only:

enter image description here

Then you check the entire chain and simulate the loop gain which is the control-to-output transfer function cascaded with the response of the type 2. You should read the correct crossover and phase margin. It's only the beginning of the process and the stability analysis, you must now check all margins are safe at low- and high-line conditions, full and light load, tolerance analysis with Monte Carlo runs and so on. Quite a job if you want to do it seriously. You have a lot of data in the seminars that I linked but I wrote a complete book on the subject if you want to go deeper in loop control.

  • \$\begingroup\$ thank you for the detailed answer. Two questions: 1. Why do I need full transfer function if controller is typically only used at a set switching frequency? 2. Why aren't chip manufacturers already providing these transfer functions, or at least a detailed application guide so it is not so much work during design phase? I mean they already did a few reference schematics, why not showing the way how they designed it? \$\endgroup\$ Commented May 3, 2020 at 17:06
  • \$\begingroup\$ Well, the control-to-output transfer function is unique to the converter you design: it depends on the structure, on the selected components and their own characteristics etc. It is your duty, as a designer, to plot it either using equations, SIMPLIS, bench measurements etc. There are many application notes available on the subject and you can try to browse the manufacturer website for documents or simply contact them directly for assistance. Good luck! \$\endgroup\$ Commented May 3, 2020 at 19:02
  • \$\begingroup\$ I read through your slides and also other papers to get an understanding. Especially in this application note (nxp.com/docs/en/application-note/AN3279.pdf) the transfer function is described. However, if I understand correctly, this trasnfer function procedure cannot be used 1 to 1 on my design, as it depends on the internal structure and controll of the IC I am using, correct? \$\endgroup\$ Commented May 3, 2020 at 19:43
  • \$\begingroup\$ If your design is a flyback converter, the transfer functions they give work except that you need to add the modulation section brought by the selected controller. Besides, the recommendation they write at section where they select the RHP zero for crossover is absolutely wrong: in a CCM flyback (or DCM by the way) crossover is limited to 20-30% of the lowest RHP zero position which, in their example would be 3.6 kHz for a 18-kHz RHPZ. If you try to go above this value, it is impossible to maintain the phase margin. Is your project for a hobby application? \$\endgroup\$ Commented May 3, 2020 at 20:57
  • \$\begingroup\$ Okay so i take the open loop transfer function for CCM and DCM, which is available for flyback converters (like boost/buck topology transfer function just with consideration of turns ratio). Then I need to model my own modulation section, where I will seek advise from LT. There is a 24V reference schematic available that uses the same transformer, so i think my schematic will not be that much off. However, i really want to understand it fully. No this is not a hobby application. \$\endgroup\$ Commented May 4, 2020 at 7:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.