Has anyone seen transorbs failing for no reason?

[MM]

Hi all,

I had recently two cases where transorbs burnt out for no obvious reason in
one of my designs. They are used on 12V rail to protect the circuit from
unlikely spikes. In both cases the boards were running in a system and would
suddenly smoke. The rest of the cards would continue to work fine. The power
supply checked after the fact was fine as well... The part number in
question is SMAJ12A made by I believe Diodes Inc.


Thanks,
/MM

 

[CBarn24050]

Subject: Has anyone seen transorbs failing for no reason?

The most likely explaination is that you have specified the wrong voltage
device for your application.

 

[MM]

"CBarn24050" <cbarn24050@aol.com> wrote in message
news:20050112141319.12040.00000080@mb-m15.aol.com...

Subject: Has anyone seen transorbs failing for no reason?

The most likely explaination is that you have specified the wrong voltage
device for your application.

What in your opinion would the right voltage to specify for a 12V rail if
the power supply spec is +/- 1%?

Thanks,
/MM

 

[CBarn24050]

Subject: Re: Has anyone seen transorbs failing for no reason?
From: "MM" mbmsv@yahoo.com
Date: 12/01/2005 19:30 GMT Standard Time
Message-id: <34lbtmF4bc5qmU1@individual.net

"CBarn24050" <cbarn24050@aol.com> wrote in message
news:20050112141319.12040.00000080@mb-m15.aol.com...
Subject: Has anyone seen transorbs failing for no reason?

The most likely explaination is that you have specified the wrong voltage
device for your application.

What in your opinion would the right voltage to specify for a 12V rail if
the power supply spec is +/- 1%?

Thanks,
/MM

There is no point using one on a regulated rail, you put them before the
regulator.

 

[MM]

"CBarn24050" <cbarn24050@aol.com> wrote in message
news:20050113113209.25300.00000103@mb-m17.aol.com...

There is no point using one on a regulated rail, you put them before the
regulator.

I agree, but this has been sort of design tradition at the place where I
work. I know what I should do in the future. The question is why the hell it
failed on the regulated rail?

Thanks,
/MM

 

[MM]

"Spehro Pefhany" <speffSNIP@interlogDOTyou.knowwhat> wrote in message
news:vtfdu0h6l4k7fl7eoh0tgvddic1j535cth@4ax.com...

Aside from capacitors, is there a source of energy that could exceed
the rating of the TVS? Any motors, solenoids, that sort of stuff?

It is basically a PC... So, the only motors are the fans and the drives. But
perhaps you are right and I should look at what they can do with the 12V
rail closer...


Thanks,
/MM

 

It is basically a PC

The 12V rail on most PC-clone supplies is very poorly regulated (it's
only regulated as a side-effect of the 5V regulation...).

The SMAJ12A you are using may trip as low as 13.3V, and I can easily
see a PC-clone supply putting that out.

Is it possible that you actually want a crowbar? Are you trying to
protect against power supply regulator failure?

Tim.

 

[MM]

<shoppa@trailing-edge.com> wrote in message
news:1105648026.811304.320970@f14g2000cwb.googlegroups.com...

It is basically a PC

The 12V rail on most PC-clone supplies is very poorly regulated (it's
only regulated as a side-effect of the 5V regulation...).

The SMAJ12A you are using may trip as low as 13.3V, and I can easily
see a PC-clone supply putting that out.

The power supply is not a cheap PC-clone. It is rather one of the best
available:
http://www.pcpowerandcooling.com/products/power_supplies/highperformance/turbocools/510-atx-pfc.htm


/MM

 

It is basically a PC
The 12V rail on most PC-clone supplies is very poorly regulated
The power supply is not a cheap PC-clone. It is rather one of the
best

Nothing against the PC Power and Cooling supplies; I like them. But
the 1% spec on the +12V is hopelessly optimistic under varying load
(especially small loads). I'm sure you're dumping more than 13.3V onto
the 12V bus occasionally. And with the 12V supply being rated at
30-some amps, I doubt that your Transzorb is dissipating the 350 watts
without burning to a crisp :-(.

Tim.

 

[MM]

"Clarence_A" <no@No.com> wrote in message
news:bvCFd.1520$8Z1.1125@newssvr14.news.prodigy.com...

The power supply is not a cheap PC-clone. It is rather one of
the best
available:


http://www.pcpowerandcooling.com/products/power_supplies/highperformance/turbocools/510-atx-pfc.htm
/MM

SO sorry, THAT supply IS a cheap PC CLONE! Made in China! Only
the +5V is regulated. The other voltages are ratioed to the +5V.

As far as I know even the most expensive non-ATX cPCI supplies use
cross-regulation. Show me a PC supply that doesn't? On the other hand where
did you find all this information, like where it is made and that it is
actually cross-regulated? There is no mentioning of China on their website,
so is it your educated guess or do you actually know this for a fact.
According to their spec +12V is regualted to 1%. As we are talking I have a
scope sitting on this 12V rail set to trigger on 12.5V. Half a day, still
waiting...

/MM

 

"MM" <mbmsv@yahoo.com> wrote:

shoppa@trailing-edge.com> wrote in message

The 12V rail on most PC-clone supplies is very poorly regulated (it's
only regulated as a side-effect of the 5V regulation...).

The power supply is not a cheap PC-clone. It is rather one of the best
available:
http://www.pcpowerandcooling.com/products/power_supplies/highperformance/turbocools/510-atx-pfc.htm

You're probably both right. "Best available" and "poor 12V
regulation" are not mutually exclusive...

 

SO sorry, THAT supply IS a cheap PC CLONE!

They're the best of the PC clones, and they claim "1%" regulation on
the 12V line. I think the original poster is reading too much into
that. I'm not sure how you do 1% regulation on a 30 amp supply without
remote sensing and/or busbars...

Look at a much better spec'ed 12V supply like a Lambda linear and they
never set the crowbar below 14.5V. I may be misreading the original
poster's intentions but a Transzorb isn't exactly a power supply
crowbar.

Tim.

 

[MM]

"Clarence_A" <no@No.com> wrote in message
news:23UFd.1750$8Z1.999@newssvr14.news.prodigy.com...

The Rep also warned about meeting the Minimum current for
regulation.

So, what is their minimum current spec? I couldn't find it in the
datasheet...

I have four of these and NOW they tell me it is not recommended
for the application I have.

If I can ask, what is your application? And why the supply is not
recommended for it?

I sized it for expansion, but to use
it I need a Resistive load in addition to the boards to assure
proper regulation. And I was only trying to answer some ones
question. Now I need to find a more appropriate supply!

Have you looked at Antec?



/MM

 

[Clarence_A]

"MM" <mbmsv@yahoo.com> wrote in message
news:34qm59F4e4kniU1@individual.net...

"Clarence_A" <no@No.com> wrote in message
news:23UFd.1750$8Z1.999@newssvr14.news.prodigy.com...

The Rep also warned about meeting the Minimum current for
regulation.

So, what is their minimum current spec? I couldn't find it in
the
datasheet...

I have four of these and NOW they tell me it is not
recommended
for the application I have.

If I can ask, what is your application? And why the supply is
not
recommended for it?

I sized for a eight card expansion, but we are shipping with only
one card in place. So the Current is too low to assure proper
regulation.

I didn't see it either. Probably should call to find out what it
is. The 1-800 number is on the web site!

I sized it for expansion, but to use
it I need a Resistive load in addition to the boards to assure
proper regulation. And I was only trying to answer some ones
question. Now I need to find a more appropriate supply!

Have you looked at Antec?

I assigned an engineer to develop a better spec for our app, then
we will resubmit to several makers.

 

for a eight card expansion, but we are shipping with only

An 8-card expansion PCI backplane? Or some custom card format?

A generic PCI card will have to be quite happy with junk on +12V,
because in a typical PC there is so much junk on +12V.

If this is some custom (analog?) application where you need high-spec
12V at lowish currents, you might go back to on-card regulators.

You still haven't really explained why you need Transzorbs on the 12V
bus in the first place. The only reason I can think of might be a
crowbar, but you haven't yet confirmed this suspicion of mine so maybe
there's some other reason.

Tim.

 

[Jim Adney]

On Wed, 12 Jan 2005 14:30:54 -0500 "MM" <mbmsv@yahoo.com> wrote:

"CBarn24050" <cbarn24050@aol.com> wrote in message
news:20050112141319.12040.00000080@mb-m15.aol.com...
Subject: Has anyone seen transorbs failing for no reason?

The most likely explaination is that you have specified the wrong voltage
device for your application.

What in your opinion would the right voltage to specify for a 12V rail if
the power supply spec is +/- 1%?

Transorbs should be selected to trip at voltages higher than anything
that rail should normally see. A typical application for a 12V rail
would be to pick a transorb that trips at around 15V. Don't try to use
these for voltage regulation; they've not very accurate and they're
only good for 1.5W continuous power. They are strictly for transient
protection, but they are extremely good for that.

I used to work at a place where over the years we used literally tens
of thousands of transorbs. The ONLY transorbs I ever saw fail were
ones which had been subjected to severe overvoltage without any
current limiting impedance. Those were a 15V pair which I installed in
a chart recorder to prevent damage. I failed to anticipate that
someone would try to hook a 120 V AC line to the 10V max inputs. Those
burned up, but the chart recorder lived.

On my next iteration I added 100K in front of the 15V transorbs, since
that would be insignificant compared to the 10 MOhm input impedance.
That was 10-15 years ago, and I believe those are still healthy today.

-
-----------------------------------------------
Jim Adney jadney@vwtype3.org
Madison, WI 53711 USA
-----------------------------------------------

 

[Clarence_A]

"MM" <mbmsv@yahoo.com> wrote in message
news:34o8d1F4dig5nU1@individual.net...

shoppa@trailing-edge.com> wrote in message
news:1105648026.811304.320970@f14g2000cwb.googlegroups.com...
It is basically a PC

The 12V rail on most PC-clone supplies is very poorly
regulated (it's
only regulated as a side-effect of the 5V regulation...).

The SMAJ12A you are using may trip as low as 13.3V, and I can
easily
see a PC-clone supply putting that out.

The power supply is not a cheap PC-clone. It is rather one of
the best
available:

http://www.pcpowerandcooling.com/products/power_supplies/highperformance/turbocools/510-atx-pfc.htm
/MM

SO sorry, THAT supply IS a cheap PC CLONE! Made in China! Only
the +5V is regulated. The other voltages are ratioed to the +5V.

 

[Clarence_A]

"MM" <mbmsv@yahoo.com> wrote in message
news:34oc3vF4bnf5uU1@individual.net...

"Clarence_A" <no@No.com> wrote in message
news:bvCFd.1520$8Z1.1125@newssvr14.news.prodigy.com...

The power supply is not a cheap PC-clone. It is rather one
of
the best
available:

http://www.pcpowerandcooling.com/products/power_supplies/highperformance/turbocools/510-atx-pfc.htm
/MM

SO sorry, THAT supply IS a cheap PC CLONE! Made in China!
Only
the +5V is regulated. The other voltages are ratioed to the
+5V.

As far as I know even the most expensive non-ATX cPCI supplies
use
cross-regulation. Show me a PC supply that doesn't? On the other
hand where
did you find all this information, like where it is made and
that it is
actually cross-regulated? There is no mentioning of China on
their website,
so is it your educated guess or do you actually know this for a
fact.
According to their spec +12V is regulated to 1%. As we are
talking I have a
scope sitting on this 12V rail set to trigger on 12.5V. Half a
day, still
waiting...

/MM

Well! Snippy=snippy!
I bought four of these for a project and the little manual in the
box was clearly labeled. However --- Where did you get the idea
that they were NOT made in china? Did you find a label saying
"MADE IN USA?" I didn't! Also, I called the manufacturer...
His factory (partner) is in Taiwan!

'Everyone' knows that the PC supplies are cross-regulated. Your
comments made me think 'you' missed it. But that is not true in
the case. Separate regulation.

On your test set up. How often do you cycle the supply on and off?
That is when the rails get crazy usually!


I presume you have met the minimum load requirement on the +5 and
+112 V outputs?
IF as you say it didn't get above 12.5V, Could it be the MOV was
out of spec? OR damaged? Also, why a MOV on the (somewhat
regulated) side of a supply?

 

[Jim Adney]

On Thu, 13 Jan 2005 17:52:41 -0500 "MM" <mbmsv@yahoo.com> wrote:

As far as I know even the most expensive non-ATX cPCI supplies use
cross-regulation. Show me a PC supply that doesn't? On the other hand where
did you find all this information, like where it is made and that it is
actually cross-regulated? There is no mentioning of China on their website,
so is it your educated guess or do you actually know this for a fact.
According to their spec +12V is regualted to 1%. As we are talking I have a
scope sitting on this 12V rail set to trigger on 12.5V. Half a day, still
waiting...

I suspect that you're right about the cross regulation, and that
should be fine since only the +5V parts are likely to be picky about
their supply voltage.

If you're monitoring the 12V rail, are you also varying the load on
the +5? I would expect the 12V rail to rise if the load on the 5V
increases.

Considering all the tolerances, I'd go higher on the transorb voltage.
Try a 15V transorb if you really think you need one here.

-
-----------------------------------------------
Jim Adney jadney@vwtype3.org
Madison, WI 53711 USA
-----------------------------------------------

 

[Spehro Pefhany]

On Thu, 13 Jan 2005 12:03:04 -0500, the renowned "MM"
<mbmsv@yahoo.com> wrote:

"CBarn24050" <cbarn24050@aol.com> wrote in message
news:20050113113209.25300.00000103@mb-m17.aol.com...

There is no point using one on a regulated rail, you put them before the
regulator.

I agree, but this has been sort of design tradition at the place where I
work. I know what I should do in the future. The question is why the hell it
failed on the regulated rail?

Thanks,
/MM

Aside from capacitors, is there a source of energy that could exceed
the rating of the TVS? Any motors, solenoids, that sort of stuff?


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com