Searching \ for '[EE] Transient Suppression - was Re: [ee] 3904 bas' 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/index.htm?key=3904+bas+EE%5D
Search entire site for: 'Transient Suppression - was Re: [ee] 3904 bas'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Transient Suppression - was Re: [ee] 3904 bas'
2006\02\06@011927 by Bob Axtell

face picon face
Spehro Pefhany wrote:

<respectful snip>

>A diode across the coil has been suggested, and is the simplest way of
>handling the coil inductance. A 1N4148 is fine for any relay you'd be
>likely to use.
>  
>
Alas, sorry,  it is NOT fine and is no longer considered good
engineering practice. I see
this idea all over the place, and this is bad information. Even relay
manufacturers include
a diode, but it is not good practice.

The 1N4148 is a "faster" general purpose silicon diode, used at one time
because nothing
any better was available. The 1N4148 cannot switch fast enough to
suppress the edges
of an inductive spike, so was most successful when a cap in series with
a resistor is in parallel with
the coil as well as the 1N4148. Moreover, it cannot handle the high
current pulse of the inductive
collapse. I've seen wholesale device failures with this setup, even with
tiny 5V relays.

Use a Transorber, Varistor,  or TVS device, with a voltage rating just
above the voltage being
applied to the relay or motor. If need be, just trust me on this one,
its the truth. Go to the
technical documents page of any transient voltage suppressor maker and
you will understand
what people are doing. These were originally invented by GE.

Slow oscilloscopes won't pickup the unsupressed spike, so this can't be
easily proven, but if you
own a small portable radio, you can pick up the unsuppressed spike  on
the AM spectrum. With a
1N4148 it will still be present, with little improvement. If you then
use a transorber or TVS placed
AT THE RELAY COIL, the spike will be so suppressed that you can't pick
it up (but if the suppressor
is not placed as close to the relay as possible, the RF will still be
generated by radiating through the
conductors going from the relay coil to the suppressor).   Try it.

--Bob

{Quote hidden}

--
Note: To protect our network,
attachments must be sent to
.....attachKILLspamspam@spam@engineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

2006\02\06@035123 by Chen Xiao Fan

face
flavicon
face


> -----Original Message-----
> From: piclist-bouncesspamKILLspammit.edu
> [.....piclist-bouncesKILLspamspam.....mit.edu] On Behalf Of Bob Axtell
>
> Spehro Pefhany wrote:
>> A diode across the coil has been suggested, and is the
>> simplest way of handling the coil inductance. A 1N4148
>> is fine for any relay you'd be likely to use.
>
> Use a Transorber, Varistor,  or TVS device, with a voltage
> rating just above the voltage being applied to the relay or motor.
> If need be, just trust me on this one,  its the truth. Go to the
> technical documents page of any transient voltage suppressor
> maker and you will understand what people are doing. These were
> originally invented by GE.

I tend to think a 1N4148/4448 will be good enough for low
duty relays since the coil resistance is fairly big in
this case. But I must agree with you that TVS/Varistors
are much better protection device than a simple and
cheap 1N4148/4448.

Regards,
Xiaofan

2006\02\06@054554 by Spehro Pefhany

picon face
At 11:19 PM 2/5/2006 -0700, you wrote:
>Spehro Pefhany wrote:
>
><respectful snip>
>
> >A diode across the coil has been suggested, and is the simplest way of
> >handling the coil inductance. A 1N4148 is fine for any relay you'd be
> >likely to use.
> >
> >
>Alas, sorry,  it is NOT fine and is no longer considered good
>engineering practice. I see
>this idea all over the place, and this is bad information. Even relay
>manufacturers include
>a diode, but it is not good practice.

It's fine from the pov of protecting the transistor. Hundreds of millions
are in daily use. You won't see datasheet performance on the relay, though.

What authority suggests that it is "not good practice"? And why would it
change? BTW, some of my designs use relays by the sh..er... pallet-load.

>The 1N4148 is a "faster" general purpose silicon diode, used at one time
>because nothing
>any better was available. The 1N4148 cannot switch fast enough to
>suppress the edges
>of an inductive spike, so was most successful when a cap in series with
>a resistor is in parallel with
>the coil as well as the 1N4148.

Distributed capacitance of the coil will limit dV/dt, although in more
sophisticated designs it may be desirable to add some external capacitance-
but not for protection reasons. You don't necessarily need a series resistor,
current will be limited by beta. For protection, even a slow-as-molasses
1N400x diode is fine. There are no nanosecond edges here to be concerned
with.

>  Moreover, it cannot handle the high
>current pulse of the inductive
>collapse. I've seen wholesale device failures with this setup, even with
>tiny 5V relays.

"High current pulse"? Where does this idea come from anyway? I have heard
this misconception before. The peak current through the diodes will never
exceed the steady-state current through the relay coil. There is no magic
here, the coil is generally just acting as an inductor (with some small
variations due to the magnetic stuff moving around, which can be ignored
for this purpose).

The "strange" thing that can cause transistor failures is violating
the SOA of the BJT during "off" switching because of the inductive
load. Small transistors  often don't have the safe operating area
explicitly defined, so it's best to err on the side of using a hefty part,
vis-a-vis the coil current, as I suggested.

See, for example, Figure 3 on this TIP31 data sheet:
http://pubpages.unh.edu/~aperkins/pdf/TIP-devices/TIP31.pdf


>Use a Transorber, Varistor,  or TVS device, with a voltage rating just
>above the voltage being
>applied to the relay or motor. If need be, just trust me on this one,
>its the truth. Go to the
>technical documents page of any transient voltage suppressor maker and
>you will understand
>what people are doing. These were originally invented by GE.

Expensive and totally unnecessary from the pov of the transistor. They are
actually *harder* on the transistor because the voltage goes higher.

{Quote hidden}

A TVS is useful where high currents are involved- for power supply
protection from load dump on automotive systems, where lightning may be
involved (that's where fast clamping is good). For relay coils, the
best approach if you feel like using a TVS is to use a diode or
regular zener and send the price difference directly to me. ;-)

One exception- if the layout is poor or if a crummy relay is used, it's
possible to have coil-contact flashover, especially from lightning. In
that case, a TVS can protect the driver (eg. telecom situation). But a
properly rated relay with solid-state (Transil, TVS, MOV) or spark
arrestors at the line is going to be more reliable long-term.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
EraseMEspeffspam_OUTspamTakeThisOuTinterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2006\02\06@071223 by Jinx

face picon face
> Use a Transorber, Varistor,  or TVS device, with a voltage rating
> just above the voltage being applied to the relay or motor

Thanks for pointing that out Bob. I'm in the process of re-vamping
a multi-relay circuit that controls motors and your advice is timely

A browse through retailers (Mouser and RS) led me to AVX

http://www.avxcorp.com/docs/catalogs/tguard.pdf

2006\02\06@080612 by Bob Axtell

face picon face
Spehro Pefhany wrote:

{Quote hidden}

I was more concerned with failure of the 1N4148 itself.


{Quote hidden}

Actually, yes, the 1N4005 would be a better choice.

{Quote hidden}

Then why are these TVS devices becoming smaller and smaller? They are
available for
$0.25 USD in 0603 sizes. There is a huge market for inductive
suppression (although
I think the small ones are to allow US products to pass EU ESD specs).

>One exception- if the layout is poor or if a crummy relay is used, it's
>possible to have coil-contact flashover, especially from lightning. In
>that case, a TVS can protect the driver (eg. telecom situation). But a
>properly rated relay with solid-state (Transil, TVS, MOV) or spark
>arrestors at the line is going to be more reliable long-term.
>
>  
>
I spent 6 weeks solving failures on financial instruments (CC readers)
and identified the
1N4148 across a telecom relay as the primary problem... it was shorting
after a few months
in service, and taking out the internal fuse on the PCB. The fix was to
use GE TVS in place
of the 1N4148.

Its OK, Spehro, Sorry I snapped...I am just reacting to my experience.
We agree on almost
everything else.  An item here and there won't hurt.

--Bob

{Quote hidden}

--
Note: To protect our network,
attachments must be sent to
@spam@attachKILLspamspamengineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

2006\02\06@081247 by Bob Axtell

face picon face
Jinx wrote:

>>Use a Transorber, Varistor,  or TVS device, with a voltage rating
>>just above the voltage being applied to the relay or motor
>>    
>>
>
>Thanks for pointing that out Bob. I'm in the process of re-vamping
>a multi-relay circuit that controls motors and your advice is timely
>
>A browse through retailers (Mouser and RS) led me to AVX
>
>http://www.avxcorp.com/docs/catalogs/tguard.pdf
>
>  
>
Wow, what an awesome assortment of ratings and sizes!
Thanks for the find, Jinx!

--Bob

--
Note: To protect our network,
attachments must be sent to
KILLspamattachKILLspamspamengineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

2006\02\06@095501 by Larry G. Nelson Sr.

picon face
You do not want to use a transorb for a repetitive application like this. Transorbs have a limited life. Every spike they see shortens this life. When they fail, and they will, they usually fail shorted causing even more problems.
Larry


---- Bob Axtell <RemoveMEengineerTakeThisOuTspamcotse.net> wrote:
{Quote hidden}

> --

2006\02\06@095949 by Spehro Pefhany

picon face
At 06:06 AM 2/6/2006 -0700, you wrote:

>I spent 6 weeks solving failures on financial instruments (CC readers)
>and identified the
>1N4148 across a telecom relay as the primary problem... it was shorting
>after a few months
>in service, and taking out the internal fuse on the PCB. The fix was to
>use GE TVS in place
>of the 1N4148.

Interesting. Were they concentrated geographically or by production lot?

>Its OK, Spehro, Sorry I snapped...I am just reacting to my experience.
>We agree on almost
>everything else.  An item here and there won't hurt.

Absolutely! No problem, mon.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
RemoveMEspeffspamTakeThisOuTinterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2006\02\06@120352 by Bob Axtell

face picon face
Spehro Pefhany wrote:

>At 06:06 AM 2/6/2006 -0700, you wrote:
>
>  
>
>>I spent 6 weeks solving failures on financial instruments (CC readers)
>>and identified the
>>1N4148 across a telecom relay as the primary problem... it was shorting
>>after a few months
>>in service, and taking out the internal fuse on the PCB. The fix was to
>>use GE TVS in place
>>of the 1N4148.
>>    
>>
>
>Interesting. Were they concentrated geographically or by production lot?
>
>  
>
No, it was all over the place. Production lots are always a worry, but
not this time.

This company leases a million of 'em.

--Bob


{Quote hidden}

--
Note: To protect our network,
attachments must be sent to
EraseMEattachspamengineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

2006\02\06@125102 by Dwayne Reid

flavicon
face
At 07:55 AM 2/6/2006, Larry G. Nelson Sr. wrote:
>You do not want to use a transorb for a repetitive application like
>this. Transorbs have a limited life. Every spike they see shortens
>this life. When they fail, and they will, they usually fail shorted
>causing even more problems.
>Larry

Hi there, Larry.

Can you point me to some data that supports this?  I know that MOVs
have an associated wear factor but I've NEVER seen anything similar
that in regards to a Transorb.

I use a LOT of Transorbs in various places and the only failures I
can associate with them is when dealing with severe load dump issues
in heavy earth-moving machinery.  Subsequent investigation revealed
that the peak power ratings on the failed devices were exceeded by at
least an order of magnitude.  In other words, not the fault of the TVS.

But I've not aware of any "wear" issues with TVS devices and would
very much like to confirm or deny that.

dwayne

--
Dwayne Reid   <RemoveMEdwaynerEraseMEspamEraseMEplanet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax

Celebrating 22 years of Engineering Innovation (1984 - 2006)
 .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-
    `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'
Do NOT send unsolicited commercial email to this email address.
This message neither grants consent to receive unsolicited
commercial email nor is intended to solicit commercial email.

2006\02\06@125735 by Michael Rigby-Jones

picon face


{Quote hidden}

Is the use of diodes for back EMF protection actualy documented as bad practice anywhere?  The only negative I've ever seen is that due to the low clamping voltage, the fall time of the current is quite long, delaying relay opening.  Worse case the increase in current when the armature seperates could potnentialy cause the contacts to be closed again, which is obvioulsy not acceptable in some applications.

{Quote hidden}

The 1N4148 is good for application up to many 10's (possibly 100's?) of MHz, is the rise time of the induced voltage really too fast? I guess that all depends on the speed of the driver but the commonly used saturated bipolar device is comparatively slow to switch off.

Fairchilds 1N4148 is rated for 1 Amp forward current for 1 second at a Vf of 1.5 volts.  I've not done the calcs but this feels like it would be adequate for the majorty of small relay applications.  Then again, I guess there may be an awfull lot of inferior devices being sold as 1N4148's that have nothing like these specs.  Still though a failure in this case could only be attributed to an under rated component.

Regards

Mike

=======================================================================
This e-mail is intended for the person it is addressed to only. The
information contained in it may be confidential and/or protected by
law. If you are not the intended recipient of this message, you must
not make any use of this information, or copy or show it to any
person. Please contact us immediately to tell us that you have
received this e-mail, and return the original to us. Any use,
forwarding, printing or copying of this message is strictly prohibited.
No part of this message can be considered a request for goods or
services.
=======================================================================

2006\02\06@133227 by David VanHorn

picon face
On 2/6/06, Michael Rigby-Jones <EraseMEMichael.Rigby-JonesspamspamspamBeGonebookham.com> wrote:
>
>   Worse case the increase in current when the armature seperates could
> potnentialy cause the contacts to be closed again, which is obvioulsy not
> acceptable in some applications.



Wouldn't that be a violation of conservation of energy?
Current would have had to decay to the dropout point for the armature to
start moving, and then you're supposing that this motion results in enough
energy to stop that same motion (already 100% yield) and then reverse it?



> The 1N4148 is good for application up to many 10's (possibly 100's?) of
> MHz, is the rise time of the induced voltage really too fast? I guess that
> all depends on the speed of the driver but the commonly used saturated
> bipolar device is comparatively slow to switch off.


I've seen some pretty abysmal layout practices, where the diode's turn-on
would be extended quite a bit.

2006\02\06@133353 by William Chops Westfield

face picon face
On Feb 6, 2006, at 9:03 AM, Bob Axtell wrote:

> identified the 1N4148 across a telecom relay as the primary problem.

I always thought signal diodes were a little fragile in this
application, and tend toward 1n400x rectifiers instead.  But
based on your first message, if speed is an issue, they might
be worse.  But if the diodes blew in that design, that's not
a diode speed issue, is it?  And are the protection diodes in
something like a ULN2003 that much faster than discreet diodes,
or are they just useless marketing ploys?

Sigh.  I though I knew what I was doing, and now I'm all confused!

BillW

2006\02\06@134148 by Larry G. Nelson Sr.

picon face
You are correct. MOVs are what I was primarily thinking of and my fingers typed transorb. I too have seen transorbs fail in automotive based applications and my mind probably extrapolated to both. I will look around to see if I can find anything relating to transorbs directly for the degradation. Load dumps in cars and trucks can cause surges A LOT higher than expected ie 200volts for 20 ms on a 24 V system in a transit bus when the AC cycled.
Larry


---- Dwayne Reid <RemoveMEdwaynerKILLspamspamplanet.eon.net> wrote:
{Quote hidden}

> --

2006\02\06@135113 by Enrico Schuerrer

picon face

-----Ursprüngliche Nachricht-----
Von: "Michael Rigby-Jones" <spamBeGoneMichael.Rigby-JonesSTOPspamspamEraseMEbookham.com>
An: "Microcontroller discussion list - Public." <KILLspampiclistspamBeGonespammit.edu>
Gesendet: Montag, 06. Februar 2006 18:57
Betreff: RE: [EE] Transient Suppression - was Re: [ee] 3904 base resistor


[rest snipped]

>
> The 1N4148 is good for application up to many 10's (possibly 100's?) of MHz, is the rise time of the induced voltage really too fast? I guess that all depends on the speed of the driver but the commonly used saturated bipolar device is comparatively slow to switch off.
>
> Fairchilds 1N4148 is rated for 1 Amp forward current for 1 second at a Vf of 1.5 volts.  I've not done the calcs but this feels like it would be adequate for the majorty of small relay applications.  Then again, I guess there may be an awfull lot of inferior devices being sold as 1N4148's that have nothing like these specs.  Still though a failure in this case could only be attributed to an under rated component.

I don't think that the problem is speed or current - the problem is the voltage. A 1N4148 is rated with 100 Volts and this could be less if you switch off a relay. From my experiences a 1N4007 as a general purpose diode with Vrrm = 1000V and much reserves in non repetitive current will do the job without problems and cost nearly nothing.

Regards

Enrico


2006\02\06@140114 by David VanHorn

picon face
>
>
> Sigh.  I though I knew what I was doing, and now I'm all confused!


Well, the 2003 diodes are better than nothing, most of the time.

I once was debugging a thermal printer that had problems with the occasional
"halt and catch fire"..  The design was in production, but was going through
a lot of printheads with no apparent cause.

After a bit of negotiation, we finally got the schematic.
The printhead elements were run from unregulated power, with an ADC to read
the voltage and correct burn time.  They were connected from unreg, through
the print elements, to a ULN2003.
Turns out they had connected the protection diode common to VCC.
This caused some really odd behaviour, since the current through the
printhead, and protection diodes was usually what was running the system,
NOT the output of the linear regulator.  The system would float around
5.24Vnominal, with VCC taking the expected hits whenever printhead
elements were
on.  I never could track things beyond that point to a specific cause, but
in all cases cutting the diode pin on the 2003 cured any problems they were
having, and we usually didn't see those units again.

I suspected that there was some problem in the regulation caused by the
output of the reg going higher than the regulation voltage, then back below.
Not a state that most linear regs are designed to operate in.

2006\02\06@142152 by Richard Prosser

picon face
On 07/02/06, Enrico Schuerrer <EraseMEenricospamEraseMEgmx.at> wrote:
>
> -----Ursprüngliche Nachricht-----
> Von: "Michael Rigby-Jones" <@spam@Michael.Rigby-Jones@spam@spamspam_OUTbookham.com>
> An: "Microcontroller discussion list - Public." <spamBeGonepiclistspamKILLspammit.edu>
> Gesendet: Montag, 06. Februar 2006 18:57
> Betreff: RE: [EE] Transient Suppression - was Re: [ee] 3904 base resistor
>
>
> [rest snipped]
>
> >
> I don't think that the problem is speed or current - the problem is the voltage. A 1N4148 is rated with 100 Volts and this could be less if you switch off a relay. From my experiences a 1N4007 as a general purpose diode with Vrrm = 1000V and much reserves in non repetitive current will do the job without problems and cost nearly nothing.
>
> Regards
>
> Enrico
>

But the flyback voltage generated by the relay coil turns the diode
on. The voltage rating of the diode only needs to match the power
supply requirement (the voltage accros the relay when on) - plus the
usual margin.


Richard P

2006\02\06@170243 by Wouter van Ooijen

face picon face
> I don't think that the problem is speed or current - the
> problem is the voltage. A 1N4148 is rated with 100 Volts and
> this could be less if you switch off a relay.

How can you even approach 100V when switching a low-voltage relais? When
the relais is energised there is just the power supply voltage (5, 12,
24V), when the transistor switches of the diode will conduct so the
voltage is just 0.6V.

Wouter van Ooijen

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


2006\02\06@170619 by Enrico Schuerrer

picon face

-----Ursprüngliche Nachricht-----
Von: "Richard Prosser" <.....rhprosserspam_OUTspamgmail.com>
An: "Microcontroller discussion list - Public." <TakeThisOuTpiclist.....spamTakeThisOuTmit.edu>
Gesendet: Montag, 06. Februar 2006 20:21
Betreff: Re: [EE] Transient Suppression - was Re: [ee] 3904 base resistor



> But the flyback voltage generated by the relay coil turns the diode
> on. The voltage rating of the diode only needs to match the power
> supply requirement (the voltage accros the relay when on) - plus the
> usual margin.

And why  1N4148s are dying, 1N4007 not?

Enrico

2006\02\06@171750 by David VanHorn

picon face
>
> And why  1N4148s are dying, 1N4007 not?


Interesting question. Obviously the 4007 can handle more energy, but that's
not likely the cause.

I'd wonder about the wiring between the transistor and relay, maybe another
sort of transient happening at much higher speed..

This is the sort of thing that gets interesting :)

2006\02\06@181728 by Brent Brown

picon face
> > But the flyback voltage generated by the relay coil turns the diode
> > on. The voltage rating of the diode only needs to match the power
> > supply requirement (the voltage accros the relay when on) - plus the
> > usual margin.
>
> And why  1N4148s are dying, 1N4007 not?

Not too many scenarios here...

Firstly, check if the 1N4148 has a current rating equal or greater than the
normal relay coil current (as other's have correctly said, this is the current the
diode sees at the instant when the relay is turned off, and it decays to zero
from there, there are no peaks above this).

Secondly, if the power supply is ever reversed this diode will be forward
biased and it or the switching transistor will die! Could be you have a fairly
rugged transistor and the 1N4148 is the weak link.

My take on this subject... inverse parallel diode is generally a more than OK
way of subduing back EMF of a relay when switched by a transistor. Keeps
EMF down out 0.6V.

TVS or similar device.... allows EMF to rise to a higher voltage (must be
designed for) in order to allow current to decay as fast as possible.

Good question...how does this effect EMI? If for example a 12V TVS was
used with a 12V supply... At turn off the coil voltage voltage would go
relatively quickly from +12V to -12V, hold there for a short while, then decay
to 0. The fast edge (24V spike) is what generates the EMI.

Compare this with a simple inverse parallel diode...At turn off the coil voltage
voltage would go relatively quickly from +12V to -0.6V, hold there for a little
longer than the above example, then decay to 0. Hmmm, a 12.6V spike
compared with a 24V spike.

I suspect that "slow" diodes (eg 1N400x series) might even produce less
EMI.... the slowness being evident as an equivalent capacitance that
suppresses dv/dt. I also could be totally wrong here :-) An RC snubber would
be a better idea.

Interesting subject...simple, but quite a few things to think about.

--
Brent Brown, Electronic Design Solutions
16 English Street, Hamilton 2001, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell/txt: 027 433 4069
eMail:  TakeThisOuTbrent.brownKILLspamspamspamclear.net.nz


2006\02\06@191721 by Enrico Schuerrer

picon face

-----Ursprüngliche Nachricht-----
Von: "Wouter van Ooijen" <.....wouterspamRemoveMEvoti.nl>
An: "'Microcontroller discussion list - Public.'" <RemoveMEpiclistspamspamBeGonemit.edu>
Gesendet: Montag, 06. Februar 2006 23:04
Betreff: RE: [EE] Transient Suppression - was Re: [ee] 3904 base resistor


> > I don't think that the problem is speed or current - the
> > problem is the voltage. A 1N4148 is rated with 100 Volts and
> > this could be less if you switch off a relay.
>
> How can you even approach 100V when switching a low-voltage relais? When
> the relais is energised there is just the power supply voltage (5, 12,
> 24V), when the transistor switches of the diode will conduct so the
> voltage is just 0.6V.
>
Because it is dynamic, not static. Due to the capacity is charging in the pn-zone. And when the transistor shuts off the current through the relay a transient spike will be the result. During this time the voltage across the diode will go much higher than the supply voltage.

Regards

Enrico

2006\02\06@192231 by Brent Brown

picon face
{Quote hidden}

Doh! My bad. I just drew out the circuit and discovered my mistake. When a
TVS or zener is used it goes ACROSS the transistor, NOT accross the relay
coil.

Now I see that it is superior a simple diode across the relay coil. When
the transistor is turned off the TVS conducts to maintain current flow
from the coil (important to note that this current path now includes the
power supply, whereas with the simple diode across the coil circuit the
current flowed in a loop).

If the TVS clamping voltage is close to the power supply voltage the EMF
across the coil will be minimized. This does rely on the TVS switching
very quickly, but as someone else said this is something they are very
good at.
--
Brent Brown, Electronic Design Solutions
16 English Street, Hamilton 2001, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell/txt: 027 433 4069
eMail:  spamBeGonebrent.brown@spam@spamspam_OUTclear.net.nz


2006\02\06@193919 by Gerhard Fiedler

picon face
David VanHorn wrote:

>> And why  1N4148s are dying, 1N4007 not?
>
> Interesting question. Obviously the 4007 can handle more energy, but that's
> not likely the cause.

Could it be that the voltage rises too fast? That it reaches values deadly
for the 1N4148 before the diode can open?

Gerhard

2006\02\06@194958 by David VanHorn

picon face
>
> Could it be that the voltage rises too fast? That it reaches values deadly
> for the 1N4148 before the diode can open?


Not in that direction, that's the forward bias direction.

2006\02\06@195139 by David VanHorn

picon face
>
>
> I suspect that "slow" diodes (eg 1N400x series) might even produce less
> EMI.... the slowness being evident as an equivalent capacitance that
> suppresses dv/dt. I also could be totally wrong here :-) An RC snubber
> would
> be a better idea.


Reminds me of a shipment of russian relays we got in once.
They had series R/C supressors across the coils.
Molded in red plastic.

(red supressors on russian relays.. am I the ONLY guy that finds this
funny?)

2006\02\06@200041 by Rich Graziano

picon face
I have not followed this thread, but typically the switching speed of the
diode needs to be faster than the transient.  Fast schottky diodes are
typically used.  There are special diodes that are designed for the purpose.
Rectifier diodes like the 1N4007 will not switch fast enough to be effective
as a transient sink for fast transients.  The 1N4148 (also 1N914) are faster
and will turn on with a fast transient because they have a faster switching
speed. The 1N4148 is trying to do the job but apparently it cannot handle
the transient current.  The 1N4007 is probably not even responding until the
transient cycle has ended or is near the end. An O-scope is not an ideal
device for transient analysis.  A spectrum analyzer is the more appropriate
tool.  But not everyone has one., so an O-scope may be all you can use. You
might pick out a schottky diode to swamp the relay coil.  Also, it depends
on how you make the field rise.  The faster it rises, the higher wider the
spectrum of the transient.  If you turn the rely on hard, you can generate
an induction kick of 1 or greater KV.  If you turn it on slower you can
minimize the transient.  A lot ( actually everything) depends on your
application.  The application spec is the starting place.  Look at your
timing diagrams and see what you need to do.  I hope this helps a little.
Sometimes a small capacitor across the diode will help.  But remember that
the relay will not turn on or off until the capacitor is charged or
discharged. Your application spec will tell you if this is an option.
Rich :-)

{Original Message removed}

2006\02\06@201320 by Bob Axtell

face picon face
Brent Brown wrote:

{Quote hidden}

Yes, TVS switching times are around 0.1 to 1 pS. The 1N4148 switching
time is 4nS.
I _THINK_ that is 4000 x faster.

--Bob

--
Note: To protect our network,
attachments must be sent to
TakeThisOuTattachspamspamengineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

2006\02\06@202928 by Bob Axtell

face picon face
Gerhard Fiedler wrote:

>David VanHorn wrote:
>
>  
>
>>>And why  1N4148s are dying, 1N4007 not?
>>>      
>>>
>>Interesting question. Obviously the 4007 can handle more energy, but that's
>>not likely the cause.
>>    
>>
>
>Could it be that the voltage rises too fast? That it reaches values deadly
>for the 1N4148 before the diode can open?
>
>Gerhard
>
>  
>
My conclusion was that with a max reverse voltage rating of only 100V,
the occasional spike
that exceeded 100V with the relay's collapsing field (these are 12V
relays), caused a punch-through
of the junction, causing the short, as we found that all failures were
shorted 1N4148's. Since the
1N4007 can handle a 1000V spike, no damage can occur. But the 1N4007 is
much slower than the
1N4148, and we needed to block the spike more, so the TVS is used.

--Bob

--
Note: To protect our network,
attachments must be sent to
attachEraseMEspamengineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

2006\02\06@221112 by Gerhard Fiedler

picon face
David VanHorn wrote:

> (red supressors on russian relays.. am I the ONLY guy that finds this
> funny?)

Hm... I have that white ceramic resistor lying here, made in USA. Any
laughs? :)

Gerhard

2006\02\06@224248 by Rich Graziano

picon face
Yes, I agree. Perhaps an RC snubber would be good.  It all depends on the
timing requirements of the application. If the 1N4007, with its junction
capacity (that actually increases as it becomes more reversed biased) does
not switch (is held off) until the transient cycle has ended, it will do
nothing to capture the transient. In this application, a 12 volt relay can
generate a very large spike. It can also destroy the switching transistor if
the cycle hits right. The snubber would slow the rise time dPhi/dt.
Rich :-)


{Original Message removed}

2006\02\06@225548 by David VanHorn

picon face
On 2/6/06, Gerhard Fiedler <RemoveMElistsEraseMEspamspam_OUTconnectionbrazil.com> wrote:
>
> David VanHorn wrote:
>
> > (red supressors on russian relays.. am I the ONLY guy that finds this
> > funny?)
>
> Hm... I have that white ceramic resistor lying here, made in USA. Any
> laughs? :)


Actually.. yes. :)

2006\02\07@060457 by Alan B. Pearce

face picon face
>> identified the 1N4148 across a telecom relay as the primary problem.
>
>I always thought signal diodes were a little fragile in this
>application, and tend toward 1n400x rectifiers instead.

That is my gut feeling too. Having used diodes in exactly this manner for a
long time, I suspect that this failure of 1n4148's is more related to some
detail about the relay.

>But based on your first message, if speed is an issue, they
>might be worse.  But if the diodes blew in that design, that's
>not a diode speed issue, is it?

I suspect it may be more to do with the magnitude of the current flow
through the diode.

>And are the protection diodes in something like a ULN2003
>that much faster than discreet diodes, or are they just
>useless marketing ploys?

These may have started off as a parasitic diode in the chip, but someone saw
the possibility of having them included as catch diodes, so built in
appropriate ones.

>Sigh.  I though I knew what I was doing, and now I'm all confused!

<VBG> but isn't that normal for us lot ... ???

2006\02\07@114541 by Herbert Graf

flavicon
face
On Mon, 2006-02-06 at 23:04 +0100, Wouter van Ooijen wrote:
> > I don't think that the problem is speed or current - the
> > problem is the voltage. A 1N4148 is rated with 100 Volts and
> > this could be less if you switch off a relay.
>
> How can you even approach 100V when switching a low-voltage relais? When
> the relais is energised there is just the power supply voltage (5, 12,
> 24V), when the transistor switches of the diode will conduct so the
> voltage is just 0.6V.

I'm no expert, but remember we are dealing with non ideal components
here.

The diode isn't an ideal device. When reversed biased the charge
separation is basically a capacitor (very useful in FM transmitters).
When the relay is turned off it'll take a while for this "capacitor" to
discharge and the diode to become forward biased and start conducting.

The relay, being an inductor, will "want" to keep the same amount of
current flowing and will do anything it can to make that happen. It'll
ramp it's voltage up so that the current continues to flow. It's during
this time of going from reverse biased to forward biased that the
voltage "spikes"; and you can EASILY see voltages WAY above the power
supply voltage.

Considering how widespread and successful the practise of putting a
reverse diode across a relay coil is, calling it "bad practise" is a
false generalization. I will agree that IN SOME CASES it is prudent to
go for a "better" device, but if the problem were that bad we would
notice.

TTYL

-----------------------------
Herbert's PIC Stuff:
http://repatch.dyndns.org:8383/pic_stuff/

2006\02\07@134539 by Richard Prosser

picon face
One other possibility - Is the feed to the relay decoupled? If there
is significant inductance in this line & it's not decoupled properly
then there will be an additional transient from this inductance. While
I can't quite see how it would damage a diode in parallel with the
relay, it could damage the switching device.

Also - is there anything special about the relay. Relay coil
inductance does change as it operates & there could be "something" to
do with the relay that produces a waveform different from what we're
talking about. For example a polarised relay will produce different
waveform from a non-polarised one.

And it is DC energised isn't it?

RP


On 08/02/06, Herbert Graf <@spam@mailinglist2RemoveMEspamEraseMEfarcite.net> wrote:
{Quote hidden}

> -

2006\02\07@163401 by Wouter van Ooijen

face picon face
> The relay, being an inductor, will "want" to keep the same amount of
> current flowing and will do anything it can to make that happen. It'll
> ramp it's voltage up so that the current continues to flow.
> It's during
> this time of going from reverse biased to forward biased that the
> voltage "spikes"; and you can EASILY see voltages WAY above the power
> supply voltage.

Yes, but at that moment the diode is either conducting or on the eve of
conducting. Maybe a very fast switch-off could produce a high-enough
spike to harm the diode. Sideline: does a diode spec mention a maximum
*forward* voltage?

I very much doubt the standard saturated biplar transistor will switch
off faster than the reaction time of an 1N4148.

Wouter van Ooijen

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


2006\02\07@184056 by Gerhard Fiedler

picon face
Wouter van Ooijen wrote:

> Sideline: does a diode spec mention a maximum *forward* voltage?

No, but a maximum forward current. As long as you stay below that, the
forward voltage is in normal ranges.

Gerhard

2006\02\07@193048 by Spehro Pefhany

picon face
At 09:40 PM 2/7/2006 -0200, you wrote:
>Wouter van Ooijen wrote:
>
> > Sideline: does a diode spec mention a maximum *forward* voltage?
>
>No, but a maximum forward current. As long as you stay below that, the
>forward voltage is in normal ranges.
>
>Gerhard

They always spec the maximum forward voltage at one or more currents, and
often give the typical voltage at a variety of currents. You can infer what
the maximum forward voltage should be, given the current (and junction
temperature).

For example, the 1N400x series has a maximum Vf of 1.1V at 1A/Tj = 25°C (0.93V
typical), and typically 1.4V @20A Tj = 25°C.

There is a current (around 10A, IIRC) where the tempco of forward voltage
is actually zero, but obviously it is of rather limited utility.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
EraseMEspeffspam@spam@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2006\02\07@195520 by Brent Brown

picon face
> > > Sideline: does a diode spec mention a maximum *forward* voltage?
> >
> >No, but a maximum forward current. As long as you stay below that, the
> >forward voltage is in normal ranges.

The more interesting question here is though..."if the diode has a finite turn
on time, is there a maximum voltage that may be applied during that period
that will not destroy the device."

(P.S. I don't know the answer)

--
Brent Brown, Electronic Design Solutions
16 English Street, Hamilton 2001, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell/txt: 027 433 4069
eMail:  @spam@brent.brownspam_OUTspam.....clear.net.nz


2006\02\07@202034 by Cristóvão Dalla Costa

picon face
On 2/7/06, Brent Brown <spamBeGonebrent.brownEraseMEspamclear.net.nz> wrote:
>
> > > > Sideline: does a diode spec mention a maximum *forward* voltage?
> > >
> > >No, but a maximum forward current. As long as you stay below that, the
> > >forward voltage is in normal ranges.
>
> The more interesting question here is though..."if the diode has a finite
> turn
> on time, is there a maximum voltage that may be applied during that period
> that will not destroy the device."
>
> (P.S. I don't know the answer)
>

The diode's turn-on time should be dependant on the forward voltage being
applied and the issue you point out would be unimportant, I think, specially
on the case of small power inductive discharges where the diodes capacitance
will absorb a (significant) part of the discharge.

2006\02\08@021205 by Bob Axtell

face picon face
Richard Prosser wrote:

>One other possibility - Is the feed to the relay decoupled? If there
>is significant inductance in this line & it's not decoupled properly
>then there will be an additional transient from this inductance. While
>I can't quite see how it would damage a diode in parallel with the
>relay, it could damage the switching device.
>
>Also - is there anything special about the relay. Relay coil
>inductance does change as it operates & there could be "something" to
>do with the relay that produces a waveform different from what we're
>talking about. For example a polarised relay will produce different
>waveform from a non-polarised one.
>
>  
>
I know. No, it was a standard 12V DC Modem relay, like you see all over
the place.
12V is applied while it is closed.

Good idea, though.

I didn't design it,  I was just hired to  find out what  was going
wrong. The failure <1%
of the installed base, and didn't fail immediately after insertion into
service. But for
these guys, that is a VERY high number. You have been to a K-Mart,
right? These
relays actuate every time the call center is dialled up, about 500 +x
per day.

What worried us the most was NOT that the diode failed & took out the
12V supply,
it was that the diode might OPEN and the RF might cause problems with
the uP,thus
flawing the transaction document being transferred.
.
The TVS solved both problems.

--Bob

{Quote hidden}

>>-

2006\02\08@034427 by Wouter van Ooijen
face picon face
> > Sideline: does a diode spec mention a maximum *forward* voltage?
>
> No, but a maximum forward current. As long as you stay below that, the
> forward voltage is in normal ranges.

In a static situation yes, but the issue here was that the forward
voltage might rise so fast that the diode does not have time to turn on
so you'd get a large forward voltage which might destroy the diode. I
doubt the reality of this scenario, but it is what (this part of) the
discussion was about.

Wouter van Ooijen

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


2006\02\08@060429 by Gerhard Fiedler

picon face
Wouter van Ooijen wrote:

>>> Sideline: does a diode spec mention a maximum *forward* voltage?
>>
>> No, but a maximum forward current. As long as you stay below that, the
>> forward voltage is in normal ranges.
>
> In a static situation yes, but the issue here was that the forward
> voltage might rise so fast that the diode does not have time to turn on
> so you'd get a large forward voltage which might destroy the diode. I
> doubt the reality of this scenario, but it is what (this part of) the
> discussion was about.

I understand. But I don't think this is usually specified.

As Spehro says, they often specify the forward voltage at a number of
(static) forward currents. The only dynamic characteristic that I think is
commonly specified is capacitance in reverse and reverse recovery.

But they actually specify the maximum forward voltage (at a certain
current) in maximum ratings sections. We probably could conclude from this
that going over this in a dynamic situation takes the diode out of the
specified operating range.

Gerhard

2006\02\08@062805 by Wouter van Ooijen

face picon face
> But they actually specify the maximum forward voltage (at a certain
> current) in maximum ratings sections.

I checked http://www.fairchildsemi.com/ds/1N/1N4148.pdf, it does not
specify this. Only the maximum current.

> We probably could conclude from this
> that going over this in a dynamic situation takes the diode out of the
> specified operating range.

The 'this' is missing, so no conclusion can be drawn...

I might be a sloppy reader, but I can't even find a switch-on time
specification, which would be the root issue here. So again, no data =>
no conclusions :(

Wouter van Ooijen

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


2006\02\08@081949 by Spehro Pefhany

picon face
At 09:46 AM 2/8/2006 +0100, you wrote:
> > > Sideline: does a diode spec mention a maximum *forward* voltage?
> >
> > No, but a maximum forward current. As long as you stay below that, the
> > forward voltage is in normal ranges.
>
>In a static situation yes, but the issue here was that the forward
>voltage might rise so fast that the diode does not have time to turn on
>so you'd get a large forward voltage which might destroy the diode. I
>doubt the reality of this scenario, but it is what (this part of) the
>discussion was about.

The forward recovery time of a slow-as-molasses diode such as the 1N4005 is
a couple hundred ns at 100mA typically. So, the voltage might overshoot a
a volt or less with a typical edge from an inductor. No big deal.
1N4006,7 use a different die and are about 50% slower. Reverse recovery of
these slugs is in the microseconds.

On the 1N4148, with 50mA switched, 20ns rise time, you'll see less than 2.5V
forward peak forward voltage (maximum guaranteed).

None of this is at all significant unless things are right at the edge.

>Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
RemoveMEspeff@spam@spamspamBeGoneinterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2006\02\08@090333 by Enrico Schuerrer

picon face
-----Ursprüngliche Nachricht-----
Von: "Spehro Pefhany" <.....speff@spam@spamEraseMEinterlog.com>
An: "Microcontroller discussion list - Public." <.....piclistRemoveMEspammit.edu>
Gesendet: Mittwoch, 08. Februar 2006 14:31
Betreff: RE: [EE] Transient Suppression - was Re: [ee] 3904 base resistor


{Quote hidden}

Yesterday I tested a simple relay switch first with a NPN switching transistor (2N2222) and then with a MOSFet. The relay was a 2 contact 12V/200 Ohms printed board relay made by Schrack and  I tested different diodes.

Conclusion for me: the quick FET produces the biggest spike, independent from the diode. The voltage runs for nearly 1 µs up to 160 Volts, measured with an good old analog Tektronix 300 MHz scope. The best diode for suppression was a BYV2100, a Schottky soft recovery diode (a little bit expensive as a diode for a relay), the second best a 1N4148. At this diode the spike was nearly 80 Volts high, so with an other relay maybe the diode will come in trouble... (80V / 200 Ohm = 0,4 A !).

The lowest spikes I measured with the NPN-transistor switching the relay, especially when I put 330pF between base and ground (22 Volts spike).

Regards
Enrico




2006\02\08@092402 by David VanHorn

picon face
>
> The lowest spikes I measured with the NPN-transistor switching the relay,
> especially when I put 330pF between base and ground (22 Volts spike).


Slowing down the turn-off will do that, at the expense of heating the
transistor some while it's in the linear region.

I had a similar problem on an SMPS once, 60V transistor was showing 58V
spikes, but the factory told me the transistors were tested to 70V.. Turned
out that when I put a 1000V mosfet in there, the spikes were about 200V!
The original transistor was getting into avalanche far faster than the scope
could see.  Lesson learned, make a check with a very high voltage part!

Solved that one by combination of light snubber network and slightly slowing
the turn-off time.

2006\02\08@104223 by Wouter van Ooijen

face picon face
> Conclusion for me: the quick FET produces the biggest spike,
> independent from the diode. The voltage runs for nearly 1 µs
> up to 160 Volts, measured with an good old analog Tektronix
> 300 MHz scope. The best diode for suppression was a BYV2100,
> a Schottky soft recovery diode (a little bit expensive as a
> diode for a relay), the second best a 1N4148. At this diode
> the spike was nearly 80 Volts high, so with an other relay
> maybe the diode will come in trouble... (80V / 200 Ohm = 0,4 A !).

I think you should not calculate like that: the voltage can rise only
because there is no current (the inductance tries to maintain the
current, and rises the voltage as needed)! And the 1N4148 I saw was
actually rated for 1A for short pulses.

Wouter van Ooijen

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


2006\02\08@132211 by Bob Axtell

face picon face
Wouter van Ooijen wrote:

{Quote hidden}

The guaranteed switching time for the 1N4148 is 4 nS. The switching time
for a unidirectional TVS
is less than 1 pS; I calculate that to be 4000 times faster.

--Bob


{Quote hidden}

--
Note: To protect our network,
attachments must be sent to
.....attachSTOPspamspam@spam@engineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

2006\02\08@134124 by David VanHorn

picon face
>
>
> The guaranteed switching time for the 1N4148 is 4 nS. The switching time
> for a unidirectional TVS
> is less than 1 pS; I calculate that to be 4000 times faster.


But if your transistor can be made to switch off in a time of about a
microsecond, neither of these will matter.

2006\02\08@134458 by Michael Rigby-Jones

picon face


>-----Original Message-----
>From: piclist-bouncesEraseMEspam@spam@mit.edu [RemoveMEpiclist-bouncesspamspamBeGonemit.edu]
>Sent: 08 February 2006 18:22
>To: Microcontroller discussion list - Public.
>Subject: Re: [EE] Transient Suppression - was Re: [ee] 3904
>base resistor
>
>
>>
>The guaranteed switching time for the 1N4148 is 4 nS. The
>switching time
>for a unidirectional TVS
>is less than 1 pS; I calculate that to be 4000 times faster.

What would be a typical rise time of a saturated bipolar relay driver?

Regards

Mike

=======================================================================
This e-mail is intended for the person it is addressed to only. The
information contained in it may be confidential and/or protected by
law. If you are not the intended recipient of this message, you must
not make any use of this information, or copy or show it to any
person. Please contact us immediately to tell us that you have
received this e-mail, and return the original to us. Any use,
forwarding, printing or copying of this message is strictly prohibited.
No part of this message can be considered a request for goods or
services.
=======================================================================

2006\02\08@145504 by Peter

picon face


On Wed, 8 Feb 2006, Wouter van Ooijen wrote:

{Quote hidden}

Ton should be about 5 nsec for 1N4148.

