Designing a voltage/power regulator

Projects General Guidance
1

I want to design a voltage/power regulator for electronic cigarettes. As you probably know, e cig is just composed of a heating (resistive) coil that is powered by a lithium battery (though usually not lithium ion)

When the battery is fully charged, it's around 4.2V. As it discharges, voltage at the battery terminals drop, sometimes to as low as 3.5V. Without any power regulation, this causes the power to be reduced at the heating element (which is a fixed resistance). Power = V^2/R.

I want to build a device that will output a fixed voltage to the heating element, say 4V or 4.5V, or 3.8V. Actually voltage value doesn't matter as long as it's fixed and consistent, since the resistance of the coil can come in different resistance to give a desired power level.

------------------The above design will give an output (at coil) at a certain fixed voltage, regardless of battery charge--------------

optional: The user should be able to adjust the voltage or power (within a range).

Here's the kicker: The power to the coil must withstand as high as 75W.

Can a prototype be built with Arduino? What else do I need?

Thanks

2

anyone comments or suggestions to get me started please?

3

You want to build a buck/boost DC-DC converter controlled by a microcontroller and appropriately beefy components.
That's what you should be researching.

You are aware that devices like this, for E-cigs, are readily available now, right? All the cool (?) kids have em now. My friend has one rated for 140W w/2 18650 batteries, and it's not even that big. LCD displays showing current operating parameters are par for the course, and many of them support nickel coils that change resistance as they heat up, so they're able to hold a specific coil temperature, rather than a specific wattage or voltage.

4

Thanks.

I'm aware that chips that does this have existed for a long time. I'm a vaper myself. I'm doing this for hobbies sake. I just need some basic questions answered that'll point me in the right direction.

  1. I understand buck converter can only go from a high voltage to a lower voltage, while a boost convert go from low to higher. Which one do you recommend using? Also, I've heard that buck converter is “easier" from a design perspective.

  2. If the output power is 75W or even 100+W, that's a pretty large current. That current does NOT go through the microcontroller (Arduino in our case) right?

  3. How exactly does the microcontroller control the voltage to the load? Is it via pulse width modulation?

  4. I'm aware of features like temp control and nickel wires. I'm not shooting for that right now. I just want to design a simple fixed wattage/voltage device that can handle large power. How hard is it to design this? Assuming I do my hw and work hard, is 3 month enough to get a rough prototype design done? Just wanna know what I'm in for here.

Thanks

5

You can learn more on this buck-boost converter, so that the input voltage can be higher or lower than the output voltage.

http://www.diyourself.ru/technology/arduino-based-switching-voltage-regulators-do-it.html

That current does NOT go through the microcontroller (Arduino in your case), it will go through the MOSFET.

paulwece:
I just want to design a simple fixed wattage/voltage device that can handle large power. How hard is it to design this? Assuming I do my hw and work hard, is 3 month enough to get a rough prototype design done?

You need alot of MOSFET too, to test out your design. Any mistake in design will burn the MOSFET.

6

Thanks, I will study those links you provided. I also bought a textbook on power electronics.

Now, are their off the shelf buck or boost converters I can buy that will meet my power requirements or do I have to design it from scratch myself? The answer to this question will affect how long I can get something working. It'll be nice if their are off the shelf stuff I can buy and just connect it with the Arduino that'll complete most of the hardware setup and then all I have to do is program the Arduino, which shouldn't take too long...

7

Here are pics of some boost converters I found:

It's pretty big, which is understandable given the high current involved.

Yet the chip that's used in the e cig market, the DNA200, which can handle 200W, is very tiny in comparison:

1280×853 147 KB

How is that possible to make a chip so compact yet capable of 200W? I thought you need beefy components?

Please advise. I'm very curious about this.
thanks

8

Well, the first photo shows something that's made for general purposes. They don't know what you're going to power with it, so the output terminals are large and the components are relatively oversized. The second photo shows something built to a totally different design aesthetic.

