why is this comparator triggering on falling edge?

[hevans1944]

A shorter answer is that the LM358 is the brick shithouse of opamps - slow, stupid, and almost indestructible.

ak

I really do like BSH components, but in today's world of disposable electronics, hardly anyone uses them except by accident.

 

[Audioguru]

The datasheet of the old LM324 and LM358 says, "Either or both inputs can go to +32V without damage (+26V for LM2902), independent of the magnitude of V+".

 

[AnalogKid]

Love that 60 V linear process line.

 

[george2525]

A shorter answer is that the LM358 is the brick shithouse of opamps - slow, stupid, and almost indestructible.

ak

Im not using an LM358! im using an LM324N. I used a tl084 in the diagram because I was too lazy to type in the model and I assumed the reasons would be the same

 

[george2525]

The datasheet of the old LM324 and LM358 says, "Either or both inputs can go to +32V without damage (+26V for LM2902), independent of the magnitude of V+".

thats interesting. Any chance you have a link to the old datasheet??

 

[george2525]

Oh ok I found it here

chrome-extension://oemmndcbldboiebfnladdacbdfmadadm/https://www.onsemi.com/pub/Collateral/LM324-D.PDF

was a simple matter of typing "old datasheet" into google ;-)

 

[AnalogKid]

http://www.analog.com/en/analog-dia...rovide-integrated-overvoltage-protection.html

http://www.analog.com/media/en/training-seminars/tutorials/MT-036.pdf

ak

 

[george2525]

http://www.analog.com/en/analog-dia...rovide-integrated-overvoltage-protection.html

http://www.analog.com/media/en/training-seminars/tutorials/MT-036.pdf

ak

Great links. Cheers

 

[Audioguru]

You are wasting our time by entering the wrong part number and asking why it does something odd.
Don't you read the datasheets for the parts you use? The LM324 is completely different to the TL084 but the case and pin numbers are the same. I think all quad opamps use this case and pin numbers.

I mentioned that the LM358 and LM324 have the same problem and max input voltage rating because their opamps are exactly the same!
The LM358 has only 2 opamps in an 8 pins case and the LM324 has 4 opamps in a 14 pins case.

 

[george2525]

You are wasting our time by entering the wrong part number and asking why it does something odd.
Don't you read the datasheets for the parts you use? The LM324 is completely different to the TL084 but the case and pin numbers are the same. I think all quad opamps use this case and pin numbers.

I mentioned that the LM358 and LM324 have the same problem and max input voltage rating because their opamps are exactly the same!
The LM358 has only 2 opamps in an 8 pins case and the LM324 has 4 opamps in a 14 pins case.

I read the parts of the data sheet that I understand and I ignore the rest. When I have to read 50 data sheets to try and find something that might work, I simply dont have the time to read them all thoroughly.

my LM324 behaved that way so I simulated with a tl084 which did the exact same thing, hence why I thought it may be a standard response for op amps and felt no need to change it.

As soon as I realised it was not a general response I repeatedly stated that I was using an LM324, to be fair

 

[hevans1944]

I read the parts of the data sheet that I understand and I ignore the rest. When I have to read 50 data sheets to try and find something that might work, I simply dont have the time to read them all thoroughly.

That doesn't sound like very good engineering practice. If you don't have time to do it right, where do you find the time to do it over? Shotgun engineering may be okay for amateurs pursuing a hobby interest, but it isn't real engineering. If you just read the parts of the data sheet you understand and ignore the rest, will you ever learn anything?

I apologize if this is just a hobby for you, but some of us here take electronics seriously and try to help others who also take it seriously. Accuracy and details are important insofar as they help those who respond to the original poster. Electronics is an art of the possible. Something that might work goes into the Star Trek file for future reference until someone figures out how to make it work.

 

[george2525]

That doesn't sound like very good engineering practice. If you don't have time to do it right, where do you find the time to do it over? Shotgun engineering may be okay for amateurs pursuing a hobby interest, but it isn't real engineering. If you just read the parts of the data sheet you understand and ignore the rest, will you ever learn anything?

I apologize if this is just a hobby for you, but some of us here take electronics seriously and try to help others who also take it seriously. Accuracy and details are important insofar as they help those who respond to the original poster. Electronics is an art of the possible. Something that might work goes into the Star Trek file for future reference until someone figures out how to make it work.

Well i'm an amateur if thats what you mean. I aim to take it seriously but these things take time. Shotgun engineering has taught me quite a lot actually and you guys on this forum are also teaching me a lot. I used to play music as a hobby and eventually I ended up doing it professionally so I equate learning with how I approached that (putting my fingers anywhere on the guitar until after a few years it started to make sense). 4 years ago I couldnt even do long multiplication or basic fractions.