Peter

2006\02\08@152621 by Spehro Pefhany

picon face
At 09:55 PM 2/8/2006 +0200, you wrote:


>On Wed, 8 Feb 2006, Wouter van Ooijen wrote:
>
> >> But they actually specify the maximum forward voltage (at a certain
> >> current) in maximum ratings sections.
> >
> > I checked http://www.fairchildsemi.com/ds/1N/1N4148.pdf, it does not
> > specify this. Only the maximum current.
> >
> >> We probably could conclude from this
> >> that going over this in a dynamic situation takes the diode out of the
> >> specified operating range.
> >
> > The 'this' is missing, so no conclusion can be drawn...
> >
> > I might be a sloppy reader, but I can't even find a switch-on time
> > specification, which would be the root issue here. So again, no data =>
> > no conclusions :(
>
>Ton should be about 5 nsec for 1N4148.

So, at 150mV/ns dv/dt at the 2N3904 collector, we've got an extra
overshoot of 0.75V.

>Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
spamBeGonespeffKILLspamspam@spam@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2006\02\08@160422 by Wouter van Ooijen

face picon face
> > I might be a sloppy reader, but I can't even find a switch-on time
> > specification, which would be the root issue here. So
> again, no data =>
> > no conclusions :(
>
> Ton should be about 5 nsec for 1N4148.

'should' is not a valid input for design. Can you locate that parameter
in the (any?) datasheet?

Wouter van Ooijen

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


2006\02\08@164951 by Bob Axtell

face picon face

>>Ton should be about 5 nsec for 1N4148.
>>    
>>
>
>'should' is not a valid input for design. Can you locate that parameter
>in the (any?) datasheet?
>
>Wouter van Ooijen
>
>
>  
>
http://www.vishay.com/docs/85521/85521.pdf  

Shows time to "reverse its polarity", same thing as switching speed.

--Bob

--
Note: To protect our network,
attachments must be sent to
attachspam_OUTspam@spam@engineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

2006\02\08@170055 by David VanHorn

picon face
>
>
> 'should' is not a valid input for design. Can you locate that parameter
> in the (any?) datasheet?
>
> Wouter van Ooijen


Neither is TBD, but way too many devices these days go through their whole
lives with a bunch of parameters in this mode.

2006\02\08@170148 by Spehro Pefhany

picon face
At 02:49 PM 2/8/2006 -0700, you wrote:

> >>Ton should be about 5 nsec for 1N4148.
> >>
> >>
> >
> >'should' is not a valid input for design. Can you locate that parameter
> >in the (any?) datasheet?
> >
> >Wouter van Ooijen
> >
> >
> >
> >
>http://www.vishay.com/docs/85521/85521.pdf
>
>Shows time to "reverse its polarity", same thing as switching speed.
>
>--Bob

trr is reverse recovery time (time to switch "off").

This Philips datasheet shows the forward recovery characteristic in terms of
overshoot voltage Vfr:

http://ronja.twibright.com/datasheets/1N4148_1N4448_5.pdf


>Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
spamBeGonespeff@spam@spaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2006\02\08@174354 by Gerhard Fiedler

picon face
Wouter van Ooijen wrote:

>> But they actually specify the maximum forward voltage (at a certain
>> current) in maximum ratings sections.
>
> I checked http://www.fairchildsemi.com/ds/1N/1N4148.pdf, it does not
> specify this. Only the maximum current.

I meant this to mean "in some maximum ratings sections", like for example
for the Diodes Inc S1A family. (However, this is a section "Maximum Ratings
and Electrical Characteristics", which doesn't make it clear what kind of
spec this is.)

>> We probably could conclude from this that going over this in a dynamic
>> situation takes the diode out of the specified operating range.
>
> The 'this' is missing, so no conclusion can be drawn...

The 'this' was meant to be the fact that some datasheets seem to specify a
maximum forward voltage. But maybe not... the one I found is not really
clear about what kind of spec this is.

> I might be a sloppy reader, but I can't even find a switch-on time
> specification, which would be the root issue here. So again, no data =>
> no conclusions :(

I think what you mean is called "Reverse Recovery Time" and is 4 ns at 10
mA forward current in this datasheet.

Gerhard

2006\02\08@194820 by Jinx

face picon face
So where are we up to with this ?

Options so far -

TVS across transistor
RC snubber across relay coil
RC on transistor base to slow switch-off
Diode (of some sort) across relay coil

Is there any consensus as to the preferred order ?

eg if cost is not an issue / if cost is an issue
eg if components are not to be grossly over-rated

2006\02\08@204410 by Brent Brown

picon face
> So where are we up to with this ?
>
> Options so far -
>
> TVS across transistor
> RC snubber across relay coil
> RC on transistor base to slow switch-off
> Diode (of some sort) across relay coil
>
> Is there any consensus as to the preferred order ?
>
> eg if cost is not an issue / if cost is an issue
> eg if components are not to be grossly over-rated

Probably not a consensus but my summary in terms of price and
performance would be...

1. Diode across relay coil (cheapest option) - Time honoured method. Must
be suitably current rated ( = coil current), voltage spikes don't exist here,
diode turn off time not an issue either, diode turn on time has been an
interesting discussion but I am yet to be convinced that a lazy-as 1N4004 etc
wouldn't be perfectly acceptable here.

2. TVS across transistor (pricier) - great for lowest EMI option, better than
above, but needs a little more design attention to select best TVS voltage for
lowest EMI. Potentially problematic if PSU voltage is unregulated or spikey
(will conduct and pass current through relay coil when voltage rises above
threshold).

3. RC on transistor base - Even just increacing base resistor value can slow
down turn off time appreciably for no extra component cost, slower EMF
spike => lower EMI.  Works with BJT and MOSFET. Add C where
necessary. There will be dynamic power dissipation in the transistor as it
turns off, need to check SOAR carefully.

4. RC snubber across relay coil. Adds the finishing touch to reducing EMI.
Low cost anyway. Add this in addition to any of the above. I rank this last
because it's two extra components which eat PCB real estate and needs to
be used in addition to some kind of diode.

--
Brent Brown, Electronic Design Solutions
16 English Street, Hamilton 2001, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell/txt: 027 433 4069
eMail:  RemoveMEbrent.brownEraseMEspamKILLspamclear.net.nz


2006\02\08@210324 by David VanHorn

picon face
> 3. RC on transistor base - Even just increacing base resistor value can
> slow
> down turn off time appreciably for no extra component cost, slower EMF
> spike => lower EMI.  Works with BJT and MOSFET. Add C where
> necessary. There will be dynamic power dissipation in the transistor as it
> turns off, need to check SOAR carefully.


Depending on implementation, this can be the lowest cost and lowest EMI.
If you can get there just adjusting the resistor value, then it's a zero
cost option.
If you don't make an emi spike, then you don't need to supress it.
All the other methods are absorbing a spike already made.

2006\02\08@212334 by Richard Prosser

picon face
> Options so far -
>
>  TVS across transistor
>  RC snubber across relay coil
>  RC on transistor base to slow switch-off
>  Diode (of some sort) across relay coil
>
> Is there any consensus as to the preferred order ?
>
>  eg if cost is not an issue / if cost is an issue
>  eg if components are not to be grossly over-rated
>
...snip

My 2c worth -

Note that anything across the switching device (TVS / snubber etc)
will include the power supply and decoupling cap in the current loop.
This is likely to induce more EMI in adjacent circuits than the
shorter path when just a diode is used.

In general I'd use a diode to take care of the relay current and
possibly add a snubber to sort out feed inductance effects. And
decouple as close to the hot side of the coil as possible.

If switching speed is important I'd look at using a TVS across the
relay coil (in place of the diode)  to kill the loop current as soon
as possible but would take a bit of care selecting the correct voltage
ratings etc. I've done a similar circuit with a switched, zener
controllered bipolar to kill the coil current quickly but that was to
meet a special speed requirement.

Another possibility is to connect a zener from the collector to base
of the switching device. A high voltage on the collector then turns
the switch on slightly and the power is dissipated in that - rather
than the zener (useful for higher power levels). You have to have a
some impedance in the base drive to avoid the current pulse feeding
back to the driver however. And a snubber to sort things out until the
zener turns on.
Adding an RC to the drive circuit will work but it can also oscillate
if you get very unlucky with stray inductance & capacitances and fast
devices.

Richard P

2006\02\08@214452 by Bob Axtell

face picon face
Jinx wrote:

>So where are we up to with this ?
>
>Options so far -
>
>  
>
My take is now:

>TVS across transistor
>  
>
if warranted for high reliability

>RC snubber across relay coil
>  
>
Actually this seems to work well.

>RC on transistor base to slow switch-off
>  
>
not familiar with this.