Look how tiny and complex the traces are. Look how many chips there are. That thing certainly has more computing power than the Apollo spacecraft and it only does one thing: heat up a coil. If you wanted to use that circuit for something else, it would perform very badly at that other task.

9

MorganS:
Well, the first photo shows something that's made for general purposes. They don't know what you're going to power with it, so the output terminals are large and the components are relatively oversized. The second photo shows something built to a totally different design aesthetic.

Look how tiny and complex the traces are. Look how many chips there are. That thing certainly has more computing power than the Apollo spacecraft and it only does one thing: heat up a coil. If you wanted to use that circuit for something else, it would perform very badly at that other task.

I get your point, but I always thought you need pretty beefy components for that kind of power (100W-200W).

You even see a heat sink on the first one, which makes sense because even a slight inefficiency causes a lot of heat.

On the DNA chip, there is nothing that I can see that can withstand 15A. Or maybe you just don't need that beefy of components like the first one?

10

Well MOSFET technology has advanced a long way in the last 10 years. That big circuit is probably using components designed 20 or 30 years ago. Most of the visible components in the photo could be older than that.

The newer MOSFETs have lower RDSon, which means less power wasted in the MOSFET and more delivered to the load. The technology for driving the MOSFETs efficiently has also improved.

But like I said before, optimizing for one task gives a very small circuit.

11

Hi,

http://www.evolvapor.com/datasheet/dna200.pdf

This is the data sheet for the DNA200, it gives the output load and power characteristics.

Note that it uses temperature feedback via a piece of resistance wire and the load resistance being 0.20 Ohms

Tom... :slight_smile:

12

thanks for the help so far everyone.

I'm going to study a bit more about boost converters and also that datasheet TomGeorge provided, so I can come up with more intelligent questions later.

This is something I want to be able to design and understand, by the end of the year hopefully.

thanks

13

Okay, I'm back. I've been studying boost/buck converters after buying a book on power electronics. Now I know the basic concept behind them. Essentially it uses switches to control the energy contained in the inductor, and the switching mechanism is key to making it work. The switching is done through a MOSFET usually, and the PWM duty cycle and frequency are important characteristics.

I guess the microcontroller (such as Arduino) is used to output the PWM signal to the MOSFET. That's essentially the basics right?

A few questions:

  1. The circuit I've seen in the textbook and WIkipedia are very simple, just a few switches, inductors, capcitors, etc... Will a real one be this simple too?

  2. A friend recommended this regulator. I think it comes close to my need:
    http://www.mouser.com/new/monolithicpowersystems/mps-mp3427-mp3428/
    Does this make my project a lot simpler? What else do I need?

Thanks

14

The people who design and build regulators like the DNA200 usually have an EE degree and several years of practical experience.

15

paulwece:
2. A friend recommended this regulator. I think it comes close to my need:
Monolithic Power Systems (MPS) Distributor | Mouser Europe
Does this make my project a lot simpler? What else do I need?

Yes, buying an IC will make your project a lot simpler. The more you pay, the simpler it gets. For example this even has the FETs and inductor inside the package, and two of them could provide 2.5V @ 30A to get the 75W you need -> http://cds.linear.com/docs/en/datasheet/4611fb.pdf

This sort of goes against the DIY you desire though, doesn't it? I'm not saying it's bad idea to buy something (these ICs exist because designing these circuits is hard), but you seemed to want this as a DIY project as opposed to solving a problem the simplest way.

Also, does your coil really not change resistance as it heats up? You say constant power/voltage in this thread which is only true if the coil does not change resistance with temperature. I find that hard to believe, although I'm not familiar with coils designed for vaporizors.

16

BigBobby:
Yes, buying an IC will make your project a lot simpler. The more you pay, the simpler it gets. For example this even has the FETs and inductor inside the package, and two of them could provide 2.5V @ 30A to get the 75W you need -> http://cds.linear.com/docs/en/datasheet/4611fb.pdf

This sort of goes against the DIY you desire though, doesn't it? I'm not saying it's bad idea to buy something (these ICs exist because designing these circuits is hard), but you seemed to want this as a DIY project as opposed to solving a problem the simplest way.

