Statpower update

H

hdavis

New member
Mar 13, 1998
518
0
0
>Henry
>
>I owe you big on this one. I would name my first born after you but
>he is 25 and might frown on that.
>
>>
>> To help them determine the root cause, can you give me the following
>> information:
>>

Thanks for all of the detailed information in one place. My manuals are in
teh coach. Does anyone know the current draw for teh Dana? I'd like to get
that to Statpower as well.

What we know so far:

1. failure cause is MOSFET (power transistor) that failed
2. fire was most likely a result of the MOSFET self-destructing
3. MOSFET failure took out circuitry that controlled the actual charger
outputs
4. failure also destroyed the circuitry that controlled the amount of
current & voltage to supply to each output

What I suspect:

1. MOSFET failure could easily be infant mortality (sorry Arch, you may
have been "it" for this particular type of failure.)
2. it's unlikely that Arch or his wiring was the root cause

Statpower QA group is anxious to discover the root cause of the problem -
whether it's circuit design, bad part, incorrect wiring, etc. I'll talk
with QA again in the morning to see if they want any other info and what
their next steps are.

Arch - they are most concerned that you get "back online" with your coach
ASAP. They don't want you or anyone else down because of their product.
(This is QA talking. A very good indicator of what the company actually
thinks. They're usually among the most conservative engineers in any
electronics company.) I personally would not have any great fears about
putting the new unit in place - but I completely understand wanting to
understand the failure BEFORE putting a replacement back in.

Anyone else with a Statpower problem PERSONALLY? If so, ship me the details
and I'll see what can be discovered. I don't have the time right now to
deal with "a friend of a friend problems."

Stay tuned.

Henry

Henry Davis Consulting, Inc / new product consulting
PO Box 1270 / product readiness reviews
Soquel, Ca 95073 / IP reviews
ph: (831) 462-5199 / full service marketing
fax: (831) 462-5198
http://www.henry-davis.com/ http://www.henry-davis.com
 
I as going to try to put together an ADMIN post about the right way to get
involved in someone else's situation. I don't think I need to now as Henry
did it by example. If you want to help out, get the individual's permission
first and all the "facts" as they see them, then contact the second party as
a "friend" of the first. Less possibility of ruffled feathers all around.

Good job Henry!
Patrick

>
> In a message dated 9/17/99 1:39:20 AM Central Daylight Time,

>
> >
> > 1. MOSFET failure could easily be infant mortality (sorry
> Arch, you may
> > have been "it" for this particular type of failure.)
> > 2. it's unlikely that Arch or his wiring was the root cause
> >
> Henry
>
> Again thanks very much. I want all of you to watch this one very
> close. This is the best example of how we can all be helped by
> sticking together. I thought I was at a dead end. Henry knew who
> to talk to. Now we are all going to learn a lot more. This is
> a very big
> lesson for all of us----because very few people have seen the net work
> like this. For those of you in business look, listen and learn---how
> can you make this beast work for you? Sorry folks but I find this
> exciting. New ground new everything. WOW! I am going to be on
> the road this weekend but I am not going to unsubscribe. Will
> check in when I can. May even see if I can find my old cellular
> modem.
>
> Take Care
> Arch 76 GB IL
>
 
Hi Henry...

>
> 1. MOSFET failure could easily be infant mortality (sorry Arch, you
may
> have been "it" for this particular type of failure.)

I'm waiting for them to get to an explanation of the shooting yellow
flame that met Arch's gaze as he was peering at the LEDs. What is the
condition of the electrolytics?

Regards,
John 74 Glacier WASH, DC GMC
 
Henry...

There was another GMCnetter who bought one of these units but didn't
install it promptly, so when it failed early in use, Statpower said
all they could offer was another refurbished unit (at full refurbished
price). The netter had returned it without an RMA so he was lucky
that is all that happened. I don't have his name (it is on another
database). So if this sounds familiar anyone, jump in... I know you
can't pick this up, based on this information Henry, but maybe someone
knows who I am referring to...

Regards,
John 74 Glacier WASH, DC GMC
 
Hi Arch...

Not trying to put my words into the description... but even with this
description, it sounds like more than a MOSFET going up, particularly
when you termed it an explosion... Tell us about the sound associated
with this. I would expect the MOSFET to make a pop, and maybe a small
flash as it went up. Thanks for clearing this up. I will be quiet
once they tell us the Capacitors are still sealed and did not explode.

Regards,
John 74 Glacier WASH, DC GMC
 
Just a guess, but doesn't the "MO" in MOSFET stand for "Metal Oxide"?
Wouldn't the hot oxide irrides when the MOSFET torched?

