Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > sci.electronics.design > #429908 > unrolled thread

Input (protection) FET or npn

Back to article view | Back to sci.electronics.design

Contents

  Input (protection) FET or npn George Herold <gherold@teachspin.com> - 2016-09-30 07:52 -0700
    Re: Input (protection) FET or npn George Herold <gherold@teachspin.com> - 2016-09-30 08:05 -0700
      Re: Input (protection) FET or npn Frank Miles <fpm@u.washington.edu> - 2016-09-30 15:28 +0000
        Re: Input (protection) FET or npn George Herold <gherold@teachspin.com> - 2016-09-30 08:52 -0700
    Re: Input (protection) FET or npn John Larkin <jjlarkin@highlandtechnology.com> - 2016-09-30 08:35 -0700
      Re: Input (protection) FET or npn George Herold <gherold@teachspin.com> - 2016-09-30 08:54 -0700
    Re: Input (protection) FET or npn Joerg <news@analogconsultants.com> - 2016-09-30 08:47 -0700
      Re: Input (protection) FET or npn George Herold <gherold@teachspin.com> - 2016-09-30 09:14 -0700
        Re: Input (protection) FET or npn Joerg <news@analogconsultants.com> - 2016-09-30 09:50 -0700
    Re: Input (protection) FET or npn John S <Sophi.2@invalid.org> - 2016-09-30 12:19 -0500
      Re: Input (protection) FET or npn George Herold <gherold@teachspin.com> - 2016-09-30 11:40 -0700
        Re: Input (protection) FET or npn Joerg <news@analogconsultants.com> - 2016-09-30 12:38 -0700
        Re: Input (protection) FET or npn John Larkin <jjlarkin@highlandtechnology.com> - 2016-09-30 13:15 -0700

#429908 — Input (protection) FET or npn

Hi all a simple input circuit take a input pulse. And stretches it a bit.
(from previous post on pulse stretching.) 
The input looks like this.. (bad asci art)  

                 +5V
                  +--------+
                  |       R(1M)   
                  _     ---+        
                  ^     |  |  
                  |   |/c  +--->'HC14    
pulse in---R(100)-+---|    |
                  |   |\e  C(.1u)
                  _     |  |
                  ^     GND+
                  |
                 GND

100 ohms in series and diodes to the rails.  
I can then either drive a FET (2n7000) or npn (2n4401) 
both work fine.
The exact width of the output pulse is not critical. 
Which is more robust?

George H.   

[toc] | [next] | [standalone]

#429910

On Friday, September 30, 2016 at 10:52:48 AM UTC-4, George Herold wrote:
> Hi all a simple input circuit take a input pulse. And stretches it a bit.
> (from previous post on pulse stretching.) 
> The input looks like this.. (bad asci art)  
> 
>                  +5V
>                   +--------+
>                   |       R(1M)   
>                   _     ---+        
>                   ^     |  |  
>                   |   |/c  +--->'HC14    
> pulse in---R(100)-+---|    |
>                   |   |\e  C(.1u)
>                   _     |  |
>                   ^     GND+
>                   |
>                  GND
> 
> 100 ohms in series and diodes to the rails.  
> I can then either drive a FET (2n7000) or npn (2n4401) 
> both work fine.
> The exact width of the output pulse is not critical. 
> Which is more robust?
> 
> George H.

Oops there is also 10k ohm to ground on the base/gate of the transistor
so that it's off when no input is connected.  

GH

[toc] | [prev] | [next] | [standalone]

#429916

On Fri, 30 Sep 2016 08:05:41 -0700, George Herold wrote:

> On Friday, September 30, 2016 at 10:52:48 AM UTC-4, George Herold wrote:
>> Hi all a simple input circuit take a input pulse. And stretches it a bit.
>> (from previous post on pulse stretching.) 
>> The input looks like this.. (bad asci art)  
>> 
>>                  +5V
>>                   +--------+
>>                   |       R(1M)   
>>                   _     ---+        
>>                   ^     |  |  
>>                   |   |/c  +--->'HC14    
>> pulse in---R(100)-+---|    |
>>                   |   |\e  C(.1u)
>>                   _     |  |
>>                   ^     GND+
>>                   |
>>                  GND
>> 
>> 100 ohms in series and diodes to the rails.  
>> I can then either drive a FET (2n7000) or npn (2n4401) 
>> both work fine.
>> The exact width of the output pulse is not critical. 
>> Which is more robust?
>> 
>> George H.
> 
> Oops there is also 10k ohm to ground on the base/gate of the transistor
> so that it's off when no input is connected.  
> 
> GH

Either should work fine.  Unless you need the fastest possible
'reset', I'd increase the input resistor value (bipolar) or
add a resistor in series with the capacitor to limit surge
current at turn-on (t.c. << minimum reset time).