>Diode (of some sort) across relay coil
>  
>
Probably normal useage

>Is there any consensus as to the preferred order ?
>
>eg if cost is not an issue / if cost is an issue
>  
>
diode is cheapest solution

>eg if components are not to be grossly over-rated
>
>  
>
--Bob

--
Note: To protect our network,
attachments must be sent to
spamBeGoneattachspam_OUTspamRemoveMEengineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

2006\02\08@215745 by Spehro Pefhany

picon face
At 03:23 PM 2/9/2006 +1300, you wrote:


>If switching speed is important I'd look at using a TVS across the
>relay coil (in place of the diode)

You'd have to use a diode in series with the TVS or a bidirectional
TVS. In the latter case you'd have less flexibility on the voltage
rating, since it would have to be less than the maximum supply voltage
under all possible conditions, lest the transistor prematurely ring up
the curtain and joint the choir invisible.

Simplest (and noisiest) solution is to use an avalanche-rated MOSFET.

>Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
.....speffspamRemoveMEinterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2006\02\08@224452 by William Chops Westfield

face picon face
>> Options so far -
>>
>>  TVS across transistor
>>
Isn't a TVS across the relay coil also an option?

BillW

2006\02\09@015817 by Wouter van Ooijen

face picon face
> (However, this is a section "Maximum Ratings
> and Electrical Characteristics", which doesn't make it clear
> what kind of spec this is.)

Most people don't know what to do with a plain "Maximum Ratings"
section. Calling a section "Maximum Ratings and Electrical
Characteristics" must be an attempt to maximise confusion!

> I think what you mean is called "Reverse Recovery Time" and
> is 4 ns at 10 mA forward current in this datasheet.

No, I mean something like "forward switch-on time", which is the reverse
of that parameter.

Wouter van Ooijen

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


2006\02\09@015817 by Wouter van Ooijen

face picon face
> This Philips datasheet shows the forward recovery
> characteristic in terms of overshoot voltage Vfr:
> http://ronja.twibright.com/datasheets/1N4148_1N4448_5.pdf

Yes, that is something I can recognise as a switch-on parameter :)

Wouter van Ooijen

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


2006\02\09@015818 by Wouter van Ooijen

face picon face
> http://www.vishay.com/docs/85521/85521.pdf  
>
> Shows time to "reverse its polarity", same thing as switching speed.

where? I see only a 'reverse recovery time', which I understand is the
switch *off* time. Is that by definition the same as the switch on time?

Wouter van Ooijen

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


2006\02\09@040411 by Alan B. Pearce

face picon face
> http://www.vishay.com/docs/85521/85521.pdf  
>
>Shows time to "reverse its polarity", same thing as switching speed.

Not quite - it only shows switch off time the way it is spec'd.

2006\02\09@041526 by Alan B. Pearce

face picon face
>1. Diode across relay coil (cheapest option) - Time
>honoured method. Must be suitably current rated
>( = coil current), voltage spikes don't exist here,

I think the discussion has determined that the voltage spikes do exist due
to diode turn on time - and most probably because of other factors such as
poor layout of the connections to the diode, which must be as directly
across the coil as possible.

>diode turn off time not an issue either,

agreed

>diode turn on time has been an interesting discussion
>but I am yet to be convinced that a lazy-as 1N4004 etc
>wouldn't be perfectly acceptable here.

Well I suspect that this may depend more on the Vce rating of the
transistor. And while a 1n400x diode may be described as lazy, the junction
capacitance may well have as much to do with limiting the voltage peak at
transistor turn off as the diode eventually conducting.

>3. RC on transistor base - Even just increacing base
>resistor value can slow down turn off time appreciably
>for no extra component cost, slower EMF spike => lower EMI.
>Works with BJT and MOSFET. Add C where necessary.

I suspect that when driving a BJT from a PIC port, the trick would be to
have a resistor BE and tristate the port. The resistor value would need to
be chosen to limit Vcb voltage shoot through from destroying the PIC though.
Same trick may work for a MOSFET.

2006\02\09@065446 by Vasile Surducan

face picon face
I'm in the gain. Look at you old guys, 70 messges for a pour diode.
:))

On 2/7/06, Wouter van Ooijen <wouterspam@spam@voti.nl> wrote:
> > I don't think that the problem is speed or current

Oh boy, again voltage without current like on the MCLR pin on
Microchip datasheet.

What the hell is happening with more than 100V transient voltage
if the load is the coil impedance (which is also the generator) and
a poor opened (or trying to switch from off to on) diode ?

Sure,  there is no current flow in that circuit and 200mA on which
1N4148 is rated, is more than enough isn't it ?
Try to switch a 10A solenoid and see what's happening.

- the
> > problem is the voltage. A 1N4148 is rated with 100 Volts and
> > this could be less if you switch off a relay.
>
> How can you even approach 100V when switching a low-voltage relais? When
> the relais is energised there is just the power supply voltage (5, 12,
> 24V), when the transistor switches of the diode will conduct so the
> voltage is just 0.6V.

Wouter, it's the same problem like someone on the jallist has with a stepper
switch supress a time ago.

The switching time of the diode multiplied with the current flow
through diode (which is not at all neglijible, in fact is a huge spike
of current) and multiplied with the voltage drop across diode (which
is much greater than 0.6V because diode is not in direct conduction
due the switching time, it's equal with the energy requested to be
eaten by the suppressor device.

A 1N4148 could supress a small inductive load. As large is the current
through coil (and not the nominal coil voltage which could be even
12V) as large is the spike current.

As an example, back in the 80 I've measured using a volt-ampermetric
methode, the resistance of 1600KVA transformers using a 12V acid
battery (load current about 80A-100A).
The time requsted for coil charge was minutes. The flame produced when
disconect the transformer from the battery was about 50cm.
What diode could supress such energy ?An 1KA diode with 1000V
breakdown voltage. Even the measurement was done at 12V. Now I think
is clear.

best wishes,
Vasile

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

2006\02\09@072539 by Spehro Pefhany

picon face
At 01:54 PM 2/9/2006 +0200, you wrote:


>The switching time of the diode multiplied with the current flow
>through diode (which is not at all neglijible, in fact is a huge spike
>of current)

IT CANNOT >>>CANNOT<<<< EXCEED THE STEADY STATE COIL CURRENT!!!!

If you can switch it with a wimpy 2N3904 it's nothing to worry about.

>and multiplied with the voltage drop across diode (which
>is much greater than 0.6V because diode is not in direct conduction
>due the switching time, it's equal with the energy requested to be
>eaten by the suppressor device.

Hardly. First, the energy is not 'eaten by the suppressor device'
when you use a diode, MOST of it (95%) dissipates in the copper resistance
of the coil. After switching off a 12V relay, there is briefly (milliseconds)
a 12V + diode drop across the coil resistance internally (opposite polarity).
Secondly, the voltage at the collector of the BJT moves rather slowly. The BJT
switches off like molasses, and there is plenty of distributed
capacitance in the coil on top of that. There are NO terahertz or gigahertz
frequencies here where nanoseconds or picoseconds make a difference.
No magic, this is very straightforward low-frequency analog stuff.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
EraseMEspeffRemoveMEspamSTOPspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2006\02\09@131002 by Peter

picon face

On Wed, 8 Feb 2006, Wouter van Ooijen wrote:

>>> I might be a sloppy reader, but I can't even find a switch-on time
>>> specification, which would be the root issue here. So
>> again, no data =>
>>> no conclusions :(
>>
>> Ton should be about 5 nsec for 1N4148.
>
> 'should' is not a valid input for design. Can you locate that parameter
> in the (any?) datasheet?

Hmm, you seem to have a point there. I can only find the trr but I infer
that ton is also of the same order of magnitude. The best data sheet is
from Philips, here:

http://pubpages.unh.edu/~aperkins/pdf/1N-devices/1N4148.pdf

another, showing trr/current plot (National):

http://www.aleeanne.org.uk/datasheets/datasheet2/1N914.PDF

I inferred the data from the max. recommended operating frequency
(100MHz). I suspect that other problems come in between by the time you
consider the 4nsec value (like lead inductance). The trr measurements
are made with 10mm leads in FR4 board. I think that this is just
hair-splitting here, beyond a certain point.

Peter

2006\02\09@132818 by Rich Graziano

picon face
The TVS is fast (Picoseconds). It is the diode of choice for protecting the
transistor because it can shunt the transient before the transistor or FET
is damaged. The idea is to be faster than the transient.  A snubber across
the relay coil will only work if it is actually across the relay coil.  If
it is on the circuit board it will be relatively ineffective.Transients are
secular events that can produce thousands of volts from a 12 volt relay
coil.  Since they are secular, they are different every time so some can be
more destructive than others. The best bet is to put the snubber across the
coil and the TVS across the semiconductor.
Rich



{Original Message removed}

2006\02\09@133052 by Peter

picon face


On Thu, 9 Feb 2006, Wouter van Ooijen wrote:

>> http://www.vishay.com/docs/85521/85521.pdf
>>
>> Shows time to "reverse its polarity", same thing as switching speed.
>
> where? I see only a 'reverse recovery time', which I understand is the
> switch *off* time. Is that by definition the same as the switch on time?

