LTSpice Synchronous Buck+Boost Converter

[CamNielsen]

Trying to get a simple mock-up of a synchronous buck+boost converter in LTSpice - I have working models of a synchronous buck and synchronous boost converter but I am quite the novice with power electronics and the union of the two is giving me some trouble. (The capacitor/inductor values are placeholders for the time being until I get a working model - for the purposes of the model I am setting the Vin to 12.6 for the first ms and to 9.6 for the second ms)

Here is the synchronous buck converter (goal is to step down 12.6 V to 5 V primarily):



Here is the synchronous boost converter (goal is to step up 9.6 V to 20 V):



I figured connection the two in series would give me what I wanted as shown below:



But alas, it was not meant to be - instead of the combination of the above waveforms I get something quite different and the analysis hits a brick wall and starts chugging hard once the voltage switches to the 9.6 volts (boost mode). I realize the circuitry isn't identical but it seems frivolous to have inductor capacitor capacitor inductor between the two circuits.



Any idea why my two functional circuits have created this demon spawn? (I attached my LTSpice files below for reference)

 

[FvM]

Did you look closer at the switching behavior of both circuit blocks, e.g. drain currents? There is terrible shoot-through due too low dead time respectively too slow gate voltage.

In the buck simulation, you are triggering converter self destruction by feeding the output with 20V after 2 ms.

I would
1. verify the operation of buck and boost block with reasonable test cases, adjusting the parameters to get clean switching waveforms
2. use a well considered test scenario for the combination circuit

Connecting a synchronous converter with voltage sources at in- and output and operating it with a fixed duty cycle without feedback control doesn't look like a useful test scenario to me.

 

[CamNielsen]

Did you look closer at the switching behavior of both circuit blocks, e.g. drain currents? There is terrible shoot-through due too low dead time respectively too slow gate voltage.

In the buck simulation, you are triggering converter self destruction by feeding the output with 20V after 2 ms.

I would
1. verify the operation of buck and boost block with reasonable test cases, adjusting the parameters to get clean switching waveforms
2. use a well considered test scenario for the combination circuit

Connecting a synchronous converter with voltage sources at in- and output and operating it with a fixed duty cycle without feedback control doesn't look like a useful test scenario to me.

Thank you for the reply FvM - I simplified the test case for all models to include an input voltage source (constant 11.1 volts), and an output load resistor (100 ohms). I'm trying to verify operation at discrete points before adding feedback control at the moment but I think it is a great point that maybe its a bit much to untangle right now. I also made the following changes to attempt to improve performance of the circuit:

-Increased dead time from 100ns to 500ns to prevent interactions between components per your recommendation
-Increased gate resistor size and added Shotcky diodes in parallel to improve driver performance

I also did this for the individual Buck and Boost converter I designed to help my understanding of where the issue may be. The waveforms for these two converters (along with the drain currents of each MOSFET) is shown below:

Synchronous Buck
**broken link removed****broken link removed**

Synchronous Boost
**broken link removed****broken link removed**

These seem pretty reasonable - a bit slow to converge but they are completing the intended task of converting the 11.1 input voltage to 5 and 20 V respectively. The combined circuit is still exhibiting similar behavior unfortunately. The drain currents are quite worrying in the combined circuit as they do not behave appropriately whatsoever. The first waveform shows Vout along with the drain currents of the two Buck MOSFETs and the second shows Vout along with the two Boost MOSFETs

Combined Circuit
**broken link removed**
**broken link removed****broken link removed**

Doesn't look good for the transistors

My control signals are (1.78V,4V) for control 1 and (0V,1.8V) for control 2 which is the duty cycles from the individual circuit and an always on/off signal for the inactive circuit. Each phase lasts 1.5ms so it should switch from Buck to Boost mode in the middle of the waveform.

 

[FvM]

The first attachments for buck and boost converter seem to be lost on the server. This happens sometimes, unfortunately.

 

[CamNielsen]

The first attachments for buck and boost converter seem to be lost on the server. This happens sometimes, unfortunately.

Here are all of the images again in case they are helpful:

Buck -

Boost -

Combo -

 

[treez]

I must admit with the boost for example, I noticed large current spikes for a particular period……though observing the gates at this period revealed gates not on at the same time……so I reckon the shoot through is from a dodgy diode within the fet…perhaps use an external schottky to solve.
I noticed the boost inductor current going backwards after a while so there is some problem there…..what load is the boost on? ……why is the boost converters output staying up when current is being drawn out of it and sent back to the input?...have you put a voltage source at the boost output?
With the buck too I would put parallel schottkys across the fets
Maybe increase the boost output cap and/or buck input capacitor to “decouple” there feedback loops from each other….as you know, if insufficient capacitor bank between two series converters then there feedback loops can fight with each other….unless the feedback loops have widely different bandwidths…then you wouldn’t need the cap bank………….

Consider a boost pfc followed by a buck….that is ok because the pfc has a ridiculously low bandwidth………….much lower (likely) than the following buck

 

[FvM]

There's a weird copy-and-paste fault. You have hidden shorts in your combined circuit by using the same labels in buck an boost section.

- - - Updated - - -

O.K., the nets haven been renamed now.

 

[CamNielsen]

There's a weird copy-and-paste fault. You have hidden shorts in your combined circuit by using the same labels in buck an boost section.

- - - Updated - - -

O.K., the nets haven been renamed now.

...oh dear.

I have spent probably 5+ hours trying to sort through this and a mis-label has been killing me the whole time.



A lesson learned I suppose - the waveforms are still a little slow to reach their steady-state waveform if I am being picky but its good to see what I would expect from the circuit for the first time.

Thanks for the keen eye on that one! I had been buried in the circuitry for so long that I completely missed that issue.

 

[treez]

Posting the situation to this forum, probably straightened it out in your head, and then the probelm became easier for you to diagnose. Done the "double-label" trick many many times myself.