# LiPo 3S 11.1V Power Regulator

I'm currently working on a small hobby robot, where I want to use a FEZ Panda II along with a RX433 to remote control it. My power source is a Turnigy 3000mAh 3S 20C Lipo which gives me 11.1V to work with, which is too much for the FEZ, but perfect for the motors. I could easily use a resistor to get the right voltage for the FEZ, but I'd like a more generic solution, so I don't have to worry about varying current consumption from different sensors and what not. Using an LM317, I can regulate the voltage to my needs. My question is this:

When I turn on/control the motors, will this affect the input of the power regulator and in turn affect the output, and if it does, how do I stabilize it? I could test it, but I don't want to risk frying my FEZ just yet.

Also, I see people talking about an LDO regulator for the Arduino, is this more appropriate than a simple LM317 regulator?

 - Motor Specs
Voltage             =  12 V
Normal Load Current = 120 mA
Stall Current       = 360 mA

- FEZ Panda II Specs
Voltage             = Through USB port or an external DC 6-9V power supply (connecting both is safe).
Active Current      = 103 mA.
Idle Current        = 65 mA.
Hibernate Current   = 3.75mA.

• The motor specs are interesting, but not that important. We need to know what current and voltage this FEZ thing requires. – Olin Lathrop Jul 5 '11 at 23:18

It's not the motor specs that matter, assuming you're happy with how it runs from your 11V source. The issue is how much current this FEZ thing needs at what voltage.

A linear regulator should work well enough as long as the motor doesn't temporarily make the 11V supply dip so low that the regulator can't maintain the output voltage anymore.

In theory, a regulator puts out a constant voltage as long as the input voltage is above some threshold. If you're trying to get 5V or 3.3V, then 11V is plenty in that regard. However, in reality regulators can only cope with a certain speed of input voltage change. Motors can cause large and sudden spikes, which can get onto the regulated output to some extent even if the regulator input stays above its minimum threshold.

Fortunately the solution to both problems is simple. Put a diode followed by electrolytic storage cap between the 11V supply and the regulator input. That will prevent negative going spikes from making it to the regulator input. The capacitor will hold up the regulator input voltage for the short duration of any spike.

In a generally noisy environment like what you describe, it's a good idea to put some high frequency filtering in front of the regulator. I don't know how much current this FEZ thing draws, but if its just a micrcontroller at let's say 200mA, then a ferrite bead or "chip inductor" followed by 10uF ceramic cap right accross the regulator input and ground terminals will do fine. This will attenuate the high frequencies that the regulator is not so good at dealing with.

Another point is that a linear regulator will be rather inefficient in this application. Even if this FEZ thing wants 5V, that's still 6V it will drop for about 45% efficiency. That by itself may not be a big deal since the power wasted as heat may be small compared to what the motors use. It's probably more a issue of dealing with the heat. Again, that depends on the specs of the FEZ. If it draws 50mA at 5V, then the regulator will only dissipate 300mW. Not a big deal for a TO-220 case in free air. If on the other hand it draws 400mA at 3.3V, then the regulator will dissipate over 3W, which needs to be specifically dealt with.

It might therefore be worth looking into a switcher. At 80% efficiency it would only dissipate 330mW with 400mA at 3.3V out.

• Any recommended ICs for a switch-mode supply? A quick google found me an LM2675, but I'm unsure exactly what I need to look for. – William Mariager Jul 5 '11 at 22:11
• @mindworx: There are many possible ICs and other solutions, but as I already asked, we need to know the current and voltage requirements of the FEZ. – Olin Lathrop Jul 5 '11 at 23:20
• Added the FEZ Panda II specs which were oddly not in the manual I linked, sorry about that. – William Mariager Jul 6 '11 at 0:14
• @mindworx: It's your job to supply the basic specs explicitly, whether they are in the link or not. I'm not likely to follow a link just to answer a question. – Olin Lathrop Jul 6 '11 at 0:22
• @mindworx: In any case, 103mA at let's say 7V would cause a linear regulator to dissipate 420mA. That's doable in free air with the right case. It would also draw the whole 103mA from the 11.1V supply, so would cost you 1.14W. If that's OK, then a linear will be easier. A 7V switcher with 80% efficiency only costs you 900mW in total power. If the extra 1/4 Watt matters, then that's worth looking into. If it doesn't, I'd stay with the simpler linear regulator. – Olin Lathrop Jul 6 '11 at 0:24

It's unlikely to do any harm.

My reasoning: You have a 20C LiPo. That means it is capable of outputting 20x its capacity (3Ah) in amps. So you could draw up to 60A from it with no ill effects. In order to do this, the internal resistance must be low.

I fly my model planes with much smaller Rhino 750mAh 3S batteries. These are capable of giving 15A continuously. Drawing around 6.5A for my motor, I notice a drop of around 0.3V. 350mA is very unlikely to have a significant effect especially from such a big battery.

However, all voltage regulators have something called line regulation and line response characterised in the datasheet. A linear regulator can be modelled as a variable resistor controlled by an op-amp to stabilise the output (ignoring, for now, the short circuit/over temperature protections.) This resistor can't change value instantly, so if you increase the voltage by 10V you will get a little spike on the output, maybe 100mV. The good news: a decrease in voltage will very likely result in a decrease of output voltage. So you're unlikely to fry your device. Line regulation is Vout change over Vin change. Line response is the amount of variation due to transient events.

To ensure maximum stability and minimum spiking, install good electrolytic capacitors on the input and output as well as some small 100n ceramics.