Also, does your coil really not change resistance as it heats up? You say constant power/voltage in this thread which is only true with the coil does not change resistance with temperature. I find that hard to believe, although I'm not familiar with coils designed for vaporizors.

Thanks. The way I see it, why reinvent the wheel? Even with these ICs out there, there are plenty of work for me I'm sure. For example, I still need to drive it with a microcontroller don't I? I also want the user to be able to change settings, and incorporate a LCD screen, etc... so there are plenty of "DIY" stuff left for me. I prefer to work with microcontrollers and come up with unique features from the software side. I don't want to deal with the hardware side, my field of specialization is not power electronics. I'm a computer engineer (recent grad).

I feel that using a off the shelf IC that was designed by experts would be "safer," because I would personally be using this that I designed. Last thing one want is have the battery explode in one's face.

But I'm happy there are ICs out there, now my job is to understand all the specs so I can incorporate other elements.

The coil typically used for e cigs is called Kanthal (iron-chromium-aluminium). I'm not sure if the resistance changes with temperature, probably not much. But assuming it does, how hard would it be to incorporate an ohmmeter function to compensate?

Thanks for your help. I appreciate it. I'm more interested in the chip that my friend told me about. It seems capable of even higher power. I'm going to study it and try to understand the specs and how it works myself in the coming days. But if you can jump start me and just tell me what is included in that package and what else I need to provide that'll be great (like inductors, FET, etc...)

17

paulwece:
The coil typically used for e cigs is called Kanthal (iron-chromium-aluminium). I'm not sure if the resistance changes with temperature, probably not much. But assuming it does, how hard would it be to incorporate an ohmmeter function to compensate?

I checked out Kanthal's documentation here.

On page 6 they have a chart of Resistivity vs Temperature. It seems depending upon which type of wire you get, you can have either have a little or a lot of change.

To make a constant power supply would actually be pretty difficult. If you go to google you'll find a million circuits for constant voltage or constant current. You won't find many for constant power.

18

Bigbuddy, can you please take a look at that regulator I talked about?

http://www.mouser.com/new/monolithicpowersystems/mps-mp3427-mp3428/

I want to order some and start playing with it right away. I hope you can tell me what other components I need. Then I will slowly digest the data sheet. But if you can give me a general overview of the regulator with your better knowledge on electronics that'll give me a head start. Like if it's capable of higher than 75W, what kind of FET I should get, and possibly on how to use Arduino with it to experiment. Anything I should be careful with.

Any advice would be highly appreciated and would get me started. I will learn the details slowly on my own, just like last time.

Thanks

PS: A newbie like me shouldn't be too ambitious at this point. Constant voltage would be good enough. But I need to be able to increase or decrease the voltage. For example, I can't just have the output always be 5V. But I guess this is something I can do through the Arduino with the PWM.

19

Yes, some wires change resistance with temperature, but most don't. The special ones that change resistance with temperature are used with "temperature control" chips that monitors and limit the temperature by measuring the resistance. So I guess the next step would be to figure out how to measure the resistance of the coil with an ohmmeter function. Because if we can monitor the resistance at all times, then it's not hard to make the power constant by varying the voltage accordingly.

But for now, I just want to make the voltage constant (albeit adjustable). If you get a chance, I respectfully ask that you skim over that regulator link I provided, and see if you can give me some pointers. I just want to be set in the right direction, thanks.

20

That IC is not suitable for someone with limited electronics experience.

Unless you have considerable experience designing PCBs, and soldering surface mount components you will not get it to work.

The layout of the PCB for a switching regulator is critical. You need to consider the loops that current flows through at different stages of the switching cycle and optimise them.
This is mentioned on page 16 of the datasheet.

Also the IC is only 3mm x 4mm in size, with 22 pins. Could you satisfactorily solder that in place?