Way out on this limb,
Patrick

>
> In a message dated 9/17/99 9:46:34 AM Central Daylight Time,
> jdolan

>
> > I'm waiting for them to get to an explanation of the shooting yellow
> > flame that met Arch's gaze as he was peering at the LEDs.
> What is the
> > condition of the electrolytics?
> >
> John
>
> Maybe I should be a little more clear. There was a blast of yellow
> fire out of the top grill. There were no lingering flames.
> Just one big
> puff.
>
> Take Care
> Arch
>
 
These the ones?

=========================================

8-22-99 Dave Lowry dlowry

Put me on the list as also having a refurb'd Statpower failure. No flames,
buy that nasty arcing sound and some smoke. Also, not messing with wiring
and on shore power at the time.

Unit is now at Statpower for evaluation. I'll post results.

ps: Sure glad I kept the old hummer.

===========================================

8-24-99 Dave Lowry dlowry

Spoke to StatPower today.

They received mine "blind" 'cuz I didn't call ahead for an RMA, so it
wasn't on the fast track.

Since it was out of warranty, the best they could do for me was another
refurb'd unit, 2nd day air to CA for $200 (US). I should have it Thurs. and
it'll get installed Sat. The 1st one didn't get installed about 10 weeks
after receipt, so didn't have mmuch warranty left. This one will get tested
right away as we leave 9/5 on a 4,00 mi. trip that will include ALBQQ. At
least if this one's gonna crap out, maybe it'll be under warranty.

BTW I'll be packin' the old hummer and an extra fire extinguisher just in case.

===============================================

>Henry...
>
>There was another GMCnetter who bought one of these units but didn't
>install it promptly, so when it failed early in use, Statpower said
>all they could offer was another refurbished unit (at full refurbished
>price). The netter had returned it without an RMA so he was lucky
>that is all that happened. I don't have his name (it is on another
>database). So if this sounds familiar anyone, jump in... I know you
>can't pick this up, based on this information Henry, but maybe someone
>knows who I am referring to...
>
>Regards,
>John 74 Glacier WASH, DC GMC
 
Arch

The LORD works in misterious ways

Wayne

> In a message dated 9/17/99 10:16:33 AM Central Daylight Time,

>
> > Tell us about the sound associated
> > with this. I would expect the MOSFET to make a pop, and maybe a small
> > flash as it went up. Thanks for clearing this up. I will be quiet
> > once they tell us the Capacitors are still sealed and did not explode.
> >
> John
>
> Sound was like a gun shot. Louder than a .22. Maybe closer to
> a .38. My ears were ringing for about 2 hours. It was enough to
> scare the He** out of me.
>
> Take Care
> Arch
 
>In a message dated 9/17/99 9:46:34 AM Central Daylight Time,
jdolan
>writes:
>
>> I'm waiting for them to get to an explanation of the shooting yellow
>> flame that met Arch's gaze as he was peering at the LEDs. What is the
>> condition of the electrolytics?
>>
>John
>
>Maybe I should be a little more clear. There was a blast of yellow
>fire out of the top grill. There were no lingering flames. Just one big
>puff.
>

Arch,

Based on this description I'd say that you were "lucky" to see the MOSFET
fail. I'll find out from Statpower what type of package the transistor is
in and can tell you more from that. One of the failure modes for MOSFETs is
a short between power and ground INSIDE the part. At best the junction is
several hundred microns in size so you have a lot of energy dissipating in
a small connection.

I'll post the FACTs of this failure as Statpower figures it out.

Henry

Henry Davis Consulting, Inc / new product consulting
PO Box 1270 / product readiness reviews
Soquel, Ca 95073 / IP reviews
ph: (831) 462-5199 / full service marketing
fax: (831) 462-5198
http://www.henry-davis.com/ http://www.henry-davis.com
 
Bill Massey...

Amazing... that's exactly who I was thinking about. I guess since we
are making a record, we should let Dave jump in and decide whether to
include his details in this or not.

>These the ones?

>=========================================

>8-22-99 Dave Lowry dlowry

>Put me on the list as also having a refurb'd Statpower failure.

Regards,
John 74 Glacier WASH, DC GMC
 
>Just a guess, but doesn't the "MO" in MOSFET stand for "Metal Oxide"?
>Wouldn't the hot oxide irrides when the MOSFET torched?
>

The MOSFETs themselves are pretty small. So, if you think about what
happens when you short a small strand of wire (think of the indiviidual
strands of a multistrand wire) you'll got a flash and some of the metal is
vaporized.

