How important is the value of the resistor used with a Zener Diode
surely the resistance will affect the voltage read by the pin ?
How important is the value of the resistor used with a Zener Diode
surely the resistance will affect the voltage read by the pin ?
You might want that 1k whether the zener was there or not for current limiting.
That zener is for over voltage protection.
The resistor must be small enough so that the voltage divider consisting of the resistor and the load resistance (digital input resistance in this case) results in enough voltage to cause the Zener to start working. The resistor must also be small enough to permit enough current to flow through the Zener to keep the Zener functioning once it does start working but it also must be large enough to limit the Zener current to a safe value while it is working.
In an application such as yours the latter condition is more critical since the digital input resistance is very large. You would need to know the power (or current) rating of the Zener diode in order to determine how small the resistor could be.
surely the resistance will affect the voltage read by the pin ?
No it won't. Normally the current is so low that the voltage drop across the resistor is negligible and the voltage read by the pin will be essentially the same as the input voltage. If the input voltage were to 'spike' then the Zener would start operating and the voltage read by the pin would be 4.7 volts.
Don
You might want that 1k whether the zener was there or not for current limiting.
No, this is a growing rubbish trend I have noticed.
You do not need or indeed want series resistors on digital inputs.
The argument about limiting the current is wrong with the high impedance of the input is just does not make a jot of difference to the current. What it does do however, is reduce the noise immunity and so make the input more susceptible to picking up interference.
What it does do however, is reduce the noise immunity and so make the input more susceptible to picking up interference.
That's not all. The resistor creates Johnson Nyquist noise of its own. (A useful phenomenon when making a white noise generator)
Grumpy mike
so i do not need a resistor at all?
Did you have a section about zener diodes on your web site?
You'd want a resistor before the zener but in parallel with the input,
like said before picked to not allow the zener to burn out
and I think the inputs have a relatively lower impedance as far as high impedance inputs go, but nothing to worry about unless your reading a high impedance voltage divider or something
so i do not need a resistor at all?
If you have a zener and you want to protect against a high voltage then you need a resistor to stop the zener burning out when that voltage is applied.
The value should be as small as possible to prevent the problems I outlined earlier.
So find out what current your zener can take, what the maximum voltage you want to protect the input from and arrange the resistor to allow this current (or maybe 80% of the maximum current) to flow when the maximum over voltage is experienced.
I talk about this here:-
http://www.thebox.myzen.co.uk/Tutorial/Protection.html
You'd want a resistor before the zener but in parallel with the input,
Not sure I understand that, it sounds wrong to me. A resistor in parallel with the zener is not going to divert much current. This is because the zener is a non linear device as so will take much more of the current as the voltage rises. The resistor is of course a linear device and will obey ohms law in passing current with a rising voltage.
Before like in series with the zener, parallel the input
basically what you said just slightly more crude english
Thanks mike
its a maximum voltage of 5v the board can only read 2.5
Regards
You'd want a resistor before the zener but in parallel with the input,
Before like in series with the zener, parallel the input
Both of these are wrong. If you are going to use a resistor then it should be exactly where it is shown in the diagram.
Don
Johnson and Nyquist were NEVER contemporary... and the resistor is to protect the Zener from "Short" noise pulses. The size of the resistor is usually empirically determined and depends on the dissipation of the Zener and the energy in the noise pulse 1K is Much too large (unless the driver supplies the voltage as a part of an interface which is poor engineering at best...). Before I would make a snap judgement like the ones I have seen so far I personally would want to know a great deal more about the application. 1K would be ok for an input voltage much higher than 5V but wholly inappropriate if it was just a noise clipper where 33 - 100 ohms would be more appropriate. Zener diodes are frequently used for noise suppression because they are Fast when they conduct and in many cases to be preferred to a Fast diode (yes they are Very different) and sometimes as interface devices, for example connecting RS232 or 422 signals to a processor or in the case of RS485 used as noise clippers. Two Completely different uses/applications. FYI the OP that compared Johnson and nyquist is mis-informed Nyquist noise is artifacts left behind in A/D sampling and is generally a product or mix of the sampling speed and the sampled signal. The Nyquist limit or the minimum period one can sample a signal and get back reliable/clean results. The period is usually taken at twice the period of the signal (minimum time). Johnson or Shot noise is related to loose electrons (Very simplified here) moving through a material that is passing a current ac or dc knocking other electrons loose and thus creating random noise. Said noise in a "Good" part is typically > 60 db down from 0 DB power levels or a factor of a Million to one. This is what I was taught it has always been good at least so far. IMO
Doc
Excellent tretises on zener OV protection and noise suppression, Mike & Doc. Thanks.