[toc] | [prev] | [next] | [standalone]

#429927

On Friday, September 30, 2016 at 11:28:24 AM UTC-4, cassiope wrote:
> On Fri, 30 Sep 2016 08:05:41 -0700, George Herold wrote:
> 
> > On Friday, September 30, 2016 at 10:52:48 AM UTC-4, George Herold wrote:
> >> Hi all a simple input circuit take a input pulse. And stretches it a bit.
> >> (from previous post on pulse stretching.) 
> >> The input looks like this.. (bad asci art)  
> >> 
> >>                  +5V
> >>                   +--------+
> >>                   |       R(1M)   
> >>                   _     ---+        
> >>                   ^     |  |  
> >>                   |   |/c  +--->'HC14    
> >> pulse in---R(100)-+---|    |
> >>                   |   |\e  C(.1u)
> >>                   _     |  |
> >>                   ^     GND+
> >>                   |
> >>                  GND
> >> 
> >> 100 ohms in series and diodes to the rails.  
> >> I can then either drive a FET (2n7000) or npn (2n4401) 
> >> both work fine.
> >> The exact width of the output pulse is not critical. 
> >> Which is more robust?
> >> 
> >> George H.
> > 
> > Oops there is also 10k ohm to ground on the base/gate of the transistor
> > so that it's off when no input is connected.  
> > 
> > GH
> 
> Either should work fine.  Unless you need the fastest possible
> 'reset', I'd increase the input resistor value (bipolar) or
> add a resistor in series with the capacitor to limit surge
> current at turn-on (t.c. << minimum reset time).

npn w/ more resistance.  Aye Captian.  

George H. 

[toc] | [prev] | [next] | [standalone]

#429919

On Fri, 30 Sep 2016 07:52:43 -0700 (PDT), George Herold
<gherold@teachspin.com> wrote:

>Hi all a simple input circuit take a input pulse. And stretches it a bit.
>(from previous post on pulse stretching.) 
>The input looks like this.. (bad asci art)  
>
>                 +5V
>                  +--------+
>                  |       R(1M)   
>                  _     ---+        
>                  ^     |  |  
>                  |   |/c  +--->'HC14    
>pulse in---R(100)-+---|    |
>                  |   |\e  C(.1u)
>                  _     |  |
>                  ^     GND+
>                  |
>                 GND
>
>100 ohms in series and diodes to the rails.  
>I can then either drive a FET (2n7000) or npn (2n4401) 
>both work fine.
>The exact width of the output pulse is not critical. 
>Which is more robust?
>
>George H.   

Why not go directly into the HC14? Then you could use a bigger series
resistor, which would protect the resistor, too.


-- 

John Larkin         Highland Technology, Inc

lunatic fringe electronics 

[toc] | [prev] | [next] | [standalone]

#429929

On Friday, September 30, 2016 at 11:36:15 AM UTC-4, John Larkin wrote:
> On Fri, 30 Sep 2016 07:52:43 -0700 (PDT), George Herold
> <gherold@teachspin.com> wrote:
> 
> >Hi all a simple input circuit take a input pulse. And stretches it a bit.
> >(from previous post on pulse stretching.) 
> >The input looks like this.. (bad asci art)  
> >
> >                 +5V
> >                  +--------+
> >                  |       R(1M)   
> >                  _     ---+        
> >                  ^     |  |  
> >                  |   |/c  +--->'HC14    
> >pulse in---R(100)-+---|    |
> >                  |   |\e  C(.1u)
> >                  _     |  |
> >                  ^     GND+
> >                  |
> >                 GND
> >
> >100 ohms in series and diodes to the rails.  
> >I can then either drive a FET (2n7000) or npn (2n4401) 
> >both work fine.
> >The exact width of the output pulse is not critical. 
> >Which is more robust?
> >
> >George H.   
> 
> Why not go directly into the HC14? Then you could use a bigger series
> resistor, which would protect the resistor, too.
> 
> 
> -- 
> 
> John Larkin         Highland Technology, Inc
> 
> lunatic fringe electronics

I've got to have this long input pulse to make sure things have enough time to 
sync-up.  (a very slow clock.)  

George H. 

[toc] | [prev] | [next] | [standalone]

#429924

On 2016-09-30 07:52, George Herold wrote:
> Hi all a simple input circuit take a input pulse. And stretches it a bit.
> (from previous post on pulse stretching.)
> The input looks like this.. (bad asci art)
>
>                   +5V
>                    +--------+
>                    |       R(1M)
>                    _     ---+
>                    ^     |  |
>                    |   |/c  +--->'HC14
> pulse in---R(100)-+---|    |
>                    |   |\e  C(.1u)
>                    _     |  |
>                    ^     GND+
>                    |
>                   GND
>
> 100 ohms in series and diodes to the rails.
> I can then either drive a FET (2n7000) or npn (2n4401)
> both work fine.
> The exact width of the output pulse is not critical.
> Which is more robust?
>

The FET is more robust. The main reason being that in this constellation 
the BE junction gets hit with the full ESD current if the spike is 
positive. The abs max base current peak for the 2N4401 is only 200mA.

http://media.digikey.com/pdf/Data%20Sheets/NXP%20PDFs/2N4401.pdf

The upper diode has no effect with the BJT in this scenario. Before that 
comes the BE junction has gone phut.

Hanging a 0.1uF emitter to ground is not so nice, high peak current. 
Won't be destructive but it isn't considered very kosher. With a FET it 
can be ok if the Rdson is guaranteed high enough so the abs max current 
won't be exceeded. Else use a small resistor in series with the drain.

-- 
Regards, Joerg

http://www.analogconsultants.com/

[toc] | [prev] | [next] | [standalone]

#429934

On Friday, September 30, 2016 at 11:47:15 AM UTC-4, Joerg wrote:
> On 2016-09-30 07:52, George Herold wrote:
> > Hi all a simple input circuit take a input pulse. And stretches it a bit.
> > (from previous post on pulse stretching.)
> > The input looks like this.. (bad asci art)
> >
> >                   +5V
> >                    +--------+
> >                    |       R(1M)
> >                    _     ---+
> >                    ^     |  |
> >                    |   |/c  +--->'HC14
> > pulse in---R(100)-+---|    |
> >                    |   |\e  C(.1u)
> >                    _     |  |
> >                    ^     GND+
> >                    |
> >                   GND
> >
> > 100 ohms in series and diodes to the rails.
> > I can then either drive a FET (2n7000) or npn (2n4401)
> > both work fine.
> > The exact width of the output pulse is not critical.
> > Which is more robust?
> >
> 
> The FET is more robust. The main reason being that in this constellation 
> the BE junction gets hit with the full ESD current if the spike is 
> positive. The abs max base current peak for the 2N4401 is only 200mA.
> 
> http://media.digikey.com/pdf/Data%20Sheets/NXP%20PDFs/2N4401.pdf
> 
> The upper diode has no effect with the BJT in this scenario. Before that 
> comes the BE junction has gone phut.
OK more base resistance then as suggested above.  
> 
> Hanging a 0.1uF emitter to ground is not so nice, high peak current. 
> Won't be destructive but it isn't considered very kosher. With a FET it 
> can be ok if the Rdson is guaranteed high enough so the abs max current 
> won't be exceeded. Else use a small resistor in series with the drain.

OK, so the FET with 10-20 ohms in the drain line.  
https://www.fairchildsemi.com/datasheets/2N/2N7000.pdf
(5v/500mA) 
Is there any reason not to increase the gate resistance too?  
(50 pF Cin.... ~50 ns TC...)

George H. 
  
> 
> -- 
> Regards, Joerg
> 
> http://www.analogconsultants.com/

[toc] | [prev] | [next] | [standalone]

#429940

On 2016-09-30 09:14, George Herold wrote:
> On Friday, September 30, 2016 at 11:47:15 AM UTC-4, Joerg wrote:
>> On 2016-09-30 07:52, George Herold wrote:
>>> Hi all a simple input circuit take a input pulse. And stretches it a bit.
>>> (from previous post on pulse stretching.)
>>> The input looks like this.. (bad asci art)
>>>
>>>                    +5V
>>>                     +--------+
>>>                     |       R(1M)
>>>                     _     ---+
>>>                     ^     |  |
>>>                     |   |/c  +--->'HC14
>>> pulse in---R(100)-+---|    |
>>>                     |   |\e  C(.1u)
>>>                     _     |  |
>>>                     ^     GND+
>>>                     |
>>>                    GND
>>>
>>> 100 ohms in series and diodes to the rails.
>>> I can then either drive a FET (2n7000) or npn (2n4401)
>>> both work fine.
>>> The exact width of the output pulse is not critical.
>>> Which is more robust?
>>>
>>
>> The FET is more robust. The main reason being that in this constellation
>> the BE junction gets hit with the full ESD current if the spike is
>> positive. The abs max base current peak for the 2N4401 is only 200mA.
>>
>> http://media.digikey.com/pdf/Data%20Sheets/NXP%20PDFs/2N4401.pdf
>>
>> The upper diode has no effect with the BJT in this scenario. Before that
>> comes the BE junction has gone phut.
> OK more base resistance then as suggested above.


Depends on the ESD rating you need and small resistors can be "jumped" 
by an ESD pulse. In the end it all depends on where this is used. I've 
had cases where ESD on outdoor equipment was extreme and caused a lot of 
grief until we fixed all areas where it could come in. For example, 
sneakers plus resin board decking (the artificial wood kind) plus a long 
enough walking distance are a recipe for an impressive zap.

>>
>> Hanging a 0.1uF emitter to ground is not so nice, high peak current.
>> Won't be destructive but it isn't considered very kosher. With a FET it
>> can be ok if the Rdson is guaranteed high enough so the abs max current
>> won't be exceeded. Else use a small resistor in series with the drain.
>
> OK, so the FET with 10-20 ohms in the drain line.
> https://www.fairchildsemi.com/datasheets/2N/2N7000.pdf
> (5v/500mA)
> Is there any reason not to increase the gate resistance too?
> (50 pF Cin.... ~50 ns TC...)
>

Sure you can. Keep in mind for your timing calculation that the gate 
capacitance is subject to process tolerances and with this FET can be 
anywhere between 10pF and 50pF.

-- 
Regards, Joerg

http://www.analogconsultants.com/

[toc] | [prev] | [next] | [standalone]

#429948

On 9/30/2016 9:52 AM, George Herold wrote:
> Hi all a simple input circuit take a input pulse. And stretches it a bit.
> (from previous post on pulse stretching.)
> The input looks like this.. (bad asci art)
>
>                  +5V
>                   +--------+
>                   |       R(1M)
>                   _     ---+
>                   ^     |  |
>                   |   |/c  +--->'HC14
> pulse in---R(100)-+---|    |
>                   |   |\e  C(.1u)
>                   _     |  |
>                   ^     GND+
>                   |
>                  GND
>
> 100 ohms in series and diodes to the rails.
> I can then either drive a FET (2n7000) or npn (2n4401)
> both work fine.
> The exact width of the output pulse is not critical.
> Which is more robust?
>
> George H.

What is the nature of the input pulse? 5V, 50V, 500V, etc? Energy 
capability? 1us, 10us, 1000us, etc?

Caution: If the input pulse is very energetic and of sufficient 
voltage/energy, it could raise your 5V rail to device destruction.

[toc] | [prev] | [next] | [standalone]

#429954

On Friday, September 30, 2016 at 1:18:56 PM UTC-4, John S wrote:
> On 9/30/2016 9:52 AM, George Herold wrote:
> > Hi all a simple input circuit take a input pulse. And stretches it a bit.
> > (from previous post on pulse stretching.)
> > The input looks like this.. (bad asci art)
> >
> >                  +5V
> >                   +--------+
> >                   |       R(1M)
> >                   _     ---+
> >                   ^     |  |
> >                   |   |/c  +--->'HC14
> > pulse in---R(100)-+---|    |
> >                   |   |\e  C(.1u)
> >                   _     |  |
> >                   ^     GND+
> >                   |
> >                  GND
> >
> > 100 ohms in series and diodes to the rails.
> > I can then either drive a FET (2n7000) or npn (2n4401)
> > both work fine.
> > The exact width of the output pulse is not critical.
> > Which is more robust?
> >
> > George H.
> 
> What is the nature of the input pulse? 5V, 50V, 500V, etc? Energy 
> capability? 1us, 10us, 1000us, etc?
> 
> Caution: If the input pulse is very energetic and of sufficient 
> voltage/energy, it could raise your 5V rail to device destruction.

Well, I get to spec the input pulse.  5V logic, ~10 us minimum pulse width.  
But this goes to students and I have to assume almost anything that can 
come out of a function generator might be put into it.  ~+/- 20 V.  
And then ESD is always possible.  
I'll have a through hole FET, (or npn) so at some point the FET becomes a fuse...
for those extrally energetic students.   

George H. 

