8051F020 series, 5V tolerant input schematic?

[Joerg]

So you have a "dirty" 5V supply?

No, rock-stable. But clamping against it will lead to up to 7V spike
amplitude if it's a huge transient. Those small clamper arrays can
easily reach 2V Vf or more for a few usec.

 

[John K]

No, rock-stable. But clamping against it will lead to up to 7V spike
amplitude if it's a huge transient. Those small clamper arrays can
easily reach 2V Vf or more for a few usec.

If you are dealing with 5V logic signals, isn't the threshold VCC/2, or
around 2.5V?

How about clamping with a 3.9V zener to ground with a small series
resistor to limit the transient current? As long as the desired signal
gets above the threshold, do you care if it goes all the way to 5V?

For a BZX84C3V9, Fig. 2 on page 3 shows it doesn't start to conduct until
you get above 3V, and 50mA doesn't even reach 5V:

http://www.pridmore.com/datasheets/BZX84CA07.pdf

Don't know what all the tolerances are, but so far nothing else seems to
have any chance of working in your application.

JK

 

[Joerg]

_Where_ are these "huge transients" coming from?

The usual, through outside cabling, induced by heavy machinery being
turned on et cetera. Unit is plastic (has to be), a decent ground
structure is difficult to impossible.

 

[Joerg]

I still haven't seen described how a "rock solid" 5V supply produces
such gross transients.

It doesn't. It's the signal lines that do.

[...]

 

[Joerg]

On Sat, 29 Jun 2013 08:00:01 -0700, Joerg <[email protected]>
Oops, sorry, your are right. At a 5V-tolerant pin it would work.
Problem is that we only have 5V in the area where the protectors
are.
If you only have 5V around driving a 5V-tolerant pin... what are
you fretting over?

The concern are surge events that make these lines carry nasty
transients, which in turn can lock up or damage the uC.
So you have a "dirty" 5V supply?

No, rock-stable. But clamping against it will lead to up to 7V spike
amplitude if it's a huge transient. Those small clamper arrays can
easily reach 2V Vf or more for a few usec.
If you are dealing with 5V logic signals, isn't the threshold VCC/2, or
around 2.5V?

How about clamping with a 3.9V zener to ground with a small series
resistor to limit the transient current? As long as the desired signal
gets above the threshold, do you care if it goes all the way to 5V?

For a BZX84C3V9, Fig. 2 on page 3 shows it doesn't start to conduct until
you get above 3V, and 50mA doesn't even reach 5V:

http://www.pridmore.com/datasheets/BZX84CA07.pdf

Don't know what all the tolerances are, but so far nothing else seems to
have any chance of working in your application.

JK
I still haven't seen described how a "rock solid" 5V supply produces
such gross transients.

It doesn't. It's the signal lines that do.

[...]

So "induced" transients? How much actual energy?

Hard to say how much arrives, all I know from reports is that it has
frozen the uC at times. So we had to do something about it. We should be
good with those clamps. I was just wondering how far things can go
above abs max and for how long. Of course we won't be sending in 7V for
hours. When I redesign stuff I like to make it not just good enoug to
pass a test, but ideally almost bullet-proof.

 

[Joerg]

Jim Thompson wrote:


On Sat, 29 Jun 2013 08:00:01 -0700, Joerg <[email protected]>
Oops, sorry, your are right. At a 5V-tolerant pin it would work.
Problem is that we only have 5V in the area where the protectors
are.
If you only have 5V around driving a 5V-tolerant pin... what are
you fretting over?

The concern are surge events that make these lines carry nasty
transients, which in turn can lock up or damage the uC.
So you have a "dirty" 5V supply?

No, rock-stable. But clamping against it will lead to up to 7V spike
amplitude if it's a huge transient. Those small clamper arrays can
easily reach 2V Vf or more for a few usec.
If you are dealing with 5V logic signals, isn't the threshold VCC/2, or
around 2.5V?

How about clamping with a 3.9V zener to ground with a small series
resistor to limit the transient current? As long as the desired signal
gets above the threshold, do you care if it goes all the way to 5V?

For a BZX84C3V9, Fig. 2 on page 3 shows it doesn't start to conduct until
you get above 3V, and 50mA doesn't even reach 5V:

http://www.pridmore.com/datasheets/BZX84CA07.pdf

Don't know what all the tolerances are, but so far nothing else seems to
have any chance of working in your application.

JK
I still haven't seen described how a "rock solid" 5V supply produces
such gross transients.

It doesn't. It's the signal lines that do.

[...]
So "induced" transients? How much actual energy?

Hard to say how much arrives, all I know from reports is that it has
frozen the uC at times. So we had to do something about it. We should be
good with those clamps. I was just wondering how far things can go
above abs max and for how long. Of course we won't be sending in 7V for
hours. When I redesign stuff I like to make it not just good enoug to
pass a test, but ideally almost bullet-proof.

The 7V limit only applies to those pins which are configurable I/O. If
it's an input-only pin, it can stand a whole lot more... some
processes can be as high as 50V.

They state a 5.8V abs max, 7V is just what one would typically see if a
monster pulse comes along and leans into the clamp diodes.

But you said uC... they're probably ALL I/O :-(

Yup, all I/O :-(

 

[Joerg]

Jim Thompson wrote:
Jim Thompson wrote:


On Sat, 29 Jun 2013 08:00:01 -0700, Joerg <[email protected]>
Oops, sorry, your are right. At a 5V-tolerant pin it would work.
Problem is that we only have 5V in the area where the protectors
are.
If you only have 5V around driving a 5V-tolerant pin... what are
you fretting over?

The concern are surge events that make these lines carry nasty
transients, which in turn can lock up or damage the uC.
So you have a "dirty" 5V supply?

No, rock-stable. But clamping against it will lead to up to 7V spike
amplitude if it's a huge transient. Those small clamper arrays can
easily reach 2V Vf or more for a few usec.
If you are dealing with 5V logic signals, isn't the threshold VCC/2, or
around 2.5V?

How about clamping with a 3.9V zener to ground with a small series
resistor to limit the transient current? As long as the desired signal
gets above the threshold, do you care if it goes all the way to 5V?

For a BZX84C3V9, Fig. 2 on page 3 shows it doesn't start to conduct until
you get above 3V, and 50mA doesn't even reach 5V:

http://www.pridmore.com/datasheets/BZX84CA07.pdf

Don't know what all the tolerances are, but so far nothing else seems to
have any chance of working in your application.

JK
I still haven't seen described how a "rock solid" 5V supply produces
such gross transients.

It doesn't. It's the signal lines that do.

[...]
So "induced" transients? How much actual energy?

Hard to say how much arrives, all I know from reports is that it has
frozen the uC at times. So we had to do something about it. We should be
good with those clamps. I was just wondering how far things can go
above abs max and for how long. Of course we won't be sending in 7V for
hours. When I redesign stuff I like to make it not just good enoug to
pass a test, but ideally almost bullet-proof.
The 7V limit only applies to those pins which are configurable I/O. If
it's an input-only pin, it can stand a whole lot more... some
processes can be as high as 50V.

They state a 5.8V abs max, 7V is just what one would typically see if a
monster pulse comes along and leans into the clamp diodes.

But you said uC... they're probably ALL I/O :-(

Yup, all I/O :-(

If you can get a handle on the energy involved (*), I can show you
some simple clamps that are reliable and tight on the limits.

(*) Just patch in a diode to +5V and current probe the data line.

I could design some clamps that are more precise but there is so little
space that just adding the teeny clamp arrays was already a challenge.
Also, it's thousands of miles away.