Let's Make Robots!

doubts regarding circuit protection fuses

So, overall I know what a circuit protection fuse is and what it's purpose. I also know that are several types of fuses, in many shapes and forms, one-time-only and resettable (self, or non-self), slow and fast acting and some of rules of thumb for when using one or the other. However, my personal experience is only limited to often trigger the circuit breakers of my mains power supply (by either drawing to much current or shorting stuff [as a kid *cough* long time ago]) and blowing a few glass fuses.

From my research I read that slow blow fuses are indicated to protect things like motors or batteries/power supplies? And well, fast blowing fuses whenever any? spike in current could lead to catastrophic consequences? I recently equiped my multimeter with a fast blowing fuse, but what I read is that catastrophes can happen "when making reading on ovens something something main power" but, anyway I'm not planning to mess around with main (AC) current as of the moment.

What concerns me mostly, as of the moment, is fuse protection when using LiPos. So far I've only used LiPo batteries with "certified safe" circuits, that is not designed or built by myself, so besides the normal precautions handling LiPos plugging such battery was never a concern to me.

I've also read that the fuse value to be used should be the expected/measure current usage of the whole circuit plus a margin but many questions still linger in my mind for instance:

- how flexible should that margin be? what main variables to consider?

- should I use just one fuse directly after the battery? should I use several fuses trying to protect the most valueable/costy parts of my system? one fuse per "sub-circuitry"? that is I have vreg rated for 1A to feed the microcontroller and radio module, should I have a 1A slow blow fuse before it?

- for instance, I've built a couple of motor drivers/h-bridges and use diodes for protection, that way the motors should be protected, no? the only foreseable problems would be... diodes current rating being overcome or stray conductive material causing a short?

Also, I have no idea how often would a LiPo "overfeed a power spike without being asked for it".

I'm probably being overthinking it and being overparanoid and furthermore expressing myself badly because I also feel that the information I've gathered on fuses it's a jumble on my mind.

So I also issue a sort of challenge, if someone out there savvy enough to write up a compreensive tip/walkthrough regarding fuses and typical application within robotics, I at least, would appreciate it very much.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

I don't care about blowing up some electronic components.  We have Ebay and China to solve that :p I'm only worried about my lipo blowing up because of a short circuit. So I put an in line fuse in the connection of the lipo and the robot. I use a 10A fuse in my Rover 5 and a 5A fuse for my little hexapod.

Something like this: http://www.ebay.co.uk/itm/In-Line-standard-waterproof-Blade-Fuse-Holder-fuses-10a-10amp-kit-car-boat-bike-/300640464789?pt=UK_In_Car_Technology&hash=item45ff916f95

My problem is not toasting one component or the other, it's toasting the whole board or an big portion of it rendering it useless and thus nullifying hours and hours of my work :(

If the motor controller boards are of a good design I can’t see why you should get a lot of current via the control lines. Nor that the LiPo should pump lots of current into your Arduino clone. But then again I’m just a happy amateur.
As you probably know, it’s bad practice to run noisy motors from the same power source as the microcontroller so you would probably be better off with a separate power source for the Arduino.

There you go "if ... are of a good design". Since I've designed them myself (although based on a bunch of schematics) I just can't be too sure of what to expect :)

I'm current specific case I'm trying to specifically avoid a second power source because there are weight constrains I must consider. 

You might get some good tips from EEVblog on Multimeter Input Protection
http://www.youtube.com/watch?v=zUhnGp5vh60&feature=g-user-u

but then again you always have the Murphy’s law;
The rule of Protection:
If you install a 50 cent fuse to protect a 100$ component, the 100$ component will blow to protect the 50 cent fuse.

Truer words never spoken.

oops

A fuse or breaker interrupts the circuit if too much current is drawn. When applying a fuse to the main power input, fuse the "hot side", as close as possible to your voltage source (e.g., your battery). This way if power at any point in your circuit gets shorted to a ground, your circuit is still protected.

As you mention, choose the value of the fuse based on the expected current draw for whatever you are protecting. Normally, you fuse the power source, so this is the total maximum expected draw of your circuit.  

If you have a branch of your circuit that could fail and draw too much current, you could fuse that separately. However, this is unusual. You could fuse your motors separately, as discussed below.

Whatever you are fusing, the wires and other components in that part of the circuit need to be rated for more than the fused value. In other words, if your component is rated for 500 mA, you need to fuse it at less than 500 mA. Otherwise, the component is as likely to fail as the fuse.

