At 00:08 2018-01-27, Brent Brown wrote:
>On 26 Jan 2018 at 15:01, Mario wrote:
>> At 23:10 2017-12-13, Brent Brown wrote:
>> >On 13 Dec 2017 at 15:54, Neil wrote:
>> >
>> >> I want to protect an ADC or op-amp input from over-voltage and I'm
>> >> wondering what common options exist for this? Google is giving me a lot
>> >> of crowbar circuits or just a zener across the input, but the max
>> >> allowed voltage is not too far away from the max measurable
>voltage I need.
>> >> I don't really want to short the input either, but gently limit it.
>> >> Could I use a crowbar-like circuit, but instead of switching on an SCR
>> >> to short the input, I would switch on a MOSFET with a series zener
>> >> instead? This way, the zener only comes into the picture when the
>> >> voltage has been exceeded, and I microcontroller would have indication
>> >> that it's in overvoltage condition.
>> >> Or is there something simpler?
>> >
>> >Attached clamp (not crowbar) circuit has good performance. Reccomended by a
>> >PIClister 10 or more years ago, can't immediately find the reference.
>> >The issue was
>> >Microchip data sheets started specifiying operating specs with reduced
>> >input range
>> >of 0.3V above and below supply rails ~ harder to achieve with diode
>clamps or
>> >zeners.
>> >
>> >The two transistors in the top circuit set up bias points 1 x Vbe
>> >above GND and 1 x
>> >Vbe below supply. The two transistors in the bottom circuit do the
>> >clamping. The
>> >key point is the 1 x Vbe required to turn each clamping transistor on
>> >is balanced out
>> >by the offset bias points, resulting in clamping relatively close to
>> >each supply rail.
>> >Also, the transistor gain makes clamping a little sharper than say a
>> >zener. The
>> >bottom circuit can be repeated multiple times for multiple inputs and
>> >use the same
>> >bias circuit. It looks a little complex, but the performance is much
>> >better than most
>> >simpler circuits. To reduce component count/save space I've sometimes used
>> >BC847BPN dual complimentary NPN/PNP transistor arrays.
>>
>> Hello, I'm probably saying something stupid, but in reality I'm asking:
>>
>> why would this circuit that you shared (and that I am re-attaching) would
>> be preferable to a simple pair of schottky diodes to the supply rails?
>> This only assuming that the input is driven by a low impedence input of
>> course, as the schottky has a significant reverse current flow, but I
>> think it can be assumed because if it was a high impedance source, then
>> there would be little need to protect the ADC input beyond the protection
>> diodes already in the PIC.
>
>Hi Mario,
>
>No, that's a good question. The reason is this circuit keeps the
>voltage within about
>+/-0.1V of the PIC supply rails (depends on Vce sat etc). By
>comparison a pair of
>Schottky diodes can struggle to keep within the +/-0.3V required by
>the Absolute
>Maximum Rating specs for the PIC I/O pins.
>
>There is some discussion about how much current can or can not be safely
>tolerated in the PIC's internal protection diodes, which presumably
>start to conduct
>somewhere around 0.3V or above. The data sheets these days specify an Absolute
>Maximum Rating (e.g. +/-20mA for a PIC16F88x) for the internal diode
>current, but
>NO current is stated for which the device will remain functional and
>completely in
>spec. It is known that bad things can happen when protection currents
>flow in the
>die, latch-up is one possibility (*). It is therefore best practice to
>minimise current in
>the PIC's internal protection diodes, keeping as close to zero current as is
>practicable.
>
>For sure, there are times and places where the simpler/cheaper option
>of Schottky
>diodes might be acceptable. E.g. with a BAT54S (2 x Schottky diodes in series,
>SOT-23) if you can keep the worst case current somewhat below 1mA then Vf will
>be less than 0.3V accross a reasonable temperature range. Or you can
>push further
>towards abs max specs and into undefined area of PIC specs (especially
>if it's only
>during fault conditions), but you'd want to verify empirically that
>PIC operation
>remains acceptable (no latch-ups or other anomalies).
>
>* Section 8.1 of data sheet for PIC16F88x refers to an overload of
>0.6V above or
>below VDD/VSS forward biasing one of the ESD protection diodes on an analog
>input pin, and says "a latch-up may occur".
Thank you Brent, I will save your circuit, it may be useful someday.
>Wikipedia reminds me that a latch-up is "the inadvertent creation of a
>low-impedance path between the power supply rails of a MOSFET circuit,
>triggering
>a parasitic structure which disrupts proper functioning of the part,
>possibly even
>leading to its destruction due to overcurrent. A power cycle is
>required to correct this
>situation."
When I was a kid I discovered this weird phenomenon on my Commodore VIC-20:
if I shorted (..blush.. I feel a monster now) with my screwdriver some chip,
it would quickly start getting very hot until I switched the power off and
then on.
I guess it was latchup. :D
On a side note, my VIC-20 survived and it still works as today. :o)
Cheers,
Mario
>
>Regards, Brent.
>
>--
>http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
>View/change your membership options at
>http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist