I’m trying to create my first synchronous boost converter using KiCad. This is my first time using this program as well, so maybe there’s something I don’t understand, but my signals make no sense. I’m thinking this is a problem with my setup of the unfamiliar program.
Some things of note:
I get errors such as doAnalyses: TRAN: Timestep too small; time = 1.0483e-10, timestep = 1.25e-17: trouble with node “xic2:probe_int_boot#branch” run simulation(s) aborted
As far as I can tell from this forum, this is a catchall error and gives me no real information. The weirdest thing is that my simulation will sometimes run, I will make a change that will cause this error, and upon undoing the change, the error will exist in the sim that ran fine at first. Sometimes in a different component. There’s no rhyme or reason to where the error will occur, which is why you see so many resistors in my sim. They are a stopgap fix, but I don’t know the root causes.
I’m using TPS28226DR gate driver, but using the library file for TPS28225 as this is the only one I can find Pspice code for and they are very similar in design and function.
Even when problem 1 doesn’t occur, I get a voltages across my entire circuit, even my bootstrap capacitor, and current in the unit of microamps.
What could I be doing wrong? Any help would be appreciated. I’m still learning this program.
I can’t share attachments on this forum yet, so here is a Google Drive link to all of the files needed to run this sim.
The files you put in your folder contain a user-specific path, so other people will not be able to use them as is.
(property "Sim.Library" "C:\\\\Users\\\\maste\\\\OneDrive - University of South Carolina\\\\Fall '25\\\\Research\\\\Phase 2\\\\slum496\\\\tps28225.lib"
Not being facetious here, but wouldn’t it be as simple as using the simulation model editor and changing the model path? I’m not sure how to share my files for others to use.
Yes, this is why I’m here. I’ve exhausted all my other options, because none of these links helped.
Simulation models may have their pins numbered differently than the corresponding symbol. For example, SPICE models for diodes usually consider pin 1 to be the anode, while schematic symbols are usually drawn with pin 1 as the cathode. Operational amplifier models are also very likely to have model pin assignments that do not match package or schematic pin numbers.
I’m a fan of 9.99, but it’s very difficult to do in that environment, and it’s very hard to keep it the same in both the situations where you have it and where you don’t. The easiest thing to do is to create a subfolder under the project, put it in there as a project specific symbol library, and then publish it, and other people will probably find it easy to reference.
project specific symbol library https://docs.kicad.org/9.0/en/eeschema/eeschema.html#managing-symbol-libraries
Thank you for the response. If I’m going to be honest, I was hoping you’d respond. You came the closest to answering my question in the similar posts I’ve found related to this problem.
I made all the changes you suggested after this line:
And while my simulation runs, it still looks nothing like yours. UGATE never actually turns on fully:
Sanity checking tells me that in a SMPS, LGATE should turn on first to charge the bootstrap cap. This leads me to believe that the pin assignment really is the last piece of the puzzle. How did you fix it on your end?
I followed the link @tama sent and the closest I could find in my version of KiCad is this, which looks fine to me.
Also now is a good time to find out that “tw” is the offtime of the signal, not the duty cycle. tw = 5.5 gets my my desired output voltage of 12V. Now I’m off to add an actual PWM chip and design my first PCB.
Hi quick question, my original circuit, a regular synchronous boost converter, worked in LT Spice with a 150 uH inductor. Why change this inductor to 20 uH? How did you calculate that value?
