I have seen a few articles in the forums discussing the circuitry required to protect an Arduino from over-voltage and noise inherent in automotive 12v supply. As a result of what I learned, I came up with the following. I would be interested to get your opinions on how robust this setup is. Have I missed anything- are there situations that will manage to 'fry' my boards?
To explain my thinking:
Fuse to protect high current draw
Diode to protect reverse polarity
zener diode to drain voltages in excess of 9v
voltage regulator to step car voltage (14.4v with engine running) to 9v (not 5v as in pic)
That's a little overkill.... but yeah that would work fine...
Any voltage spikes would be dissipated via the regulator, the caps filter both sides of the wave form, a flyback diode/ scotschy you have is it needed? Unless you expect some inductive voltages?
I've run sensitive electronics on a lot less lol, I once used a zener 12v and it worked flawlessly for many years monitoring the car battery voltage from a bargraph ic.
Many thanks for the response. I know there are many posts on this around and about, but the trouble is that none of them seem definitive. Ive seen some posters saying you need to protect against 100v --> looking at the datasheets for the components I listed I am sure 100v would do some damage to my circuit (if the fuse doesnt go first of course).
Does anyone have any links to a definition of the voltage/current environment in a car? This would seem like a good starting point for designing a circuit to ensure clean supply.
(The reason for the zener in my circuit is to protect against over-voltage to voltage regulator input)
cjdelphi:
That's a little overkill.... but yeah that would work fine.
Umm, no, it won't!
kissenger:
The reason for the Zener in my circuit is to protect against over-voltage to voltage regulator input
Smoke!
Well now, firstly you have the Zener the wrong way round - it will simply directly conduct and blow the fuse.
So, connect it the right way round? No. You specified a 1N4732, which is a 4.7V Zener which will again, conduct heavily and blow the fuse. Were you even to connect it to the output of the 9V regulator (or even a 5V regulator), it would draw a heavy current and shut down the regulator.
Arguably, a 1N4745 specified at 16V would be appropriate. On the other hand, better to make it more like 27V, a 1N4750 which will clamp at a voltage the regulator is stil rated (35V) to handle comfortably (at least as a transient).
cjdelphi:
That's a little overkill.... but yeah that would work fine.
Umm, no, it won't!
kissenger:
The reason for the Zener in my circuit is to protect against over-voltage to voltage regulator input
Smoke!
Well now, firstly you have the Zener the wrong way round - it will simply directly conduct and blow the fuse.
So, connect it the right way round? No. You specified a 1N4732, which is a 4.7V Zener which will again, conduct heavily and blow the fuse. Were you even to connect it to the output of the 9V regulator (or even a 5V regulator), it would draw a heavy current and shut down the regulator.
Arguably, a 1N4745 specified at 16V would be appropriate. On the other hand, better to make it more like 27V, a 1N4750 which will clamp at a voltage the regulator is stil rated (35V) to handle comfortably (at least as a transient).
I did not see a zener voltage so I figured it was a fast switching diode.. my bad
You are totally right, thanks for looking at the detail, really appreciate it. In fact I noticed this, but didn't bother posting an update. What I have now is below:
The voltage regulator is an TS7809CZ - My theory going for 15v zener rather than higher is to avoid overheating the regulator, but i guess it will only be working when there is an overvoltage coming from the car.
I guess what I really want to know is ... are there situations that can happen in a car that this circuit will not protect against?
There are Zener diode surge supressors specifically designed for automotive environments. For example the 1.5KE18 costs less than $1 and can protect against 1500W transients.
I note that there is no impedance within your circuit to limit current absorbed by the zener during over-voltage transients. Any over-voltage will cause current through the isolating diode and the zener, limited only by impedance outside your circuit.
Providing a few ohms in the 14.4V line sufficient to drop say another 0.5 volt under normal use would be adequate to protect the series diode and the zener.
I see what you mean, but having looked at it for a while, I'm struggling to size that resistor. If I take the maximum reverse current that the zener can take it is 27mA. Then if we assume that the car is going to supply peaks up to 100v (not sure if this is real, but several of the threads on this forum suggest you need to protect against this), the maximum voltage across the zener would be 85v. Then, to limit the current to 27mA, the resistor would need to be at least 3148 Ohms?? Seems high, this would surely drop the voltage to the regulator to below its input requirement. What am I doing wrong?
With a surge Zener designed for the purpose (they can pass 100 amps), you don't need a current limiting resistor. Instead, you need a fuse. I've attached a good all-around transient, overvoltage and reverse polarity protection circuit for automotive electronics. The in-line fuse is not shown -- it should preceded the Zener -- and should be chosen considering your ciruit's absolute maximum current draw (as you have done for the circuit you proposed).
kissenger:
The voltage regulator is an TS7809CZ - My theory going for 15v Zener rather than higher is to avoid overheating the regulator, but i guess it will only be working when there is an overvoltage coming from the car.
