Go Down

Topic: Arduino buck converter shield 2.0 (Read 18756 times) previous topic - next topic

antman85

Hi all

I was contacted by a few people who wanted some info on my original buck converter shield. Unfortunately i didn't get the time to finish the project and haven't touched it since. I have decided to update it using better parts and refining the design and i thought I'd post my progress.
Specs are
Aluminium pcb for easy heatsinking
Synchronous converter for over 95% efficiency
Single PWM input with built in dead time compensation
40v maximum input
Input and output voltage dividers for in and out voltage measurement

Here is the layout so far.

antman85

Original buck converter thread
http://forum.arduino.cc/index.php?topic=190642.0

antman85

I have done some more work but still not finished. It is very hard to do a design like this on a 1 layer aluminium board and still optimize the layout. Proper placement of buck converter components can maximize efficiency and reduce unwanted ripple currents.

I'd love to see some examples of surface mounted mosfets on aluminium boards but there just isn't much info out there.

antman85

The board is now finished.
I have added input current measuring and output current measuring. I hope to use it in a small MPPT so this is needed!
Iv'e also added thermal vias so the pcb can be mounted to a heatsink to remove the heat.
I have done away with electrolytic capacitors altogether. I'm hoping some x7r ceramics will do the job by themselves.
Now I just need to order the PCB so I can build and test!

Rusdy

Hi Antman,
Wow, very professional work! I'm honoured to be credited by you for my software work! Anyhoo, you did mention in your previous post (http://forum.arduino.cc/index.php?topic=190642.msg1414301#msg1414301) that your inductor saturates at 11A. I've been noticing loads of people use toroid for their buck converters. However, from my experience, toroids starts saturates on very low current. Is this because your core material is different? Iron powder?

I've been reading lots on the gapped core transformer, due to problem with by boost MPPT. I've concluded I needed gapped core for buck/boost converter as it requires stored energy in the inductor (http://www.ti.com/lit/ml/slup127/slup127.pdf). However, looking at your post history, you seem not having problem with this? Or, I wonder, if you use gapped core, will you be able to increase your current easily?

Rusdy

antman85

Hey Rusdy

Thanks for the comment on my work! Im using a blue/green material inductor. See this this PDF for a great guide on what inductors to use. http://www.micrometals.com/pcparts/PC_L.pdf
I'm no expert to be honest but I have learned alot from this guide. I noticed a huge swing in efficiency if the wrong frequency is used! In another project I have used a custom wound 400 uf 25 amp inductor with a T184-26 core. At around 10 amps I'm above 95% efficiency. The inductor doesn't even rise above 50 degrees at 25 amps load.
Once I build the newest board i will be able to test out the saturation current. I am using a different inductor on this one which I havent used before. It's a PM2120-560K

polishdude20

Hey I recently became very interested in buck converters and am having trouble finding easy to understand resources on how to design them. I don't mean to hijack your post but do you know of any good resources/tutorials? Could you maybe provide a schematic of your design?

antman85

Hey Polishdude

I've been experimenting with buck converters for about the last 6 months. They are very hard to understand at first but with heaps of reading and experimenting i am slowly beginning to understand them. Board layout, capacitor and inductor selection are the most important things. The Mosfet driver I am using in this board is the IR2184.
http://www.irf.com/product-info/datasheets/data/ir2184.pdf
I have not tested this driver yet but ill report on it once tested. I also wouldn't recommend prototyping a buck converter. It is very hard to get a proper layout with proper design tools let alone on veroboard.
I have attached a circuit diagram just in case you want to try it.

antman85

Hey everybody.
Today I received the PCB's and quickly threw one together. Please excuse the lack of alignment of the components as i didn't have much time. The buck converter works quite well except for the lack of input capacitance. I had to add a bulk electrolytic capacitor to stop ringing in the mosfets at high loads.
Test was done with down light globes as a load. At 85 watts load the current was around 11 amps which was the limit of the inductor. The mosfets get to about 50 degrees without heat sinking. I briefly tried clamping the board down to a heat sink with an insulating mat in between and the temperature dropped to a bit above ambient temperature.
I also tried bursts of 330 watts which was about 20 amps on the output. The unit handles it fine but will heat up very quick.

antman85

Update...
I spent some time today playing with the buck converter and using some MPPT code from here
http://epxhilon.blogspot.com.au/2014_06_01_archive.html

It tracks very well but is a bit slow at startup. Once it finds the maximum power point but it does hold it very well.
https://www.youtube.com/watch?v=GhdECGGpoMI

antman85

Update...
I have updated the code to respond faster. The unit tracks much faster and finds the maximum power point in around 10 seconds. I think it needs to slow down a little to keep the duty cycle a bit more stable. See video below.
https://www.youtube.com/watch?v=HuFVVMy6BCY&feature=youtu.be

PedroDaGr8

Why did you use only a bulk capacitor to stop ringing instead of using a snubber? Which is a bulk capacitor in series with a resistor.  Snubbers are great for killing ringing.

I am still learning more about SMPS via sites like DIYSMPS and Coursera. It's an interesting topic but one that is VERY VERY VERY dense and not easily approachable by the novice.

antman85

Hey Pedro

Thanks for the suggestion. Bulk capacitors are needed to cancel out the inductance in the supply leads. I thought ceramics would be enough but it turns out it wasn't.  I actually have places for snubbers on both mosfets but havent used them yet as the circuit is running great already. Once i revise the board again i will tune it with the snubbers and gate resistors.

You are right about SMPS's. It seems when you are a novice (like i was and still partly am) they run on some sort of black magic!

lat50

Hi Antman,

I'm a newbie Arduino tinkerer who is very interested in this project.

Please update me with your current progress.

I would like to know how I could get one or more of these to use in a 12v home solar power system (current systems is 480 watts of 12v panels, with 880 AH battery bankā€¦ to be expanded later).

Thx.

amml2000

Hi Antnam,

I am also watching your project with great interest. Please keep us posted!

Regarding input and output voltage, how hard would it be to adapt your design so it would work to charge a 24v or 48v system? Or to have it handle the 80v from a solar panel such as the sunpower E20/435 SOLAR PANEL?

Furthermore, it might make sense to integrate a boost converter, what you think? Eg so it can still charge the battery on a cloudy day when the volts would be too low to charge the battery otherwise.

Best,
Adrian

Go Up