A similar thing happens to transistors under soem failure conditions. Power
shorts (or nearly shorts) to ground. The junction heats up quickly, melts
the oxide that insulates power from ground, makes a better short, and then
the dead short packs enough energy into the device to vaporize the short.

The color of the flame, amount of smoke, "bang", and other obvious
reactions depends on a lot of factors including the package type, the exact
device type, power to the device, etc.

Henry

Henry Davis Consulting, Inc / new product consulting
PO Box 1270 / product readiness reviews
Soquel, Ca 95073 / IP reviews
ph: (831) 462-5199 / full service marketing
fax: (831) 462-5198
http://www.henry-davis.com/ http://www.henry-davis.com
 
Greetings!

Here's the lastest update from Statpower. Arch's unit is now in the
"problems" queue at Statpower for inclusion in their full Failure Analysis
efforts. They have determined that the cause of the failure was in fact the
MOSFET device. That caused the upstream and downstream failure of other
components. Such is the life of a MOSFET in power circuits.

Their working hypothesis is that there may have been a startup load when
the compressor cycled that initiated the failure. This could happen if the
engine battery was low in charge, the electrical connections were high
resistance etc AND the converter (charger) provided the start surge
current. Failure sometime after the initial inrush is not an unlikely
scenario. The part could also have just failed due to a latent
manufacturing defect (no fault of Statpower's and in general not detectable)

Statpower would like to receive compressor specs and model number for the
GMC so that they may create another test environment. Due to customs
issues, they would prefer to buy a new unit. I'm looking for advice from
those of you on the list who have some electrical background on what to
tell them. The Dana is out of production (I think). I'd like to give them
the original GM number and the Dana/Hadley numbers.

Next steps. Statpower and I will talk every few weeks while they work
through their problem logs. We may not get a detailed reason for the
failure since it will be difficult to recreate Arch's conditions. (Arch, I
think that installing and running your Statpower is a good thing to do. You
may consider trying the converter with the compressor running to either
duplicate the failure or convince yourself that it's OK. Your call.)
Nevertheless, Statpower wants to try to determine the cause since they want
their equipment to be as reliable as can be achieved within the realm of
commercial equipment.

Everyone should keep in mind that converter and battery chargers are a
design compromise between complete safety (from circuit failure) for the
converter and allowing devices like compressors to turn on and run while
the battery is charging. One of the compromises is between circuit
protection and premitting startup loads to pass through the unit so that it
can be used in practical situations. (startup loads are often many times
the steady state loads, so power supplies frequently permit short
excursions beyond the steady state rating).

Bottom line: sometimes converters and inverters are going to fail unless we
as users are extra conservative.

That's it for now.

Henry
Henry Davis Consulting, Inc / new product consulting
PO Box 1270 / product readiness reviews
Soquel, Ca 95073 / IP reviews
ph: (831) 462-5199 / full service marketing
fax: (831) 462-5198
http://www.henry-davis.com/ http://www.henry-davis.com
 
>In a message dated 9/28/99 7:31:49 PM Central Daylight Time,

>
>>
>> Their working hypothesis is that there may have been a startup load when
>> the compressor cycled that initiated the failure. This could happen if the
>> engine battery was low in charge, the electrical connections were high
>> resistance etc AND the converter (charger) provided the start surge
>>
>Henry
>
>First let me say thank you so very much for all of the help you have
>been through all of this.

My pleasure - someday I'll need another netter's unique professional
experience!

Then let me defend my lady. I admit battery
>was low when I started. Thats why I waited till the fan shut off on the
>charger---then waited a little longer till charge level was down to 10-12
>amps before hitting the key to turn on the compressor. I am not going
>to say that there is no way there could be a high resistance connection
>in the system but I have replaced all of the #4 battery cables with #2
>both positive and negative. I will admit that the wiring going to the
>compressor is stock. I dont know if any of that helps but keep
>asking and I will tell you what I know.

I don't think that there was a load problem as a result of your wiring (or
the battery for that matter). I think, based on what information they've
shared with me, that this is a simple case of a part giving up the ghost
early in the life curve. Now if you get a repeat, I think Statpower will be
extremely interested in reproducing the failure! I DON'T expect you to
havbe more problems.

Let us know how the plug in goes!

Henry

Henry Davis Consulting, Inc / new product consulting
PO Box 1270 / product readiness reviews
Soquel, Ca 95073 / IP reviews
ph: (831) 462-5199 / full service marketing
fax: (831) 462-5198
http://www.henry-davis.com/ http://www.henry-davis.com
 
Arch...

>Let us know how the plug in goes! Henry

Arch? Arrrrrch?

