LM393

[pereczes]

POST REMOVED AT USERS REQUEST

[Rudane]

Could you post a schematic?

[T3sl4co1l]

Circuits that are rated for 5V are to be used at 5V 

[c4757p]

It's not a stupid question (though you really need to give more information next time!). The LM393 indeed should be able to handle up to 36V, and I don't really see what the problem is.

At 12V, the 1k resistors will dissipate about 100mW - a bit high, but shouldn't be enough to pop them. That was my first thought, maybe they're burning open circuit. But no.

Are you sure it's actually an LM393, and they didn't use another similar chip on your board that might not be rated for 36V? The LMV393, for example, is only rated to 5V.

Are you sure you didn't do anything else to mess it up? You'll piss it off if you connect a voltage directly to the output, for instance.

As for the MOSFET - 12V at 1A is hardly anything for a modern FET, just about any TO-220 MOSFET will do. Just make sure the gate threshold is low enough for it to switch on at 5V (keyword: "logic level"), and obviously, check all the ratings in the datasheet.

[David_AVD]

Keep in mind that the output from the LM393 is open collector, not push pull.  This means that it can't "output 12V" as such.  It can only pull it's output pin to ground.

Have you read the data sheet for the LM393?  Lots of example circuits in the National (now TI) one.

[kizzap]

did you draw over the rated output current? might have fried something that way.

[T3sl4co1l]

did you draw over the rated output current? might have fried something that way.

 

(A "nodding" emote would be even better here)

[T3sl4co1l]

Note the short circuit duration is not "unlimited" (another frequent difference between op-amps and comparators).  Actually... what they say is sneaky.  Read the statement, and the note, carefully...

The output saturation voltage curve shows typical current over a range of voltages.

Tim

[David_AVD]

when I increased voltage to 12V the 1ohm in series with the LED on the output managed to produce enough current to burn the output transistor.

Do you really mean 1 Ohm or 1K (1000) Ohms ?

If you meant 1K, I doubt you'd kill the LM393 by having it sink a little over 10mA of LED current.

[David_AVD]

this is the module in question:
http://www.ebay.com/itm/Soil-Moisture-Sensor-Soil-Hygrometer-Detection-Sensor-Module-LM393-/330856046606?pt=LH_DefaultDomain_0&hash=item4d088ea00e

After looking at the circuit in that link, I noticed that the 2nd comparator is left floating.  From the National data sheet:

All pins of any unused comparators should be grounded

[c4757p]

indeed, though not visible under the chip, measured, and they are indeed floating. But ...hm... can that really harm in this case?

It won't hurt it generally*. Some comparators (digital stuff in general) don't like to be left floating or left in a "halfway" state. LM393 doesn't mind, you can even use it as an op amp if you like (it's just hard to tame, because there's no internal compensation capacitor). The TI datasheet notes that ESD protection is "limited".

*It will be more sensitive to static discharge. Positive voltages especially will be "trapped" on the emitter of Q1 or Q4. Negative voltages will hit the current sources, which probably won't be pleased either.

[David_AVD]

In general, floating inputs can lead to oscillation or latchup.  This may result in either incorrect operation or permanent damage.  If the maker says to ground them, you do it.