Searching \ for '[PIC] Missing silicon errata ?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/microchip/devices.htm?key=pic
Search entire site for: 'Missing silicon errata ?'.

Exact match. Not showing close matches.
PICList Thread
'[PIC] Missing silicon errata ?'
2007\03\05@132256 by Peter P.

picon face
I have a serious problem. PIC16F54-I/P with 3.58 MHz CF and internal caps, 5V
supply. The chip does not come out of reset and stops working as soon as any pin
configured as input is pulled to -0.7V or so (via a resistor, i.e. making use of
the 'inherent' internal clamp diode). Removing the negative voltage starts the
chip. Note that only a Schottky diode brings relief, a normal 1N4148 has too
much Vf !

The datasheet [1] specifies +/-20mA clamp current on any pin except TOCKI and
-0.6V..VDD+0.6V on any pin. My injected current is about -7mA (1k to -8.2V). How
do I interpret the datasheet wrt 'clamp current' in this context ?!

[1] "PIC16F5X Data Sheet" (DS41213C), PDF document
Page 61, "11.0 ELECTRICAL SPECIFICATIONS FOR PIC16F54/57", "Voltage on all other
pins with respect to VSS ... -0.6V to (VDD + 0.6V)", "Input clamp current, IIK
(VI < 0 or VI > VDD) ... 20 mA"

Imho the clamp current specification (input *AND* output, no less) is useless
and wrong (and dangerous - I have to modify my board). I would expect the maker
to test such things from time to time and issue a datasheet errata. In the past
this has happened to PICs with A/D where the voltage on any input is now given
as +/-0.3V and the same clamp current data (+/-20mA) is given (how come ?! - eg.
PIC16F62X datasheet - 17.0 electrical specs). Why do *I* have to find these
things ?

Any comments are welcome.

Peter


2007\03\05@160705 by Jan-Erik Soderholm

face picon face
Read the grey box on the same page.

Regards,
Jan-Erik.



Peter P. wrote:
{Quote hidden}

2007\03\05@162441 by Peiserma

flavicon
face
piclist-bounces@mit.edu wrote:
> Imho the clamp current specification (input *AND* output, no
> less) is useless and wrong (and dangerous - I have to modify
> my board).

[snip]

> Any comments are welcome.

The specs you listed come from Absolute Maximum ratings. The datasheet very clearly says this is a stress rating only and functional operation is not implied. The absolute max ratings only list the guaranteed limits under which the device will not emit the magic smoke.

Note there is still some debate wrt to the input protection diodes. I never use them intentionally, since doing so implies i'm violating the input voltage specs. The spec is Vss-0.6V. You say you put -8.2V on an input. Thats clearly out of spec. The only reason you might be measuring -0.6V is because the diodes are already conducting...


2007\03\05@165404 by Peter P.

picon face
Thanks for answering, but,

<peiserma <at> ridgid.com> writes:
> The specs you listed come from Absolute Maximum ratings. The datasheet very
> clearly says this is a stress
> rating only and functional operation is not implied. The absolute max ratings
> only list the guaranteed
> limits under which the device will not emit the magic smoke.
>
> Note there is still some debate wrt to the input protection diodes. I never
use > them intentionally, since
> doing so implies i'm violating the input voltage specs. The spec is Vss-0.6V.
> You say you put -8.2V on an
> input. Thats clearly out of spec. The only reason you might be measuring -0.6V
> is because the diodes are
> already conducting...

No, I put -8.2V *through* 1k on the pin. Even so, I don't think that it's
reasonable for the limit to be so stringent that a Schottky clamp must be used.
What's the use of a $0.5 MCU if it needs $1 in Schottky clamps to make 8 IO pins
that can be connected to usual external wires without freezing the chip. As you
know, external wiring will conduct glitches etc and these should be clamped by
the diodes. Yet in this chip they also freeze it. I am not sure if the WDT is
also frozen, I have not tested it enough.

By the way putting in an old 16F84 for testing works fine without clamps of any
kind (as expected).

thanks,
Peter P.


2007\03\05@172154 by Peter P.

picon face
Jan-Erik Soderholm <jan-erik.soderholm <at> telia.com> writes:

> Read the grey box on the same page.
>
> Regards,
> Jan-Erik.

Hi Jan-Erik,

yes I have read the box on the page and I am not new to this. What is new is
this Mimosa Pudica PIC that requires a Schottky clamp (normal silicon is not
enough as I showed, and a 16F84 fitted in its place for testing works great).
Now either the clamp spec is wrong (+/-20mA) or the input voltage range is
wrong (-0.6 etc). I.e. imho it is not possible to specify a clamp current of
20mA in a datasheet when one KNOWS that it cannot be attained without grossly
exceeding another parameter (the input clamp voltage). I had expected the spec
to mean that at about 0.6V clamp voltage the current will be 20mA ish and that
these two parameters are related rather intimately. Essentially the input clamp
parameters are the specs for the internal bulk diodes in the output Mosfets,
and as such, should be diode specs like If @ Vf. And if one KNOWS that the
chip is sensitive then one specs 0.3V clamp voltage, as in the 16F628 etc
datasheets, which is a clear warning to the developer to use a Schottky or
active clamp !

The voltage spec can only be attained with external clamps, and e.g. 1N4148 is
not enough, at 7mA (less than half the specified maximum clamp current) I still
have the problem, so a Schottky or active clamp must be used. This is not
reasonable imho. Also the oscillator amplitude seems to be affected, it drops to
~1Vpp at OSCOUT, with Vol ~= Vss. This would explain why the chip stops. This
does not happen with many other PICs where a normal silicon clamp is enough. I
would say that this is a bug, and that on this chip the core is too sensitive to
the voltage on external pins. I chose this chip specifically for low price and
adding $1 in clamps to a $0.5 chip will NOT go down well.

Speaking of extreme specs, Vpor = Vss as minimum start voltage to ensure reset
is also not raesonable imho. In an On/Off/On cycle the bus voltage seldomly goes
that low in low power devices (see table 11.1 page 84 DS41213C 16F5x datasheet).
I take standard measures like a discharge network and use WDT etc and safe
programming techniques to ensure reliability but some thing you just can't
design against imho.

Peter P.


2007\03\05@174223 by Jan-Erik Soderholm

face picon face
Peter P. wrote:

> No, I put -8.2V *through* 1k on the pin. Even so, I don't think that it's
> reasonable for the limit to be so stringent that a Schottky clamp must be used.

Fine.
Either you follow the specs in the data sheet, or you don't.
It's your design.

Best Regards,
Jan-Erik.

2007\03\05@184441 by Peter P.

picon face
Jan-Erik Soderholm <jan-erik.soderholm <at> telia.com> writes:

> Peter P. wrote:
>
> > No, I put -8.2V *through* 1k on the pin. Even so, I don't think that it's
> > reasonable for the limit to be so stringent that a Schottky clamp must be
> > used.
>
> Fine.
> Either you follow the specs in the data sheet, or you don't.
> It's your design.

Thanks for answering. This is very simple: There is basically no change in
written specs between the 16C54 and 16F54 in this area. The 16C54 (and a 16F84 I
also tested) withstands the use in this circuit perfectly, with -7mA drained
from an input pin, the voltage is less than -0.6V and everything is fine. With
16F54 the voltage is higher although the current is lower. So either I am
missing something very obvious OR something has indeed changed and the datasheet
needs updating. Imho.

Peter


2007\03\05@185646 by Peter P.

picon face
Jan-Erik Soderholm <jan-erik.soderholm <at> telia.com> writes:

> Fine.
> Either you follow the specs in the data sheet, or you don't.
> It's your design.

Come to think of it, you are right. But I still think that this chip is far too
sensitive and that the datasheet and the migration guide should indicate this
clearly. With the exception of analog input PICs I have met very few ICs where
the input voltage spec was tighter than 0.6V from the rails. It is immediately
obvious that any chip not withstanding those 0.6V must be clamped externally
with non-simple diode devices at least. So having very sensitive IO pins on an
embedded controller which is likely to work with external IO all the time sort
of defeats the purpose imho. And if so, there should be a clear warning in the
datasheet and in the migration guide.

thanks,
Peter


2007\03\06@020725 by Wouter van Ooijen

face picon face
> Thanks for answering. This is very simple: There is basically
> no change in
> written specs between the 16C54 and 16F54 in this area. The
> 16C54 (and a 16F84 I
> also tested) withstands the use in this circuit perfectly,
> with -7mA drained
> from an input pin, the voltage is less than -0.6V and
> everything is fine. With
> 16F54 the voltage is higher although the current is lower. So
> either I am
> missing something very obvious OR something has indeed
> changed and the datasheet
> needs updating. Imho.

