Potting horror...
Nico Coesel
A customer of mine has a problem with a PCB potted in polyurethane.
The circuit consists of some inputs (clamped & filtered), a
microcontroller with internal RAM / flash and some power electronics
to switch loads (up to 40V). 10 PCBs where produced and tested. All
worked fine. 4 of those PCBs where potted and they started to show
erratic behaviour like RAM corruption, software hang-ups. I already
milled one of the modules to expose the microcontroller to check
whether it is hot, but it stays cold. So its not a heat problem.
The polyurethane used is non-conductive and shouldn't shrink, the
potting was done by a company that specializes in potting electronics;
they know about ESD precautions.
Any ideas where to start looking?
--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
"If it doesn't fit, use a bigger hammer!"
--------------------------------------------------------------
Dirk Bruere at NeoPax
> A customer of mine has a problem with a PCB potted in polyurethane.
> The circuit consists of some inputs (clamped & filtered), a
> microcontroller with internal RAM / flash and some power electronics
> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> worked fine. 4 of those PCBs where potted and they started to show
> erratic behaviour like RAM corruption, software hang-ups. I already
> milled one of the modules to expose the microcontroller to check
> whether it is hot, but it stays cold. So its not a heat problem.
>
> The polyurethane used is non-conductive and shouldn't shrink, the
> potting was done by a company that specializes in potting electronics;
> they know about ESD precautions.
>
> Any ideas where to start looking?
>
Capacitance changes?
--
Dirk
http://www.transcendence.me.uk/ - Transcendence UK
http://www.theconsensus.org/ - A UK political party
http://www.onetribe.me.uk/wordpress/?cat=5 - Our podcasts on weird stuff
WangoTango
>
> A customer of mine has a problem with a PCB potted in polyurethane.
> The circuit consists of some inputs (clamped & filtered), a
> microcontroller with internal RAM / flash and some power electronics
> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> worked fine. 4 of those PCBs where potted and they started to show
> erratic behaviour like RAM corruption, software hang-ups. I already
> milled one of the modules to expose the microcontroller to check
> whether it is hot, but it stays cold. So its not a heat problem.
>
> The polyurethane used is non-conductive and shouldn't shrink, the
> potting was done by a company that specializes in potting electronics;
> they know about ESD precautions.
>
> Any ideas where to start looking?
>
>
You milled away the compound and it stays cool 'now'.
Does the unit still lock up that you cleared the compound away from the
MCU?
Could the compound have flowed somewhere it wasn't suppose to and it is
now mechanically causing an intermittent connection?
Dirk Bruere at NeoPax
legg
wrote:
>
>A customer of mine has a problem with a PCB potted in polyurethane.
>The circuit consists of some inputs (clamped & filtered), a
>microcontroller with internal RAM / flash and some power electronics
>to switch loads (up to 40V). 10 PCBs where produced and tested. All
>worked fine. 4 of those PCBs where potted and they started to show
>erratic behaviour like RAM corruption, software hang-ups. I already
>milled one of the modules to expose the microcontroller to check
>whether it is hot, but it stays cold. So its not a heat problem.
>
>The polyurethane used is non-conductive and shouldn't shrink, the
>potting was done by a company that specializes in potting electronics;
>they know about ESD precautions.
>
>Any ideas where to start looking?
Repeat the process, retested at various stages, including shipping, to
ensure that potting really is the significant factor.
Right now you seem to have only verbal assurances that proper
processes were used on the first 4.
Any open magnetics, resonators or electromechanical components
on-board?
RL
Nico Coesel
>WangoTango wrote:
>> In article <49948dda....@news.planet.nl>, ni...@puntnl.niks says...
>>> A customer of mine has a problem with a PCB potted in polyurethane.
>>> The circuit consists of some inputs (clamped & filtered), a
>>> microcontroller with internal RAM / flash and some power electronics
>>> to switch loads (up to 40V). 10 PCBs where produced and tested. All
>>> worked fine. 4 of those PCBs where potted and they started to show
>>> erratic behaviour like RAM corruption, software hang-ups. I already
>>> milled one of the modules to expose the microcontroller to check
>>> whether it is hot, but it stays cold. So its not a heat problem.
>>>
>>> The polyurethane used is non-conductive and shouldn't shrink, the
>>> potting was done by a company that specializes in potting electronics;
>>> they know about ESD precautions.
>>>
>>> Any ideas where to start looking?
>>>
>>>
>> How do you know the MCU stays cool when it is potted?
I only milled away a small part of the compound (just enough to reach
the 48 pin package and get to the pins) and covered the hole while
testing. The MCU is not consuming a lot of power.
>> You milled away the compound and it stays cool 'now'.
>> Does the unit still lock up that you cleared the compound away from the
>> MCU?
Yes, the unit still bahaves odd.
>> Could the compound have flowed somewhere it wasn't suppose to and it is
>> now mechanically causing an intermittent connection?
That could be the case, but that wouldn't explain odd behaviour on
multiple modules.
>Also, how is the clock generated?
A 10MHz crystal in a HC49 SMD metal case with the usual capacitors
(around 20pf). I also measured the crystal's frequency on a potted and
non-potted PCB, the readings where correct. I didn't measure the
amplitude yet.
John Larkin
wrote:
>
>A customer of mine has a problem with a PCB potted in polyurethane.
>The circuit consists of some inputs (clamped & filtered), a
>microcontroller with internal RAM / flash and some power electronics
>to switch loads (up to 40V). 10 PCBs where produced and tested. All
>worked fine. 4 of those PCBs where potted and they started to show
>erratic behaviour like RAM corruption, software hang-ups. I already
>milled one of the modules to expose the microcontroller to check
>whether it is hot, but it stays cold. So its not a heat problem.
>
>The polyurethane used is non-conductive and shouldn't shrink, the
>potting was done by a company that specializes in potting electronics;
>they know about ESD precautions.
>
>Any ideas where to start looking?
Potting is always a horror. It's expensive, messy, often thermally
horrible, can wreck parts, makes debugging almost impossible. Some
potting compounds exotherm big time, or shrink and crush parts. Even
when it seems like a good idea, it's usually not.
John
RFI-EMI-GUY
> A customer of mine has a problem with a PCB potted in polyurethane.
> The circuit consists of some inputs (clamped & filtered), a
> microcontroller with internal RAM / flash and some power electronics
> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> worked fine. 4 of those PCBs where potted and they started to show
> erratic behaviour like RAM corruption, software hang-ups. I already
> milled one of the modules to expose the microcontroller to check
> whether it is hot, but it stays cold. So its not a heat problem.
>
> The polyurethane used is non-conductive and shouldn't shrink, the
> potting was done by a company that specializes in potting electronics;
> they know about ESD precautions.
>
> Any ideas where to start looking?
>
terminated properly (pulled up/down) and now you have a big slab of ESD
producing plastic?
--
Joe Leikhim K4SAT
"The RFI-EMI-GUY"©
"Use only Genuine Interocitor Parts" Tom Servo ;-P
Eeyore
Nico Coesel wrote:
> A customer of mine has a problem with a PCB potted in polyurethane.
> The circuit consists of some inputs (clamped & filtered), a
> microcontroller with internal RAM / flash and some power electronics
> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> worked fine. 4 of those PCBs where potted and they started to show
> erratic behaviour like RAM corruption, software hang-ups. I already
> milled one of the modules to expose the microcontroller to check
> whether it is hot, but it stays cold. So its not a heat problem.
>
> The polyurethane used is non-conductive and shouldn't shrink, the
> potting was done by a company that specializes in potting electronics;
> they know about ESD precautions.
>
> Any ideas where to start looking?
I suggest they stop potting. What an ANCIENT idea.
Graham
Eeyore
Nico Coesel wrote:
> A customer of mine has a problem with a PCB potted in polyurethane.
> The circuit consists of some inputs (clamped & filtered), a
> microcontroller with internal RAM / flash and some power electronics
> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> worked fine. 4 of those PCBs where potted and they started to show
> erratic behaviour like RAM corruption, software hang-ups. I already
> milled one of the modules to expose the microcontroller to check
> whether it is hot, but it stays cold. So its not a heat problem.
>
> The polyurethane used is non-conductive and shouldn't shrink, the
> potting was done by a company that specializes in potting electronics;
> they know about ESD precautions.
>
> Any ideas where to start looking?
Is the potting compound lead-free compatible ?
Graham
Eeyore
John Larkin wrote:
Couldn't agree more.
Graham
Eeyore
RFI-EMI-GUY wrote:
Whoever's responsible, I hope they get a good arse-kicking unlike the
bankers.
Graham
Dirk Bruere at NeoPax
Just wondering if the potting compound might change the waveform.
Bob Eld
"Nico Coesel" <ni...@puntnl.niks> wrote in message
news:49948dda....@news.planet.nl...
I assume the boards all tested ok and worked properly before potting. Were
all functions tested?
Is there an unpotted working control board?
How were the boards cleaned and dried before potting?
Was the urethane properly and sufficiently mixed?
Was the urethane thermally cured, what temp?
Could there be any other surface contamination, flux or other contaminants?
Could there be tin whiskers from unleaded solder propagating along the
urethane-board interface.
Could there be other creepage issues within the urethane-board interface
such as moisture, etc.
Are you sure about the urethane, could it be a thermal but poorly insulating
compound, aluminum filled for example.
Was the potting vacuumed or could there be bubbles on internal surfaces with
possible condensation or other contamination.
Many questions here but one main one is the proper operation of a working
unpotted control board to insure that problem is in the potting and not just
the circuit or software such as open pins, etc.