Anyway, sorry if anyone feels mislead by my schematic. I was just being honest about why I used a random op amp. I use this forum to pick ups tips and learn stuff. Maybe one day I will be able to help someone else. Anyway now you know where im coming from and can choose to help me or not

 

[(*steve*)]

No worries. Learning what everything on a datasheet means is not something that happens overnight.

One of the important things you've learned from this is that when you go outside the specified limits, weird stuff can happen.

This is one of those really interesting ones where they weird stuff is totally not what you would intuitively suspect. In fact it's so weird it's actually comparatively well known.

Other times the performance may be really poor, or device failure is greatly accelerated.

It's well worth challenging yourself to try to understand more and more of your datasheets. Feel free to ask here

 

[george2525]

Feel free to ask here

Ok then I will...

Im looking at 2 datasheets for the LM324 op amp

this one is for the ON semiconductor produced version
link www.onsemi.com/pub/Collateral/LM324-D.PDF

that shows the following:

There is no mention of "maximum input voltage" in this datasheet.

Then this one is from the ST produced model
link www.st.com/resource/en/datasheet/lm324.pdf

That shows this

There is no mention of "maximum common mode voltage" in this datasheet.

so... given that Im expecting all versions of the LM324 to be the same, this seems to be telling me that common mode voltage and input voltage means the same thing?

I was under the impression that common mode for op amps was (V+ + V-)/2

This implies, to me anyway that in the first datasheet one of the inputs could be at 64V if the other was at 0V! but then given that the differential input max is 32V this cant be right, right?

Also as mention by another poster, the ST model datasheet states this:



so is this saying that the inputs can go to +/- 32V no matter what, or only if the positive supply is at 32V? and what does this extra 32V limit signify? does it mean that the op amp may not work BUT at least it wont explode? And again what is meant by "common mode" here. The classic (V+ + V-)/2 or does it realy mean the actual input on each terminal?

its all a bit confusing to me

 

[(*steve*)]

There is no mention of "maximum input voltage" in this datasheet.

See note 5 on the book of page 3.

There is no mention of "maximum common mode voltage" in this datasheet.

See note 1 on page 3.

In this case the device allows the inputs to go pretty much from rail to rail, either separately or in combination. There are some devices in which the differential voltage should not exceed a far smaller figure. Likewise the are some devices where the common mode range is less than the full voltage range. In this case neither is the case. Sure it would be nice if it was mentioned more explicitly, but in a datasheet section dedicated to limitations, the lack of a limitation not bring reported is not so surprising.

It is more interesting what happens when the inputs are between Vcc-1.7V and Vcc.

Reading both datasheets (as far as I have so far) suggests this is an interesting region where the chip is safe, but...

Sometimes it is useful to read datasheets from different manufacturers to get a broader understanding.

so... given that Im expecting all versions of the LM324 to be the same

Well... No. You shouldn't expect that. Different manufacturers may have designed their own version of the chip. They design it to exhibit the same performance within the specified limits. Most of the time this produces cos that are completely compatible, but not always. There are some logic chips where different manufacturers have subtly different implementations. Whilst I can't think of an exact example of the top of my head, I was asked recently that the Schmitt trigger input on a clock pin was not universal across manufacturers.

I was under the impression that common mode for op amps was (V+ + V-)/2

Common mode is where the same signal appears on both inputs. The output should not be affected by this. The common mode range is the voltage range that you can tie both inputs together, and to any voltage in that range and not affect the output (a small differential signal should also appear unchanged on the output). Incidentally, the difference in this case between absolute maximum ratings and maximum ratings are that in the former you're protecting the chip, and in the later you're ensuring performance.

so is this saying that the inputs can go to +/- 32V no matter what, or only if the positive supply is at 32V?

The ST datasheet specifies a max (single ended) supply voltage of 30V. I would interpret this as allowing the inputs to go to Vcc + 2V. It might mean that even with a +3V Vcc you can take the inputs to +32V, but I strongly suspect that's not true. My general feeling is that I should assume the more conservative reading of something if I perceive ambiguity. I might be wrong, but I'll also be safe.

Reading the onsemi datasheet, it says "... but either or both inputs can go to +32V without damage, independent of the magnitude of Vcc".

That makes it clear. 32V above the most negative rail is always OK. This is very useful in case where the signal might be present at the inputs when the device is not power up.

 

[(*steve*)]

Additionally, the TI datasheet gives very clear information on the input voltage range in the absolute maximum Anna electrical characteristics sections.

Not all datasheets are created equal.

Notwithstanding the comment I made previously about sight differences between the implementation between manufacturers, the LM324 is such a jellybean device made by so many manufacturers that you would expect them to be very similar.