But, not to go too far off topic, isn't the symbol in the OP's schematic that of a 'Fast Recovery' Diode, not a Zener??
FYI the OP that compared Johnson and nyquist is mis-informed
You better tell Wikipedia
To be fair he didn't compare them at all.
isn't the symbol in the OP's schematic that of a 'Fast Recovery' Diode, not a Zener
Well it is what I have been using as a zener symbol for 40 years.
Do a google image search for ** zener symbol ** plenty of people agree with me.
Not to me.. I would draw and interpret it as a zener, a schottky is 'curly' on the ends of the cathode bar and an abrupt one has a slightly different symbol but I think it is more the country you come from. I've seen many different symbols for parts over the years English and German being the two I remember most readily as compared to ones drawn in the US. Schematics drawn in GB still look as funny to me as ours no doubt do to them. The other reason why I wouldn't consider the symbol as 'other' is the application. It makes a limited amount of sense to interface a higher voltage in that manner. Chancy because of possible and IMO likely grounding issues but imminently usable. IMO
Doc
Grumpy_Mike:
You might want that 1k whether the zener was there or not for current limiting.
No, this is a growing rubbish trend I have noticed.
You do not need or indeed want series resistors on digital inputs.
The argument about limiting the current is wrong with the high impedance of the input is just does not make a jot of difference to the current. What it does do however, is reduce the noise immunity and so make the input more susceptible to picking up interference.
You agree you'd want that resistor if that was the base of a transistor, right? It says "sinking NPN interface", and I didn't notice it said "digital input".
(Though maybe lower value)
RE Nyquist I read the article and I am CORRECTED And I I sincerely regret the error. My knowledge of Nyquist is pretty much limited to the sampling theorem and I haven't used that in years. There is a better noise generator as I remember it uses a zener junction in avalanche mode and works well but statistically the noise is weighted a little as it is sorta rolled off due to the structure of the zener junction. Makes an excellent audio and low RF test generator. For true noise generation there are specially doped pn junctions, they are good to 4 - 6 ghz. They replace the old gas tubes created just for noise. Someone in the 30's or early 40's discovered that gas voltage regulators (big neon bulbs, essentially) like an 0B3 made a lot of noise and a small company investigated various gas mixes and the noise generator was a science. At least until solid state devices really came into their own. IMO
and I would like to apologize again for jumping the gun and not verifying my fat mouth before I shot it off.
Doc
wow, it always amazes me to hears stories like that, all the old school outdated stuff I would never have heard about unless people on this forum reminisced about it
Grumpy_Mike:
You do not need or indeed want series resistors on digital inputs.
The argument about limiting the current is wrong with the high impedance of the input is just does not make a jot of difference to the current. What it does do however, is reduce the noise immunity and so make the input more susceptible to picking up interference.
I just want to be sure I'm understanding this. So I found an article on sinking and sourcing sensors. With the pin in INPUT_PULLUP mode, the internal pullup is enabled, it will read HIGH, unless the NPN is on, in which case it's shorted to ground. In this case, I don't grok what the zener / resistor would be doing here at all.
(BTW, did the PinMode function change? IIRC, it used to be that mode INPUT enabled the pullup.)
I grok that waiting is necessary here as the fullness of understanding will come. There are 2 main reasons for a zener diode on a digital input pin one is a clipper (there are 2 diodes in the proc from any input pin one from Vcc to the input cathode to Vcc and one from input to ground cathode to input. The purpose of the diodes is to clip inputs to Vcc+.6V and to ground +.6 Vdc) this is done to keep other devices internal to the processor from being driven to voltages that could cause the processor to fail or operate improperly. If high speed spikes are coupled in and not restricted to those limits they can again interfere... thus a zener from input to ground in this case the series resistor value will be small 10 - 100 ohms (Typ) this is a clipper. If the port is used to monitor a higher voltage (interface with a voltage > 5V the resistor will be in the range of 1k to 10K (Typ) and the resistor/zener is a 'level' shifter as the resistor limits current to the zener from the higher voltage input (typ 10 - 15V) source to the few ma necessary to properly bias the zener and thus limit the voltage presented to the input so as to not drive the protection diodes into conduction. This is a typical poor mans RS232 - interface W/O a level shifter IC (MAX232). This interface is workable but provides no isolation for ground and thus is prone to grounding or noise issues. As i mentioned before if any one is familiar with the picaxe... it uses that scheme to directly drive the chip from a 'raw' RS232 serial port (non ttl) W/O an interface ic.
The scheme will work fine for local connections but is IMO poor engineering or engineering on "the Cheep", chip?, cheap?
Doc