Electronics > Projects, Designs, and Technical Stuff

Continuity tester circuit

<< < (4/4)

Zero999:
I don't believe the LM393 contains ESD protection diodes. Bipolar ICs generally don't have ESD protection didoes because they're generally less sensitive to ESD, than MOS devices.

ESD diodes connect the input to both supply rails, so they're reverse biased during normal operation: anode to the -V and cathode to +V. It's easy to test for them using a multemeter's diode function.

Do you have a link to the application node you're talking about? I can't see how ESD protection diodes will cause errors, unless the impedance is high enough for leakage to be an issue.


https://www.electronics-notes.com/articles/constructional_techniques/electrostatic-discharge/esd-circuit-design-guidelines.php

OM222O:
https://www.analog.com/en/analog-dialogue/articles/op-amp-input-overvoltage-protection.html

I got the figure wrong, it was 15mV not 25 and it was at a high temprature, but still really significant! the normal offset of the comparator is 2mV max, so even diodes match that, it's a 100% error added, just by using protection diodes! this is due to the reverse leakage current (which schottky diodes that are usually used, are a lot worse compared to regular silicon diodes) and the input resistor. Now there are things that I can do to minimize this, like use 1K resistors instead of 5K, use lower leakage diodes etc, but I think I need to do some error calculations before finalizing the design. I also need to somehow clamp the diodes to VCC-2V which seems like a really odd requirement, but it's specified in the LM393A datasheet  ??? :-//

I was wondering if I can use regular or even low leakage diodes instead of schottky to reduce the error, but AFAIK the reason they aren't used is because of slower acting and higher voltage drop (0.7 to 1V instead of 0.3).

Do you think bipolar ICs don't need as good of a protection compared to MOS counter parts? they seem to match the ESD requirements without including any diodes which seems odd to me.

Zero999:
You didn't say Schottky diodes, which of course have a huge leakage and contribute to errors, especially if the input impedance is high. The ESD diodes integrated into an IC have a very low leakage current and are seldom an issue. The bias current specification of a MOS input op-amp will account for the ESD protection diodes, so there's no need to consider them separately, unless you've added external ones.

You seem to be confusing the common mode range, with the maximum ratings. The LM393's common mode range only extends to +V-2V, but its inputs can withstand 36V, irrespective of the supply voltage. Refer to page 4 of the data sheet.
http://www.ti.com/lit/ds/symlink/lm2903-n.pdf

Bipolar parts don't need as good ESD protection, compared to MOS devices because they're inherently more resistant to ESD. This is because the gate of a MOSFET is an insulating layer of metal oxide, which will be damaged if the voltage across it exceeds its breakdown voltage. Bipolar devices have a PN diode junction at their inputs which will non-destructively avalanche over, when subject to high voltages. Of course there is a limit: a BJT can be damaged by ESD, but it's much less likely to fail, than a MOSFET when subject to the same abuse.

OM222O:
yes, you are correct! I misread the common mode voltage and the maximum input voltage  :palm:

I'm not sure what kind of diodes they use integrated in the IC, but pretty much everyone uses shcottky for external diodes.
I can see how the PN junction of the BJTs would be more immune to ESD, but that also means if I use normal diodes with about 0.7V forward voltage, they would be completely pointless, since the BJTs base starts to conduct around the same voltage, so it's a lucky choice between the base and the external diode conducting(whichever that has a lower forward voltage). is my understanding correct and I should just ignore using external diodes, and just rely on the input resistor?

Zero999:
The internal ESD protections diodes will be ordinary silicon, not Schottky and the error specifications on the data sheet will obviously take them into account. An external silicon diode will have a larger junction, so will still pass the lion's share of the current, especially if there's another resistor in series with the input.

The BJT input is only one diode. The LM393 has PNP inputs, refer to page 1 of the data sheet, so will conduct if the input voltage is negative, but not positive. Positive voltages which exceed the reverse voltage rating of the BJT's PN junction, (much greater than 36V in this case as it's the maximum rating) will result in avalanche breakdown. If the current/energy is low, then no damage will be done, but a large pulse will damage it. If you're worried about damage, then a diode between each input and the positive supply will provide some extra protection. A low leakage silicon diode will not pass enough current to cause an error, as the input bias current to the LM393 will be much higher than that.
http://www.ti.com/lit/ds/symlink/lm2903-n.pdf

Navigation

[0] Message Index

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod