# SMPS: Non-isolated DC-DC Buck Converter

#### KhaledOsmani

##### Full Member level 6
Hi

Considering the following interleaved DC-DC buck converter:

I've tried to built the above circuit via Proteus, and fed HIN and LIN of the two IRS2003 MOSFET drivers by a configurable function generator.

The PV+_DWN of the two sub-divisions are fed from a 24VDC supply (this is assumed theoretically that a PV panel is outputting a constant 24VDC).

Regardless the option of choosing an interleaved non-isolated buck converter, but, with a 50% PWM (from configs of pulse generator) isn't the output voltage must be seen at 12VDC?

After branching a DC voltmeter across VOUT_UP Im seeing non-stable voltage values maximizing at +7.03V.

What am I missing here?
Khaled

#### KlausST

##### Super Moderator
Staff member
Hi,

What am I missing here?
You missed to post your simulation conditions and simulation results.

Klaus

KhaledOsmani

### KhaledOsmani

points: 2

#### KhaledOsmani

##### Full Member level 6
Hello Klaus,

Nice hearing from you. I believe the simulation conditions/results were typed. Here's a graphical interpretation:

1) For Part A: the equivalent circuit on Proteus is as below:

2) For Part B: the equivalent circuit on Proteus is same as in Part A (the choice for an interleaved buck converter was meant to reduce ripple currents).

3) For Part C: the equivalent circuit on Proteus is as below:

>> Now for the simulation, I must see across DC voltmeter a value of 12V based on 50% PWM for the 24V of the input, but this is what I get:

As for L,C parameters calculation, the following link helped in calculating required values:https://www.onsemi.com/pub/Collateral/AND9135-D.PDF

The project is to later on be PWM supplied by Arduino according to a certain algorithm. This circuit configuration was only a test bench for the converter efficiency. Why is the output voltage 6V not 12V?
Khaled

##### Super Moderator
Staff member
Looking at your first schematic, it looks as though your high-side N devices Q1 Q3 have a 24V supply. However you're gating these with 12V signals. It's not sufficient bias to turn on those mosfets.

The gate needs 24V or more to turn on.

points: 2

### Easy peasy

points: 2

#### KlausST

##### Super Moderator
Staff member
Hi,

At least I now know that you yse 15kHz and 1.34mH inductance....

But why no scope pictures? If you expect 12V and see something else you need to look for details ... a voltmeter simply is not suitable to find the cause.
We need to see ringing, phase alignments, signal levels...
Both input signals, both bridge output signals, gate signals...

Your PV supply needs HF decoupling capacitors and a bulk capacitor ... at least in the real circuit.

KhaledOsmani

### KhaledOsmani

points: 2

#### KhaledOsmani

##### Full Member level 6

I increase the MOSFET driver, Vcc to +24VDC instead of +12VDC. Output was more convenient (+11.2VDC). But what is the relation between the MOSFET driver supply circuit and output voltage coming out of a PV panel?

One thing left, the input/output relation (output of buck converter voltage vs. input voltage coming from PV module that is here fixed at 24VDC) is dictated by the following equation:

Vo/Vin = D
Where D is the duty cycle. Assuming a 100% converter efficiency, when D is (manually in the pulse generator) is set to 30%, the output voltage must be equal to 7.2VDC, whereas in Proteus simulation it starts by increasing from 0 gradually to around +9.8VDC. What is meant here is that
Vo/Vin = D
not 100% satisfied. Any reason?

Thanks
Khaled

- - - Updated - - -

Hi again Klaus,

Sorry for the imbalanced pictures. It's been a while being off this forum, and the PrtSc/Paste seems ineffective.

Please take a look upon this modified circuit: https://imgur.com/OEiSbTi (hope it uploaded successfully).

I've fed both MOSFETs with 180 phased PWM signals (this is to ensure that both MOSFETs are never mutually ON/OFF, hence spare away an act of short-circuit).