[toc] | [prev] | [next] | [standalone]

#429959

On 2016-09-30 11:40, George Herold wrote:
> On Friday, September 30, 2016 at 1:18:56 PM UTC-4, John S wrote:
>> On 9/30/2016 9:52 AM, George Herold wrote:
>>> Hi all a simple input circuit take a input pulse. And stretches it a bit.
>>> (from previous post on pulse stretching.)
>>> The input looks like this.. (bad asci art)
>>>
>>>                   +5V
>>>                    +--------+
>>>                    |       R(1M)
>>>                    _     ---+
>>>                    ^     |  |
>>>                    |   |/c  +--->'HC14
>>> pulse in---R(100)-+---|    |
>>>                    |   |\e  C(.1u)
>>>                    _     |  |
>>>                    ^     GND+
>>>                    |
>>>                   GND
>>>
>>> 100 ohms in series and diodes to the rails.
>>> I can then either drive a FET (2n7000) or npn (2n4401)
>>> both work fine.
>>> The exact width of the output pulse is not critical.
>>> Which is more robust?
>>>
>>> George H.
>>
>> What is the nature of the input pulse? 5V, 50V, 500V, etc? Energy
>> capability? 1us, 10us, 1000us, etc?
>>
>> Caution: If the input pulse is very energetic and of sufficient
>> voltage/energy, it could raise your 5V rail to device destruction.
>
> Well, I get to spec the input pulse.  5V logic, ~10 us minimum pulse width.
> But this goes to students and I have to assume almost anything that can
> come out of a function generator might be put into it.  ~+/- 20 V.
> And then ESD is always possible.
> I'll have a through hole FET, (or npn) so at some point the FET becomes a fuse...
> for those extrally energetic students.
>

I was designing electronics for one of the institutes at my university 
to up my student's budget a bit. On one of the first projects I asked 
"Is this a one-off for one of our labs or small series?" ... "Oh, yes, 
small series so it needs to be Ph.D.-proof". Then I knew what to do 8-)

You might want to place a big TVS across the supply and some sort of 
fuse, in case someone reverses polarity or turned the power supply on 
and the guy before had it on 24V. That could blow the lid off of the 
74HC14. I try to use CD4000 logic wherever possible because a TVS isn't 
a very exact spec part.

-- 
Regards, Joerg

http://www.analogconsultants.com/

[toc] | [prev] | [next] | [standalone]

#429968

On Fri, 30 Sep 2016 11:40:13 -0700 (PDT), George Herold
<gherold@teachspin.com> wrote:

>On Friday, September 30, 2016 at 1:18:56 PM UTC-4, John S wrote:
>> On 9/30/2016 9:52 AM, George Herold wrote:
>> > Hi all a simple input circuit take a input pulse. And stretches it a bit.
>> > (from previous post on pulse stretching.)
>> > The input looks like this.. (bad asci art)
>> >
>> >                  +5V
>> >                   +--------+
>> >                   |       R(1M)
>> >                   _     ---+
>> >                   ^     |  |
>> >                   |   |/c  +--->'HC14
>> > pulse in---R(100)-+---|    |
>> >                   |   |\e  C(.1u)
>> >                   _     |  |
>> >                   ^     GND+
>> >                   |
>> >                  GND
>> >
>> > 100 ohms in series and diodes to the rails.
>> > I can then either drive a FET (2n7000) or npn (2n4401)
>> > both work fine.
>> > The exact width of the output pulse is not critical.
>> > Which is more robust?
>> >
>> > George H.
>> 
>> What is the nature of the input pulse? 5V, 50V, 500V, etc? Energy 
>> capability? 1us, 10us, 1000us, etc?
>> 
>> Caution: If the input pulse is very energetic and of sufficient 
>> voltage/energy, it could raise your 5V rail to device destruction.
>
>Well, I get to spec the input pulse.  5V logic, ~10 us minimum pulse width.  
>But this goes to students and I have to assume almost anything that can 
>come out of a function generator might be put into it.  ~+/- 20 V.  
>And then ESD is always possible.  
>I'll have a through hole FET, (or npn) so at some point the FET becomes a fuse...
>for those extrally energetic students.   
>
>George H. 

Maybe use one of the gate-protected 2N7002s, and a biggish series gate
resistor, and maybe a cap to ground for luck.




-- 

John Larkin         Highland Technology, Inc
picosecond timing   precision measurement 

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

[toc] | [prev] | [standalone]

Back to top | Article view | sci.electronics.design

csiph-web