I think the company that made that MOS FET is in the wrong business...
If that much of any substance can make a sound like a .38, they ought
to be in the Flash Bang business.

Regards,
John 74 Glacier WASH, DC GMC
 
Well... as Arch said, we are learning how things go when the net gets
involved.

Here's my take on all this so far...

The explanation given to Henry by Statpower seems to correspond better
with the description of failure events recounted by Dave Lowry as to
what happened when his unit failed. Remember, there was a little bit
of 'button, button who's got the button' when they tried to make Arch
go away because it had mysteriously disappeared between warehouses and
they were sorry but we were not going to learn anything more from all
this.

Ask the quality guy what the serial number is, on the unit he is
evaluating. Dave's explanation of what happened is also consistent
with my experience of what happens when a MOS FET like this one fails
(nasty arcing sound and smoke). It certainly does not sound like a
.38, but let me tell you from personal experience, an electrolytic
exploding sounds very much like a .38. Here is a replay of Dave's
information, courtesy of Billy Massey.

>>
8-22-99 Dave Lowry dlowry
Put me on the list as also having a refurb'd Statpower failure. No
flames,
buy that nasty arcing sound and some smoke. Also, not messing with
wiring and on shore power at the time.

Unit is now at Statpower for evaluation. I'll post results.
ps: Sure glad I kept the old hummer.
===========================================

8-24-99 Dave Lowry dlowry
Spoke to StatPower today.
They received mine "blind" 'cuz I didn't call ahead for an RMA, so it
wasn't on the fast track.
Since it was out of warranty, the best they could do for me was
another refurb'd unit, 2nd day air to CA for $200 (US). I should have
it
Thurs. and it'll get installed Sat. The 1st one didn't get installed
about 10
weeks after receipt, so didn't have much warranty left. This one will
get
tested right away as we leave 9/5 on a 4,00 mi. trip that will include
ALBQQ.
At least if this one's gonna crap out, maybe it'll be under warranty.

BTW I'll be packin' the old hummer and an extra fire extinguisher just
in case.
===============================================

>Henry...
>
>There was another GMCnetter who bought one of these units but didn't
>install it promptly, so when it failed early in use, Statpower said
>all they could offer was another refurbished unit (at full
refurbished
>price). The netter had returned it without an RMA so he was lucky
>that is all that happened. I don't have his name (it is on another
>database). So if this sounds familiar anyone, jump in... I know you
>can't pick this up, based on this information Henry, but maybe
>someone knows who I am referring to...
>

Regards,
John 74 Glacier WASH, DC GMC
 
>touchpad on your laptop. ( Thats a joke)

Good deductive skills (no joke)

I know we're all thinking are we being fair to StatPower... I don't know
what to
think at this point because it was going in a really bad direction there for
a while, and they have done a 180 and are behaving 100% First Class now. I
haven't ruled out the possibility that one of their distributors, heard
about this (or read about this) and said, 'don't leave them with a bad taste
in their mouth if you know what's good for yourself.'

Dave Lowry is probably on his way to ALBQQ and we may not hear from him for
a while, so I'm just going to wait and see what turns up in terms of other
experiences with these units. Go on Arch, put in your ear plugs, and your
Eye Protection and plug-er-in. They wouldn't send you one that would do it
again... no way!

Regards,
John 74 Glacier WASH, DC GMC
 
>Dave Lowry is probably on his way to ALBQQ and we may not hear from him for
>a while, so I'm just going to wait and see what turns up in terms of other
>experiences with these units.

Just to keep everyone on teh same page -

I understand that Dave's unit is a straight 40 and not the 40+. So, we
can't make any observations about one form the other since they are
different internal designs. From my conversations with Statpower, the 40+
was a redesign that added features that custoemrs were asking for - and
eliminated some weaker points of the original design. Before anyone gets up
in arms about the early vs late designs (40 vs 40+) from what I can gather
the 40 is a generally reliable unit. Companies like Statpower want to
eliminate as many field returns as possible.

By my count we have 1 dead 40+ (Arch's) and 1 dead 40 (Dave's). Any other
folks with dead Statpower units?

Henry
Henry Davis Consulting, Inc / new product consulting
PO Box 1270 / product readiness reviews
Soquel, Ca 95073 / IP reviews
ph: (831) 462-5199 / full service marketing
fax: (831) 462-5198
http://www.henry-davis.com/ http://www.henry-davis.com
 
>By my count we have 1 dead 40+ (Arch's) and 1 dead 40 (Dave's). Any other
>folks with dead Statpower units?