Regarding slow blow or fast blow, it depends on the component. Electronics like ICs can be pretty sensitive. They could fail even if there is a very short increase in current, like a sudden spike.

Motors, solenoids, and other heavy duty items can take more abuse. Plus the inductive nature of these devices means they may introduce their own noise or even Electro Motive Feedback (EMF) that could pop a fast blow fuse. That is where a slow blow fuse is helpful. You could provide separate fuses for your motor lines, and use slow blow fuses on that. 

The diodes on the motors and motor drivers are to eliminate the back EMF when the current through the motor (or any coil) changes suddenly. This happens when you start, stop or change the direction of the current through the coil. So it is nothing to do with fusing, but instead protects your circuit from unexpected reverse voltages across your motors.

While you are on the subject of circuit protection, you can also consider a large power diode on the power input of your main supply. Since a diode only allows current to pass in one direction, this protects against accidently reversing the power leads on your circuit, which as been the death of many a circuit. Keep in mind that you will get a voltage drop (about 0.7V) across your power diode, so pick your battery or other power source with that in mind.

 

The only (current) doubt that persists somehow relates to:

"Whatever you are fusing, the wires and other components in that part of the circuit need to be rated for more than the fused value. In other words, if your component is rated for 500 mA, you need to fuse it at less than 500 mA. Otherwise, the component is as likely to fail as the fuse."

Bellow I present a sort of mock circuit diagram of how I plan to connect together 3 modules (3 boards)

So, a common GND is a given and a must but will particular routings of either V+ and/or GND will help protect a given module in the case of another one failing?

For instance, let's assume module C goes crazy and starts drawing current like mad, from what I'm told the UBEC is supposed to fry as a safeguard but given that V+ to the UBEC comes from a fork at the battery would the rest of the system be compromised? Even if module C was the faulty one? (I think I going to direct connect the V+ from the battery there, so no UBEC).

Or, in either way be it B or C failing since everything is connect through a common GND would fry anyway starting with the module A where the vreg there is only rated for 1A? More so, in normal conditions... both modules B and C are expected to draw at least 1.5A which is superior to module A vreg rating... have I incurred in a noobish design flaw?

Maybe this will help.

Think about how current flows through the parts of this circuit. All the current flows into and out of the battery Vt, eventually.

The voltage through R2 is found by following the path from Vt+ to Vt-. It will have to go through R1 and R2. So the current is given by Ohm's Law:

I2 = Vt / (R1 + R2)

Similarly, the current through R3 also has to flow through R1. Out comes Ohm's Law again:

I3 = Vt / (R1 + R2)

Now the current through R1, since there are no other loads or paths for the electricity to flow, is the sum of the two currents I2 and I3:

I1 = I2 + I3

That is the total current being drawn from the battery. If you wanted to fuse the whole circuit, you'd base the fuse size in I1, and place the fuse between Vt+ and R1.

If you wanted to fuse just R2 (which could represent the load of a motor), you would base the fuse size on I2 and place it inline with R2. Note that if you wouldn't want this fuse placed between R1 and R2 inline with the top line of the circuit, because then it would see the full current I1 of the circuit.

Also note that the point directly below R2 is GND. However, only I2 is flowing though this part of ground. As soon as that line connects to the bottom line of the circuit, you add in I3 and it is carrying the whole current load.

Same goes for R3. You could drop a fuse inline with R3, and it would only see current I3.

How do we apply this to your circuit?

Each of your boards and the UBEC needs a connection to V+ and GND. Let's say you have one big GND that runs serially from the battery, to your 'duino, to the H-Bridge at B, to the Driver at C, to the UBEC. That GND connection, including all wires, circuit traces, etc, needs to support the full current you expect to run. If board C loses its mind and starts to pull heavy current, it will flow through the shared GND path of all the boards.

On the other hand, you could provide a separate GND wire to each board, and it will only see the current through that board. Just make sure they all tie back to battery GND. If board C goes nuts and starts pulling lots of current, it will only affect that board. Of course if it is a bad short, it will affect the whole battery.

That's why you want at least one big fuse for the whole battery. As long as you have separate V+ and GND lines for each board, the other boards won't be zapped if one board starts to draw crazy amounts of current, because that current is not flowing through the GND of all the boards.