If I understand you right: You used a chip outside its specs and it
worked. You use the newer chip this way and it fails. And you are
surprised?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2007\03\06@022416 by Peter P.

picon face
Wouter van Ooijen <wouter <at> voti.nl> writes:

> If I understand you right: You used a chip outside its specs and it
> worked. You use the newer chip this way and it fails. And you are
> surprised?

That is not what I was saying. What I was saying was, that the spec is lying or
wrong. The actual permitted Vclamp is probably Vdd+0.3V/Vss-0.3V as opposed to
the stated 0.6V. Iow, that chip cannot clamp at all, since its bulk diodes will
never open at +/-0.3V!. Then, clamping with a silicon diode (1N4148) did not fix
the problem, and I had to use Schottky. With the silicon diode I measured -0.59V
or so which is within spec, and the chip still did not work. Worse, more exact
Vclamp specs are not to be found in that datasheet. Of course you realize that
such a chip fitted with standard tranzorbs and suppression networks on IO pins
and connected to a normal circuit can and will be subjected to input voltage
excursions to -0.6V in normal operation (e.g. hum, induced voltage, capacitive
effects, galvanic bias from humidity, ground current etc), where the reverse
conducting tranzorb will 'catch' them, but too late, after they will have caused
a clock loss by the core ? So one would have to use a Schottky tranzorb or
something even more expensive ?

thanks,
Peter P.


2007\03\06@032014 by Wouter van Ooijen

face picon face
> With the silicon
> diode I measured -0.59V
> or so which is within spec, and the chip still did not work.

Within the working parameters or within the absolute maxima?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2007\03\06@043033 by Alan B. Pearce

face picon face
>So either I am missing something very obvious

Well, you are definitely doing that.

>OR something has indeed changed and the datasheet
>needs updating. Imho.

IMNSHO the data sheet is correct. As someone else pointed out, you are
looking at ABSOLUTE MAXIMUM ratings, NOT operating ratings.

The current rating for the diodes allows static discharge during handling -
not current while operating. These diodes are not specific components put
there for the purpose you are attempting to use them for, but intrinsic to
the manufacturing process, and made to have specific characteristics that
allow them to become protection against static charges during handling. As
such they have connections to "funny places" elsewhere in the chip
substrate, and as you observe, when the diode conducts it does other funny
things to the circuit which stops it working.

In short, despite other chips "apparently" working correctly, you cannot
rely on this, and must use an external schottky diode to protect the input
and stop the internal diode from conducting. Your talk of using a more
expensive component to protect a cheap microprocessor is not an uncommon
occurrence.

2007\03\06@043821 by Alan B. Pearce

face picon face
> [1] "PIC16F5X Data Sheet" (DS41213C), PDF document
> Page 61, "11.0 ELECTRICAL SPECIFICATIONS FOR PIC16F54/57", "Voltage on all
> other
> pins with respect to VSS ... -0.6V to (VDD + 0.6V)", "Input clamp current,
> IIK
> (VI < 0 or VI > VDD) ... 20 mA"

Further to my previous reply, check the DC Characteristics, page 64. This is
the bit you are obviously missing.

What is the "Input Low Voltage" parameter? In the very first parameter on
the table the minimum is given as Vss - NOT Vss-0.6V. The next parameter for
"Input High Voltage" similarly gives Vdd - Not Vdd+0.6V.

So the data sheet is correct, and you are trying to operate the chip outside
the limits.

2007\03\06@104947 by William Chops Westfield

face picon face

On Mar 6, 2007, at 1:38 AM, Alan B. Pearce wrote:

> So the data sheet is correct, and you are trying to
> operate the chip outside the limits.

I dunno.  If the chip behavior has changed and the datasheet
HASN'T changed, that seems like a microchip-side problem to me.
The inability to adequately clamp inputs with ordinary silicon
diodes seems like a pretty serious issue.  (Are you sure you
saw that with chips that hadn't been subject to the internal
diodes being used as clamps?  It's possible that THOSE chips
were damaged somehow...)

BillW

2007\03\06@110007 by Peiserma

flavicon
face
piclist-bounces@mit.edu wrote:
>> The spec is
>> Vss-0.6V. You say you put -8.2V on an input. Thats clearly out of
>> spec. The only reason you might be measuring -0.6V is because the
>> diodes are already conducting...
>
> No, I put -8.2V *through* 1k on the pin. Even so, I don't
> think that it's reasonable for the limit to be so stringent
> that a Schottky clamp must be used.