You can find the actual switch on time in the 1N4148 spice model afaik.

Peter

2006\02\09@134923 by David VanHorn

picon face
On 2/9/06, Rich Graziano <RemoveMErgrazia1KILLspamspamTakeThisOuTrochester.rr.com> wrote:
>
> The TVS is fast (Picoseconds). It is the diode of choice for protecting
> the
> transistor because it can shunt the transient before the transistor or FET
> is damaged. The idea is to be faster than the transient.



I hate to keep yanking this chain, but I think the idea is not to generate
the transient in the first place.

First rule of noise supression, Don't make it!

2006\02\09@141610 by Spehro Pefhany

picon face
At 08:30 PM 2/9/2006 +0200, you wrote:


>On Thu, 9 Feb 2006, Wouter van Ooijen wrote:
>
> >> http://www.vishay.com/docs/85521/85521.pdf
> >>
> >> Shows time to "reverse its polarity", same thing as switching speed.
> >
> > where? I see only a 'reverse recovery time', which I understand is the
> > switch *off* time. Is that by definition the same as the switch on time?
>
>You can find the actual switch on time in the 1N4148 spice model afaik.
>
>Peter

No. Typically SPICE diode models turn on instantly when forward biased.

>Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
spamBeGonespeffspam@spam@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2006\02\09@143808 by David VanHorn

picon face
>
> No. Typically SPICE diode models turn on instantly when forward biased.


Score another one for simulation!  :)

Isn't this an interesting little problem?  Amazing how something that seems
this simple can be so tough to put a finger on.

2006\02\10@160735 by Peter

picon face

On Thu, 9 Feb 2006, David VanHorn wrote:

>> No. Typically SPICE diode models turn on instantly when forward biased.
>
> Score another one for simulation!  :)
>
> Isn't this an interesting little problem?  Amazing how something that seems
> this simple can be so tough to put a finger on.

It is not so simple because the diode is already sort of 'on' when made.
It's just that without an exernal voltage source it is *reverse* biased.
So to turn on it has to recover from that (0.6V) reverse bias situation.
So a ton rating would be trr for I=0.

But I am fairly sure that the inductance of the leads and such prevents
any picosecond speeds. Usually a maximum speed is given for devices. For
1N4148 this is 100MHz afaik. This implies both ton and toff not longer
than 5nsec.

Peter


2006\02\10@213705 by Rich Graziano

picon face
Does that mean that a 12 DC relay being switched cannot produce a spike of
more than one KV? Has it been shown that the diode was destroyed by causes
other than the safe operating parameters having been exceeded?

{Original Message removed}

2006\02\11@173948 by Rich Graziano

picon face
I apologize if I missed some basic part of the problem, but is the transient
the result of the fast rise or collapse of flux of the relay coil, which is
what generates the transient spike?  Or is the transient from something
other than the relay coil?  If the transient is not generated from switching
the relay coil, can you please explain where the transient is being
generated.  Would this circuit still have the transient spike and the diode
problem if there was no relay coil?  What is the diode for if there is no
relay coil?    I thought that the coil would appear as a dead short for a
very brief time before the flux density increased, and an inrush of current
would rapidly expand the field. But perhaps I am wrong. I am curious, now. I
would like to learn something.  :-)

Regards,
Rich


{Original Message removed}

2006\02\11@222138 by Bob Axtell

face picon face
Rich Graziano wrote:

>I apologize if I missed some basic part of the problem, but is the transient
>the result of the fast rise or collapse of flux of the relay coil, which is
>what generates the transient spike?  Or is the transient from something
>other than the relay coil?  If the transient is not generated from switching
>the relay coil, can you please explain where the transient is being
>generated.  Would this circuit still have the transient spike and the diode
>problem if there was no relay coil?  What is the diode for if there is no
>relay coil?    I thought that the coil would appear as a dead short for a
>very brief time before the flux density increased, and an inrush of current
>would rapidly expand the field. But perhaps I am wrong. I am curious, now. I
>would like to learn something.  :-)
>
>  
>
Er, well, the recommendation was made that a 1N4148 would suffice as
transient protection
to a 2n3904 transistor.

I challenged that- my experience was that while problem-solving on a
troubleshooting contract, I found a high-
volume customer using 1N4148 across a relay that were failing by
punch-through (shorting). This took out
the power supply fuse the next time our transistor tried to pull in the
relay. Our switching IC was much stouter
than a 2N3904 (I think it was a SN75150? or something like that)..

I studied the literature, and replaced it with a unipolar TVS with a
voltage rating just over 12V (15 I think).
I assumed at the time that TVS's were now the accepted present state of
the art.

We speculated that transient spiking above the 100V max reverse voltage
rating must have damaged
the junction, but there was no formal inspection (there are companies
that examine faults like this for
a fee). It was not a lot failure as these diodes were from several
different manufacturers.

The relay used had a very high duty cycle, as it was used for on-hook
and pulse-dial..

This was a solid fix in my view, and no later failures with the TVS's
were ever passed to me.



On the PIClist, I seemed to have been corrected, as other folks still
use them in volume and never have
a failure. In fact, using a 2N3904 (a very slow switch) might prevent or
reduce transients because of
slow flux changes. I never use a 2N3904 for anything; I use 2N2222A or a
MOSFET instead.

Perhaps something else was occurring, but if so, installing the TVS
covered that over as well... that
happens sometimes in electronics.

I still believe the fix was the right thing to do... the failure rate
was too high  for my client to tolerate
under his circumstances. In my own designs, I have used TVS, Varistors,
etc all over, and I've never
had any failures whatever. I do not use diodes as protective devices,
mostly because  TVS has  a
much faster clamping speed (< 1 pS vs 4nS or slower for diodes)  which
helps EMC compliance.

Then other folks offered a few other solutions, like RC snubbers, etc

That's the story.

--Bob

--
Note: To protect our network,
attachments must be sent to
RemoveMEattachspam_OUTspamengineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

2006\02\12@062014 by Gerhard Fiedler

picon face
Bob Axtell wrote:

> Rich Graziano wrote:
>
>> I apologize if I missed some basic part of the problem, but is the
>> transient the result of the fast rise or collapse of flux of the relay
>> coil, which is what generates the transient spike?  [...]

> Er, well, the recommendation was made that a 1N4148 would suffice as
> transient protection to a 2n3904 transistor.
>
> I challenged that- my experience was that while problem-solving on a
> troubleshooting contract, I found a high- volume customer using 1N4148
> across a relay that were failing by punch-through (shorting). [...]

> That's the story.

And the rest of the story was trying to figure out why exactly those 1N4148
fail or failed; IOW what exactly happens in that brief time between
something switching off the relay current and the diode conducting
normally. It seems that most of the explanations for this are based on
assumptions and nobody has really hard knowledge or could dig up relevant
semiconductor specs.

Gerhard

2006\02\13@055650 by Alan B. Pearce

face picon face
>The relay used had a very high duty cycle, as it
>was used for on-hook and pulse-dial..

I suspect this was the real reason for the failure, effectively the pulse
dialling exceeded the repetitive surge current rating, causing the diode to
fail. The only ones that failed were probably close to the min spec limit,
with most devices having enough margin that they do not fail, although on
the basis of the failure rate, one would possibly need to look at long term
reliability.

>This was a solid fix in my view, and no later
>failures with the TVS's were ever passed to me.

And I do not think anyone can fault this fix.

I suspect that most of the discussion has resulted from TVS devices being
perceived (probably without foundation) as harder to find, and potentially
more expensive. In this instance I feel the extra expense (if any) could be
justified, and certainly the time taken to source the device is justified,
and would certainly justify changing new build to fit the TVS instead of a
diode.

2006\02\13@082020 by Bob Axtell

face picon face
Alan B. Pearce wrote:

>>The relay used had a very high duty cycle, as it
>>was used for on-hook and pulse-dial..
>>    
>>
>
>I suspect this was the real reason for the failure, effectively the pulse
>dialling exceeded the repetitive surge current rating, causing the diode to
>fail. The only ones that failed were probably close to the min spec limit,
>with most devices having enough margin that they do not fail, although on
>the basis of the failure rate, one would possibly need to look at long term
>reliability.
>  
>

agreed.

{Quote hidden}

thanks.

>I suspect that most of the discussion has resulted from TVS devices being
>perceived (probably without foundation) as harder to find, and potentially
>more expensive.
>
When I did this fix, TVS _were_ fairly new and more costly. Today, the
costs are low
(but not as low as a 1N4148 diode).

> In this instance I feel the extra expense (if any) could be
>justified, and certainly the time taken to source the device is justified,
>and would certainly justify changing new build to fit the TVS instead of a
>diode.
>  
>
Later layouts put the TVS in permanently.

--Bob

--
Note: To protect our network,
attachments must be sent to
attachspamspamengineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

More... (looser matching)
- Last day of these posts
- In 2006 , 2007 only
- Today
- New search...