Is this the BEST Voltage Converter? Trying to build a Synchronous Converter!

This is my Synchronous voltage converter that I built in the past week. Now you might be asking yourself what makes it so special since you can already get switched mode voltage converters like buck and boost converters everywhere on the internet. And by the way for those just starting out with electronics such converters can take a DC input voltage and convert it efficiently into a higher or lower DC voltage that we can then use. But getting back to the question what makes this synchronous one so special and to answer that we can have a look at this beautiful overview chart from Würth Elektronik of almost all switched mode power supply typologies. As you can see their efficiency is around 85 to 90% at max which is not bad but a synchronous converter can supposedly reach 95% efficiency which actually sounds to good to be true. That is why I built my own to measure its efficiency but before I do that why don't we go back in time together so that I can show you how exactly I made mine and explain why this converter type is quite a bit more complicated to build than you might think. Let's get started! This video is sponsored by JLCPCB which is a PCB manufacturer that I can truly trust. I actually ordered the PCBs for this project from them which was super simple to do by uploading my Gerber files, selecting my PCB preferences, paying very little money and waiting for a week and just like that I got beautiful purple PCBs that you can also get if you try out their service. First off I had to decide what kind of topology I wanted to transform into a synchronous one and I went with the buck converter because in the case of a switch fault we either only get the full input voltage or nothing at all at the output which is great for testing. In comparison if a boost converter would act faulty we could get a rather high destructive DC voltage on the output which I wanted to avoid for my first tests. But anyway, next we need to understand how the circuit works and if we simplify it as much as possible then we could say that the switch creates a PWM voltage with a high frequency that gets smoothed out and thus lowered by an LC low pass filter. It does not get simpler then that but to be more precise and accurate let's imagine the switch really gets controlled by a high frequency PWM signal and right now it is closed. Now current flows through the circuit and supplies the load while charging up the capacitor and building up a magnetic field around the inductor which by the way opposes the flowing current and thus only let it rise slowly. This continues until the Switch opens at which point the inductor uses it electromagnetic field energy in order to let current flow the same way as before by now acting as our energy source. This is possible because our Diode which blocked the current flow before now becomes part of our current loop. But of course the current is this time linear decreasing and depending on a couple of factors the current can either stay above zero before the switch once again closes or fall to zero before that at which point the capacitor has to supply the load for a short time. But either way depending on the utilized duty cycle the output voltage can be fine adjusted between 0V and the supply voltage. So all in all the circuit itself even when converting to real electrical components as well as the voltage and current waveforms are not that hard to grasp, right? I thought so as well which is why next I started to create my own DIY version on a perfboard in order to show you exactly where energy gets wasted. What I didn't expect though was that my first circuit idea with an N-Channel MOSFET and Gate-Drive Transformer didn't feel like working with me for mysterious reasons. So after 3 hours of troubleshooting I gave up and instead went with this more straightforward P-Channel MOSFET alternative according to this final schematic which once again proved that everyone can have a bad day with their projects, no matter the experience. And after writing a bit of code for the Teensy microcontroller in order to create an 80kHz PWM signal whose duty cycle I can change with a potentiometer, it was time to connect a 100uH inductor to the circuit, attach a 4.7 ohm load to the output, power everything with 12V and have a look at the oscilloscope while playing around with the duty cycle. As you can see the previously discussed theory pretty much corresponds with the practical application and thus we can fine adjust the output voltage by simply changing the duty cycle. So next we have to find out how this circuit can be more energy efficient by once again having a closer look at the main components. During the first switching state we only got the MOSFET, inductor and capacitor which all have a very low losses so there is not much to improve. During the second switching state however, current flows through the diode whose voltage drop we can measure in the practical circuit. As you can see it is around 200mV peak so around 100mV on average which multiplied by the practical average current that is around 400mA equals a power loss of around 40mW which does not sound too bad but wait until you hear the synchronous design power loss. Because with it you simply replace the diode with another MOSFET switch and it is called synchronous because in phase one the upper MOSFET is on while the lower one is off and in phase two that simply flips around synchronously. And since the MOSFET I want to use comes with a resistance of 2.8mΩ the new power loss through it should be around 1% of the diode power loss which is quite a big difference. So all in all a pretty good concept but how can we easily turn on and off the 2 MOSFETs alternating considering that we are dealing with high side switching and the MOSFETs are also never allowed to turn on at the same time. The solution is of course such an IR2184 Half-Bridge MOSFET driver that not only requires one PWM signal in order to turn on/off both MOSFETs alternating with bootstrapping but it also in cooperates a dead time so that both never turn on at the same time. And if you are completely confused right now then I would recommend you to watch my video about MOSFET drivers or basically every video I did so far about switched mode power supplies or voltage converters. But anyway, after I added complementary components around the IC, I turned this schematic into a proper 2 layer PCB design which I ordered at JLCPCB. After one week of waiting I was not only greeted with my beautiful PCBs but also with all the required components which means it was time to solder them all in place. And after 1 hour of SMD and THT soldering, the circuit was almost done. So I connected the PWM signal generated by the Teensy to the PCB and was about to solder in the previously used inductor when I realized that we might have a problem. Before we saw that with this inductor while drawing only a bit of power the current fell to zero during the second switching state which only changed when we drew more current on the output. Those two modes are called discontinuous conduction mode aka DCM and continuous conduction mode aka CCM and they both come with their own unique advantages and disadvantages. Now before in DCM mode the diode prevented current from flowing from the capacitor and through the inductor. But now that we got a MOSFET here this current flow can actually happen which will once again decrease the efficiency. The solution is of course to turn off the lower MOSFET when the current flow reaches zero but that is easier said than being done which is why there are actually lots of synchronous converter ICs out there that do this complicated job for you and all you have to do is simply add a couple of complementary components. For my DIY attempt however I simply ignored this fact because only by using a rather small inductance of for example 29uH and a low current draw this negative current flow was noticeable on the oscilloscope but with the other 100uH inductor this problem was pretty much neglectable. So with the inductor in place, it was time for the first tests and I was super happy to see that nothing blew up and the synchronous converter basically acted like a buck converter. It was also fascinating to discover that while drawing quite a bit of power and thus heating up the 50W load resistor, the MOSFETs pretty much stayed at room temperature. So it was time to measure some input and output power values and calculate a couple of efficiencies which in my case peaked at almost 92% with a 9V output voltage which is of course not the 95% I was hoping for but not bad at all. To achieve that we would have to fine adjust every component and also hit the load current sweet spot which as you can see in this datasheet can be certainly possible. Now of course at this point I could implement a feedback system into my design so that I get a stable output voltage which I might do in the future and meanwhile I encourage you to check out the video description where I linked really good articles about synchronous converters and buck converters in general. With that being said I hope you enjoyed this video and project. If so consider supporting me through Patreon so that I can continue producing more videos like this. As always don't forget to like, share, subscribe and hit the notification bell. Stay creative and I will see you next time.

Channel: GreatScott!

Views: 511,695

Keywords: buck, converter, regulator, switch, switched, mode, power, supply, smps, boost, topology, synchronous, non, diy, do, it, yourself, efficiency, tutorial, guide, make, beginner, beginners, project, gerber, pcb, files, mosfet, diode, schottky, voltage, current, drop, theory, circuit, perfboard, solder, teeny, microcontroller, uC, how, to, inductor, coil, capacitor, load, feedback, conduction, continious, discontinious, advantage, IC, ic, emulation, greatscott, greatscott!, electronics, frequency, best, step, up, down, measure, würth

Id: W4i2FRZ8gXc

Length: 11min 16sec (676 seconds)

Published: Sun Oct 31 2021