0
\$\begingroup\$

For educational reasons, I want to build physical DC-DC converters. I would start with simple buck or boost topology. For start, I wish to have a simple gate-drive for switches, without PWM modulation. After successful build, I would like to be able to include a simple control circuit with PWM modulation an topology I built.

Can you recommend any tutorial that would also ease components selections? Or, do you have good recommendation for gate-drivers and PWM modulation ICs that are simple enough for first-time builders? (it doesn't necessarily needs to be a buck or boost converter topology)

Thanks.

\$\endgroup\$

closed as off-topic by winny, Andy aka, Voltage Spike, Dmitry Grigoryev, Lior Bilia Jul 31 '18 at 14:49

This question appears to be off-topic. The users who voted to close gave this specific reason:

If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    \$\begingroup\$ For educational purposes: then why not simply use a circuit simulator? Sure you can try and build / design a DCDC converter from separate components but it will be a challenge to get it working properly. In a simulator you will have much more control to change whatever you want and also it will be much easier to see any signal (current/voltage) you want without influencing the working of the circuit. Also, read a book about DCDC converters. \$\endgroup\$ – Bimpelrekkie Jul 29 '18 at 21:08
  • 1
    \$\begingroup\$ I already have a theoretical background about DCDC converters. I would like to build an actual circuit. I would simulate my circuit in program like LT Spice before building it. \$\endgroup\$ – DenR Jul 29 '18 at 21:22
  • \$\begingroup\$ Fairchild, now part of ON semiconductor, makes tons of gate drivers for both high and low side gate drive. If you are doing this for instructional purposes, maybe you could sens voltages using ADC's on a micro. Same for currents. Use a shunt and shunt amplifier. \$\endgroup\$ – mkeith Jul 30 '18 at 3:56
  • \$\begingroup\$ Well if it's simple, then why not build one that doesn't need a gate driver? I've used a logic level MOSFET in a boost converter to create +12V for signalling, I'm sure it was inefficient but I didn't need much power and it worked fine. \$\endgroup\$ – immibis Jul 30 '18 at 5:12
2
\$\begingroup\$

What you're wanting to start work on is along the lines of what I'm working on, so allow me to share what I've learned. Thus far, I've been exclusively investigating buck topology, and working on a system that allows me both to control output voltage and drive that output voltage as a PWM. I definitely found breadboarding some components and playing around accelerated my learning process.

So, if you are looking to efficiently step down voltages, you'll need a few components, and there are various properties of them you'll have to weigh against each other, so this is a bit of what I've been learning and looking at:

Switch(es):

You can use transistors or mosfets at lower input voltages, and mosfets and IGBTS as you get into higher voltages and currents. I like mosfets for my LED and/or small motor related plans. If you're looking to measure or control the voltage after it's been stepped down,(for anything but a static load you should, and probably should anyways) keep in mind that if you switch it on the P side, you will get a voltage off the N rail, and vice versa.

The lower the voltage drop over your switch when it's ON, the better. As far as I've seen transistors and diodes are rated in Vf at their rated current in the on state, and mosfets are rated in On resistance.

The faster your switch can switch, the better. It will waste power by simple conduction, based on it's on resistance, and in switching losses, which are the bits of power that are lost by the switch when it is between on and off states. A faster switch will let you run at a higher frequency and use a less powerful inductor and output capacitor.

I'm not great with transistors, but mosfets can be quite nice. The gate on the Mosfet is a capacitor, and charging and discharging it switches the mosfet. Generally speaking a mosfet with lower ON resistance has higher capacitance. You can get all kinds of gate drive voltages for mosfets, right down to ones that can be powered off of 3.3v or 5v logic. P side mosfets tend to have much higher capacitance for the same current carrying ability. The gate capacitance and gate voltage together determine how much energy it takes to switch the mosfet.

You can switch a Mosfet faster or at a different voltage than your logic circuit by using a mosfet driver IC like the MIC4426/7/8. I've not played with these yet, but when they arrive I'll do so. I understand you can also get mosfet drivers to drive N channel mosfets on the P-side of the circuit by producing a high enough gate voltage. I've not tried any, but they appear to have lower drive currents than similarly packaged normal mosfet drivers, but the lower capacitance of the N mosfets I think is used to compensate for that.