15V is still wrong as it is too close to the normal running voltage of the car. In no way will the Zener ever limit overheating in the regulator.
Note that the device specified in the post above is rated at 18V.
@Paul__B: Sorry could you explain why it is wrong. I would have thought that the more work the VR has to do to step down the voltage, the less power it has to dissipate, and the less hot it will get. By selecting a zener above the working voltage (14.4v) supplied by the car, but not by much, I thought the zener would be helping to reduce spikes before they got to the VR and thereby minimising the work it has to do. Happy to be wrong, but want to understand why!
And, I'll check the code for the components, but was intending on using a 15v zener.
@jremington: Thanks, I've seen that circuit on other threads. I will consider a surge zener as you suggested, but still want to understand how the current limiting resistor should be sized. If my calculation above is correct, not enough voltage is going to get to my VR!
I would describe myself as a noob with a technical brain - so a little hand-holding should be all I need! Much appreciate all the help given to date.
Been pondering and I don't understand why the alternative circuit uses 18v zener. In that circuit, the zener is the only thing limiting the voltage, so a surge could easily take the outpyt voltage outside the arduino recommended voltage input range.... Or do i miss something?
Do MOV (Metal oxide varistor) get produced for low voltage ? - if the same principal applied to the 120/240ac versions but only in low voltage DC form, that could be used to protect the circuit from voltage spikes.
Sorry for digging up an oldish post, but I am still struggling and would like to put out to the experts for some advice. You will see the circuit I am working on higher up in the thread; I am now looking at the 1.5KE series of TVS diodes to protect against transient overvoltages (thanks to poster for sending me in that direction).
My trouble is that I believe I still need a series resistor with this setup to limit the current going through the TVS when it is conducting. My calculation is thus:
I select a 1.5KE18A with minimum breakdown voltage of 18v, clamping voltage of 25.2v and peak current of 59.5A. So I need to limit current to 59.5A in the case of breakdown, to do this I need a resistor of (125v - 25.2v)/59.5A = 1.6 Ohm; 1.8 Ohm is a standard size so I select that (note 125v is my maximum voltage from various datasheets about loaddump). This small resistor would only drop the voltage to my voltage regulator by 0.1A * 1.8 Ohm = 0.18v (where 0.1A is roughly my maximum normal operating current). This is great and removes the previous problem I had with too great a voltage drop. However, the resistor would need to be capable of dissipating 59.5A * 0.18v ... this is > 10W of power. This seems to me to be a lot of power.
Is my calculation correct? Do I need a power resistor in my circuit?
You do not need a current-limiting resistor if you use the 1.5KE18 Zener surge suppressor diode. The surges against which it is designed to protect are infrequent, short-lived and the diode can safely dissipate over 1000 watts of power during that very short time. The wiring has some small impedance and that will limit the current. But why not add a fuse?
Of course it is possible that there will be larger surges -- like a lightning strike. Then the entire system is toast.
The 1.5KE18A specifies that the 'Maximum Peak Pulse Current, Ipp' is 59.5Amps (it varies slightly depending on which datasheet you look at). The peak transient voltage is known (125v-18v=107v), but I do not know the impedance of the line leading to the TVS. As this will be on a PCB, and the only thing between the TVS and the supply line will be a fuse, I would imagine the impedance to be very low. Hence, if I am to limit the current through the TVS to Ipp, I will need to add a series resistance. Won't I?
You simply can't protect your equipment from all possible catastrophic events, and in any case, how reliable is that 125V figure and how frequently is a peak like that seen?
A 10 amp fast blow fuse between the auto and the equipment would solve your problem much better than a resistor.
Yes, but I am not trying to design a circuit that will protect against all catastrophic events, just this one!
You are totally right about the 125v figure, could be totally made up, I got it from a TVS manufacturer's application note so it could be taken with a pinch of salt. However, that is the figure I'm working with. Google tells me that a very fast acting fuse would be quick enough to catch a load dump transient, but it seems a bit of a blunt instrument.
kissenger:
Anyways, I'm still not sure whether I need that resistor or not!
Are you going to use the purpose-designed 18V protection Zener, or an ordinary one (of the same voltage, of course) - presumably one rated at 5W or greater?
It depends on what sort of surges you anticipate. Short surges will not be much of a problem, they will cause minimal heating in the Zener. Long surges would be more of a problem and suggest the use of a fuse.
If on the other hand, your load current is fairly predictable and you would like to lower the power dissipation of your regulator anyway, then a resistor calculated to drop the input voltage to the regulator to about 8V at the maximum load current, would be a good idea anyway and make the circuit particularly rugged. There is little or no point using a resistor and fuse in series as the resistor will limit the current well below that which would blow the fuse.