Hi Peter,

I think there's a fundamental misunderstanding here. I do understand that you're using a 1K resistor between the input and the pin. Let me try to explain: As you already know, the PIC input pin is high-impendance. This means that the voltage at the pin is going to be very close to -8.2V (actual value depends on the pin's impendance & characteristics of your input source). So there you are violating the specs. Now what happens is the protection diode conducts, which of course clamps the voltage at the pin. But that doesn't mean you are now suddenly within spec!

If you use e.g. the external diode you mentioned, or even a resistive divider, then the pin would not have seen a voltage outside specified range, the protection diodes would not have conducted, and life is good.

I hope this helps.

2007\03\06@112920 by Herbert Graf

flavicon
face
On Tue, 2007-03-06 at 07:24 +0000, Peter P. wrote:
> Wouter van Ooijen <wouter <at> voti.nl> writes:
>
> > If I understand you right: You used a chip outside its specs and it
> > worked. You use the newer chip this way and it fails. And you are
> > surprised?
>
> That is not what I was saying. What I was saying was, that the spec is lying or
> wrong. The actual permitted Vclamp is probably Vdd+0.3V/Vss-0.3V as opposed to
> the stated 0.6V.

Correct me if I'm wrong, but the specs you're looking at are in the
absolute maximum section of the datasheet. Those ratings simply state
you won't KILL the part by doing what the ratings describe, they say
NOTHING of whether the part will function or not.

For normal functions I can't see why you'd expect going below GND would
be a good idea.

The older chips worked? So what, you were out of operating spec with
them and got lucky, now it's biting you. I've been there myself.

I have a board with a voltage supply for some DRAM. Everything worked
fine, until a recent run of the boards resulted in no Vdd for the DRAMs.
Comparing the working board with the non working boards only revealed
one difference: the regulator chip was a rev. C on the new boards, and
was rev. B on the old boards. Turns out I had strapped the chips out of
spec, the old chips were fine with that, the new chips were not. Lesson
learned, don't go out of spec.

TTYL

2007\03\06@113511 by Herbert Graf

flavicon
face
On Tue, 2007-03-06 at 07:49 -0800, William Chops Westfield wrote:
> On Mar 6, 2007, at 1:38 AM, Alan B. Pearce wrote:
>
> > So the data sheet is correct, and you are trying to
> > operate the chip outside the limits.
>
> I dunno.  If the chip behavior has changed and the datasheet
> HASN'T changed, that seems like a microchip-side problem to me.

But you're talking about a change in behavior OUTSIDE of the specs.
That's like complaining that older parts could put a hole in the ceiling
when you powered them with 120V, and the new parts only give a puff of
smoke.

> The inability to adequately clamp inputs with ordinary silicon
> diodes seems like a pretty serious issue.  (Are you sure you
> saw that with chips that hadn't been subject to the internal
> diodes being used as clamps?  It's possible that THOSE chips
> were damaged somehow...)

But if you follow the specs the PIC should never see voltages on it's
pins outside of the Vdd-GND range, so why is this an issue?

TTYL

2007\03\06@114153 by Marcel duchamp

picon face
William Chops Westfield wrote:
> On Mar 6, 2007, at 1:38 AM, Alan B. Pearce wrote:
>
>> So the data sheet is correct, and you are trying to
>> operate the chip outside the limits.
>
> I dunno.  If the chip behavior has changed and the datasheet
> HASN'T changed, that seems like a microchip-side problem to me.
> The inability to adequately clamp inputs with ordinary silicon
> diodes seems like a pretty serious issue.

The PIC chip itself is made from, as you call it, "ordinary silicon"
Bill.  Adding an external "ordinary silicon" clamp diode simply puts two
 "ordinary silicon" diodes in parallel.

Which one will turn on first?

Flip a coin. In fact, since the PIC is powered up and dissipating energy
as heat, it will be (likely) warmer and have a lower Vf than the
external one so the external one will tend not to help.

Schottky clamps are "de rigeur" for clamping SILICON microcontrollers.




2007\03\06@114948 by Peiserma

flavicon
face
piclist-bounces@mit.edu wrote:
> The current rating for the diodes allows static discharge
> during handling - not current while operating. These diodes
> are not specific components put there for the purpose you are
> attempting to use them for, but intrinsic to the
> manufacturing process, and made to have specific
> characteristics that allow them to become protection against
> static charges during handling. As such they have connections
> to "funny places" elsewhere in the chip substrate, and as you
> observe, when the diode conducts it does other funny things
> to the circuit which stops it working.
>
> In short, despite other chips "apparently" working correctly,
> you cannot rely on this, and must use an external schottky
> diode to protect the input and stop the internal diode from
> conducting. Your talk of using a more expensive component to
> protect a cheap microprocessor is not an uncommon occurrence.