LC filter

If you don't want to waste too much power, rather than a RC filter, you'll want to make an inductive/capacitive filter with it's resistance as low as possible. The lower the frequency you use, the more inductance you need, which means either a larger inductor of the same resistance or an equal sized one with more resistance, so the higher the frequency, the better. Disadvantage to this is that the higher the frequency the more you have to worry about accidental antennas and keeping your circuitboard traces short.

Control mechanism:

You'll need something to actually control the switch. You can make this as simple or complicated as you want, but I'd suggest generating a clock frequency that you can reasonably switch at, and a circuit to switch on when it receives a clock pulse, and back off when the output voltage reaches a reference voltage. You can do this fairly easily with a 555 timer, another 555 timer or a flip flop or a latch or a few logic gates, and a comparator. You need something to generate the reference voltage, so a 7805 and a voltage divider or a 78l05 and a voltage divider would do it, or a zener circuit.

You can also buy ICs designed to accomplish this purpose, and simply provide them with the other circuit components, like an inductor and capacitors or resistors that they need. If the IC contains the switch to be used, it is called a voltage regulator, and if it drives an external switch, it will be a voltage controller.

Asynchronous parallel drive:

The above will get you far enough to breadboard some useful DC to DC converters, and so far as I've seen the next logical step is an asynchronous buck converter, similar to what a motherboard uses. I'll take going from one phase as described above to four phase as an example.

The control mechanism becomes more complicated as you'll want to run the four separate channels at the same frequency, but offset 90 degrees from each other.

The advantages are pretty great. Assuming you used 4 of the same inductor you would have above and kept the frequency the same, you'll end up with 1/4 the peak to peak voltage in your output, 1/4 the disturbance on your input voltage. You don't have to keep the parts of the circuit the same though, and because you've split the necessary current into four, you can use a switch with 1/4 the current rating, and the switch and inductor can add up to four times the resistance and still get about the same losses.

Thing about mosfet switches is that at a higher resistance, they tend to have lower gate capacitance and you can switch them faster. That means you can use a higher frequency at the same switching losses, which means you need less inductance for each of your (now 4) inductors, which means they can have less resistance at the same physical size.

If you just want to start messing around, I'd grab a solderless breadboard, some voltage source (I use a battery pack because my variable DC supply isn't ready yet) a few basic voltage regulators like 7805s, 7905s and 7812s, 7912s, some alligator clips, a mixed bag of through hole inductors up to maybe a few hundred uH, some 555 timers, D and J/K flip flops, a few logic gates, comparators, some N and P channel 5v and 10v Vgs mosfets, maybe some MIC4426/7/8 drivers, mixed kit of resistors, capacitors. Then maybe some high power LEDS or 5V LED tape to use as practice loads. If you want you can get a lot of this by desoldering old electronics, particularly resistors, capacitors, inductors and high power leds. A multimeter helps of course, but getting a crappy 30-50 shmeckel oscilloscope made a huge difference, just to be able to see things like the frequency at which my output voltage stabilizes, etc.

\$\endgroup\$
  • 1
    \$\begingroup\$ Wonderful. Some very good advice, and very nicely written. I was looking for something like this. Thanks. \$\endgroup\$ – DenR Aug 1 '18 at 1:30
  • \$\begingroup\$ I'm just about ready to start asking my own questions regarding my project, so getting some of this straight in my head first is helping me =) Thanks \$\endgroup\$ – K H Aug 1 '18 at 1:39
0
\$\begingroup\$

Or, do you have good recommendation for gate-drivers and PWM modulation ICs that are simple enough for first-time builders?

I've used one of these in a buck converter: -

enter image description here

It doesn't have to run at 1 MHz - you can set the frequency down into the hertz region. It's a good building block.

\$\endgroup\$
  • \$\begingroup\$ That is pretty cool! \$\endgroup\$ – mkeith Aug 4 '18 at 6:00
  • \$\begingroup\$ Can't upvote you because of my reputation, but it I will probably give a try to this IC. THanks! \$\endgroup\$ – DenR Aug 4 '18 at 14:53

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