The circuit worked perfectly with Vo = Vin x D.

The previous circuit has a problem with the MOSFET driver. What kind of drivers do you recommend using when willing to output the PWM signals from Arduino pins?

Khaled

#### KlausST

##### Super Moderator
Staff member
Hi,

whereas in Proteus simulation it starts by increasing from 0 gradually to around +9.8VDC.
A third time: show us such simulation results. Until now we see no single chart and no scope picture...
Textual description is not suitable to find the problem.

Klaus

KhaledOsmani

### KhaledOsmani

points: 2

#### Easy peasy

Also a boot-strap high side driver IC needs the low side to be gated on for ~ 10uS to allow for charging of the boot strap upper driver cap ...! only 12V needed for that - from the input LV side ...

##### Super Moderator
Staff member
what is the relation between the MOSFET driver supply circuit and output voltage coming out of a PV panel?
Notice you want 24V to be at the bottom leg of Q1 as it conducts fully. However since it's an Nmos it can never be higher than the gate voltage. Hence it never is driven to conduct fully.

Similarly if you used a Pmos, it can never turn off entirely if gated by 12v. However it's possible to make it work by attaching a resistor to supply V (24V). Then pull the gate low via Nmos to ground. 12Vv is sufficient to turn on the Nmos gate.

Or, generate 24V for the high-side Nmos by making a bootstrap circuit recommended in post #8.

KhaledOsmani

### KhaledOsmani

points: 2

#### KlausST

##### Super Moderator
Staff member
Hi,

The circuit of post#2 show a MOSFET driver IC IRS2003 which includes a bootstrap circuit to provide sufficient gate voltage to fully turn ON the high side MOSFET.

Picture2 of post#3 also shows a MOSFET driver with bootstrap feature.
But IR2101 here is not wired correctly, because it has a different driving scheme than IRS2003. It can´t work this way.

Which circuit the OP uses is a riddle to me. OP´s informations are incomplete and confusing to me.

Klaus

KhaledOsmani

### KhaledOsmani

points: 2

#### KhaledOsmani

##### Full Member level 6
But IR2101 here is not wired correctly, because it has a different driving scheme than IRS2003. It can´t work this way.
Hi, How to wire the IR2101 instead of IRS2003?

- - - Updated - - -

Any available library for IRS2003 model?

#### KlausST

##### Super Moderator
Staff member
Hi,

one IC has LIN inverted the other not. You can´t drive them equally.

What library do you need?

Every semicondutor manufacturer provides datasheets, design examples, typical circuits as well as complete application notes and design notes with detailed explanations and calculations.

Use a simulation tool or not, in either case it is very important to read throgh datasheets and understand the operation of the circuit.
A datasheet tells all you need to know and it is more reliable than a simulation.

Klaus

KhaledOsmani

### KhaledOsmani

points: 2

#### KhaledOsmani

##### Full Member level 6
Your PV supply needs HF decoupling capacitors and a bulk capacitor ... at least in the real circuit.
Hi

Is this what you mean (C1,C2,C3): https://imgur.com/l6pzKC0

For the MOSFET driver, can you re-arrange circuitry of post #3 to accomodate IR2102 instead of IRS2003 (since the IRS2003 library isn't available for Proteus)?

- - - Updated - - -

one IC has LIN inverted the other not. You can´t drive them equally.Klaus
Yes Klaus, I read the datasheet of IR2102 but the problem is that Im unable to adjust the wiring of post #2 for that driver. Can you help rearranging this circuit:

https://imgur.com/IwB0pHW

#### KlausST

##### Super Moderator
Staff member
Hi,

can you re-arrange circuitry of post #3 to accomodate
No. I don´t use proteus. I use simulators very rarely at all.

PC capacitors:
Basically yes. But use capacitors that are suitable for your switching frequency ... plus their overtones. Many electrolytics are not fast enough.
You may connect faster capacitors in parallel to medium fast bulk capacitors.

Klaus

×