That's a nice explanation. This topic should probably be part of a FAQ.

2007\03\06@133705 by Peter P.

picon face
Marcel duchamp <marcel.duchamp <at> sbcglobal.net> writes:

> William Chops Westfield wrote:
> > On Mar 6, 2007, at 1:38 AM, Alan B. Pearce wrote:
> >
> >> So the data sheet is correct, and you are trying to
> >> operate the chip outside the limits.
> >
> > I dunno.  If the chip behavior has changed and the datasheet
> > HASN'T changed, that seems like a microchip-side problem to me.
> > The inability to adequately clamp inputs with ordinary silicon
> > diodes seems like a pretty serious issue.
>
> The PIC chip itself is made from, as you call it, "ordinary silicon"
> Bill.  Adding an external "ordinary silicon" clamp diode simply puts two
>   "ordinary silicon" diodes in parallel.
>
> Which one will turn on first?

The one that is more strongly doped. In particular, the 'reverse diodes' in
MOSFETs such as those in PICs are poor diodes as diodes go and the the external
diodes should turn on first. But with die shrinks doping can become violently
high and exceed that of external diodes. I tried to compare 16F54 and 16C54 bulk
diodes using a DVM on diode scale. Results are not very conclusive but the 16F84
seems to be different enough. Testing was with the DVM + at pin 5 (Vdd) and DVM
- at pin 6 (PORTB.0 - the pin that first gave me trouble). I also compared this
with 1N4148. I can't say that this is conclusive.

> Flip a coin. In fact, since the PIC is powered up and dissipating energy
> as heat, it will be (likely) warmer and have a lower Vf than the
> external one so the external one will tend not to help.
>
> Schottky clamps are "de rigeur" for clamping SILICON microcontrollers.

I strongly disagree. All commercial tranzorbs clamp at -0.6V (silicon Vf).
There are specialty tranzorbs that clamp lower. I accept that the voltage may
not go below -0.6V iff that can be ensured with an ordinary clamp. I know how
to make high speed accurate clamps that clamp to 50mV of the rails, but I don't
want to use such things in a project that uses a $0.5 micro. Even the two
external Schottky diode clamps I have to add for this project increase the
price by 3 cents, lus labor, plus board, plus plus.

And yes, I tested with 'virgin' chips and with chips that have seen the current
drain. The problem is the same. I did this fairly methodically, first w/o
diodes, then with 1N4148, then with 1N6263 and then BAT42. A 16C54(C?)JW works
even without diodes, ditto 16F84. 16F54 does not work w/o diodes, does not work
with 1N4147 (Vf is about 0.6V), and works with any Schottky.

One more thing: The OSCOUT amplitude on these chips is low, about 2Vpp, and
rides on GND. This is unusual. I use XT mode with ZTT series 3.58 MHz crystal
filters. With a different chip (16C54, 16F84) the OSC amplitude is healthy and
over 4Vpp. All this gives me a bad feeling about long term reliability. The
programmer seems to like the chips and they verify at high and low voltage but I
must say that they behave differently from most PICs I've used over the past ~13
years.

Peter P.


2007\03\06@134507 by Peter P.

picon face
<peiserma <at> ridgid.com> writes:

> If you use e.g. the external diode you mentioned, or even a resistive divider,
> then the pin would not have
> seen a voltage outside specified range, the protection diodes would not have
> conducted, and life is good.

Yes, it seems that one HAS GOT to use clamps external to the chip all the time.
But I had not expected the clamps to need to be Schottky. As I said before,
previous drop-in compatible chips were happy with silicon clamps, or none at all.

thanks all,
Peter P.




2007\03\06@165615 by wouter van ooijen

face picon face
> In short, despite other chips "apparently" working correctly, you
> cannot rely on this, and must use an external schottky diode to
> protect the input and stop the internal diode from conducting. Your
> talk of using a more expensive component to protect a cheap
> microprocessor is not an uncommon occurrence.

> That's a nice explanation. This topic should probably be part of a
FAQ.

No, what should be part of that FAQ is what "absolute maxima" means.
BTW: does anyone have a good webpage explaining that concept? IIRC the
definition is in a closed-source standard document :(

Wouter van Ooijen
(signature lost due to XP crash)

More... (looser matching)
- Last day of these posts
- In 2007 , 2008 only
- Today
- New search...