I hope not... if we are anywhere near a random GMC sample, that is already a
statistically high percentage of recent GMC-Statpower purchases (and /or two
completely random flukes coming together within just weeks of each other.
With the Rally season on, we will hear more if there is more to hear.

I'm just saying, the explosion Arch went through (with sound magnitude more
or less equivalent to a 38 caliber revolver) with ears ringing for hours, is
not your everyday typical semiconductor failure. Dave's was.
(ZZzzzzzitTT).

Regards,
John 74 Glacier WASH, DC GMC
 
>>By my count we have 1 dead 40+ (Arch's) and 1 dead 40 (Dave's). Any other
>>folks with dead Statpower units?
>
>I hope not... if we are anywhere near a random GMC sample, that is already a
>statistically high percentage of recent GMC-Statpower purchases (and /or two
>completely random flukes coming together within just weeks of each other.

Actually we can't draw a heck of a lot of inferences from the sample size
we have.We know that the total possible sample from this list is around
200. But not everybody has a Statpower unit - I'd guess less than 50 people
on the list have either a 40 or 40+ based on my recollection of earlier
e-mails. So, our apparant failure rate is 4% - but that isn't how we tell
failure rates. Electronic failure rates are measured (generally) as
failures per unit time under known conditions. And the units under test
must be the same configuration. So, what we have is ONE failure for the 40
and ONE failure for the 40+ with an unknown number of hours on the total
population.

FWIW, semiconductors undergo accelerated life test during the product
qualification process. Different companies have different procedures, but
they all work to correlate early life failure with specific manufacturing
defects. So the business units that I ran permitted a certain number of
specific types of failures in the first 100 hours of accelerated life test
(more than that number had been shown previously to coorelate to specific
kinds of manufacturing problems). From the failures recorded during the
accelerated life test we calculated a failure rate. Parts were tested to
1000 hours at elevated temperatures (accelerates the aging process) during
the qualification process. All parts remained on accelerated life test
until they stopped working. Samples from different manufacturing lots also
went into life test to add more data.

All of this long-winded stuff is just my short lead up to reminding
everyone that we have too small a sample to draw any reliability
conclusions. AND the 40 and 40+ are two different products that we can't
lump together.

In short, I'm not concerned about Arch's failure from what I know so far.
(I have a 40+ that I've run for over 1000 hours in my Palm Beach).

>With the Rally season on, we will hear more if there is more to hear.
>
>I'm just saying, the explosion Arch went through (with sound magnitude more
>or less equivalent to a 38 caliber revolver) with ears ringing for hours, is
>not your everyday typical semiconductor failure. Dave's was.

Most semiconductors never create any amount of noise that could be heard in
a GMC if there's even the slightest background noise. Power transistors are
a different story - especially MOSFETs. I've had metal can transistors used
in power circuits blow up and propel the can across the room. One of my
techs went to the emergency room in 1977 when a power supply transitor
became a fuse for the supply. The metal can had hit his safety glasses and
embedded itself in his check. I've also had power transistors die without a
wimper - just a puff of smoke and bad smells.

More FWIW(not too technical and not complete) - FETs in general have
several failure modes that can result in spectacular events if there is an
adequate power supply present. The structure of the transistor depends on a
number of factors, but oxide is one of the materials used to separate
active regions. Oxide is used to insulate the gate (between source and
drain). The oxide sometimes gets pinholes due to very small particles on
tooling or in the air, or other defects in manufacturing. A small pinhole
results in a leakage current between the source and drain. Over time the
leak can get worse as the current flow makes the defect in the oxide bigger
(if there is enough flow). Eventually the whole thing fails. One common
failure mechaism is due to static electricity discharge. When a static
discharge hits the semiconductor, the energy goes somewhere. One of the
ways that static discharges is to jump the oxide barrier (the oxide is
typically about a thousand angstroms thick - less for some advanced
computer chips) The part may seem to work fine after the static zap, and it
may be perfectly fine. However, sometimes static ruptures the gate oxide or
causes a defect to occur that causes leakage, that gets worse, that fails
one day.

Moral of the FWIW story - follow the static protection guidelines provided
with your equipment. You may or may not "get away" with sloppy procedures.

regards,

Henry

Henry Davis Consulting, Inc / new product consulting
PO Box 1270 / product readiness reviews
Soquel, Ca 95073 / IP reviews
ph: (831) 462-5199 / full service marketing
fax: (831) 462-5198
http://www.henry-davis.com/ http://www.henry-davis.com
 
You must log in or register to reply here.
Get your shine on at www.cruzingear.co
Applied GMC

Share this page

Share this page
Top Bottom
Back