MooseFET
> A customer of mine has a problem with a PCB potted in polyurethane.
> The circuit consists of some inputs (clamped & filtered), a
> microcontroller with internal RAM / flash and some power electronics
> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> worked fine. 4 of those PCBs where potted and they started to show
> erratic behaviour like RAM corruption, software hang-ups. I already
> milled one of the modules to expose the microcontroller to check
> whether it is hot, but it stays cold. So its not a heat problem.
>
> The polyurethane used is non-conductive and shouldn't shrink, the
> potting was done by a company that specializes in potting electronics;
> they know about ESD precautions.
>
> Any ideas where to start looking?
Was this a two part polyurethane that has big trouble if there is any
water about? I assume it is.
Never apply power to the board until the reaction is totally finished.
Did someone make the mistake of using "no clean" flux? If they did
abandon all hope of potting the boards in question. Make some more
with washed flux and wash them. Beware of the water from washing
them.
Is this polyurethane colored or is it the sort of clear stuff? The
sort of clear works better than the stuff with pigments.
Is the failure heat related?
What silly fool decided to pot them? Perhaps you can delete the
potting and et on with production.
mako...@yahoo.com
did the boards work before potting
did you unpot one of the bad boards
did it start to work after unpotting
what speed is the uc
what is the dielectric constant of the potting
are there ANY hi impedance nodes anywhere in the system including
power supply, reset circuit, or anywhere else
what about the busses, do they go hi Z sometimes
was the potting fully cured and dry before powering up
let us know what you find
Mark
Martin Riddle
"Nico Coesel" <ni...@puntnl.niks> wrote in message
news:49948dda....@news.planet.nl...
>
> A customer of mine has a problem with a PCB potted in polyurethane.
> The circuit consists of some inputs (clamped & filtered), a
> microcontroller with internal RAM / flash and some power electronics
> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> worked fine. 4 of those PCBs where potted and they started to show
> erratic behaviour like RAM corruption, software hang-ups. I already
> milled one of the modules to expose the microcontroller to check
> whether it is hot, but it stays cold. So its not a heat problem.
>
> The polyurethane used is non-conductive and shouldn't shrink, the
> potting was done by a company that specializes in potting electronics;
> they know about ESD precautions.
>
> Any ideas where to start looking?
>
> --
Boards were dirty or contaminated. The potting probably never fully
cured.
That’s really the first place to look.
Cheers
Phil Allison
"Eeyore"
John Larkin wrote:
>>
>> Potting is always a horror. It's expensive, messy, often thermally
>> horrible, can wreck parts, makes debugging almost impossible. Some
>> potting compounds exotherm big time, or shrink and crush parts. Even
>> when it seems like a good idea, it's usually not.
>
> Couldn't agree more.
** Ever see inside a H-H " Dual 500 " power amp from the mid 70s ???
The two drive PCBs ( 6 x 3 inches) were fitted into a plastic tray filled
with hard black resin leaving only a bunch of pins poking out.
Madness.
..... Phil
John Larkin
That's a good suggestion. We've received boards from outside
assemblers that would drip water if you shake them.
It's hard to find assemblers nowadays who use solvent wash. And
sometimes they use dirty water, or don't completely dry the boards,
especially under parts. Potting will lock in any wet crud, and
electrolytic corrosion will eventually mess things up.
John
John Larkin
<OneBi...@InfiniteSeries.Org> wrote:
> You really are dumb, John.
At least I have a name.
>
> The potting materials discussed here have elasticity, not hard epoxy
>crush capacity, like 'stycast' has.
>
> Properly chosen media, and properly utilized, potting has a VERY
>SPECIFIC place in the electronics industry.
Sometimes. But more often it's a bad idea, often driven by paranoia.
>
> Sad that you are disconnected from the facts to the point that your
>opinions make you a really bad 'engineer'.
I'm a pretty good, degreed engineer, without quote marks. I've posted
lots of links to, and pics of, my stuff, so people can judge for
themselves.
You never post any. You don't even have a name. I think you're a
frustrated tech.
>
> That's OK. You only recently got out of plain cardboard stock bins.
>Sad. Do you think there is no use for ESD precautions as well?
Why sad? We made a ton of money out of the parts that were stored in
those bins. Several tons, actually.
And our ESD controls are better than they need to be. No problems at
all.
The reshuffle of the stock room was driven by new inventory/BOM
software and a total renumbering of all the parts in stock. It all
worked great... 4800 parts, about 1.6 million pieces, relabeled and
rearranged on the shelves, about 600 BOMs converted to the new number
system, all the computers changed to use the new system, all done in
two days. We love it.
All the parts are now described in a standardized way, and all part
numbers are logical. For example, all resistors are now in order by
both part number and resistance. Datasheets and notes/pictures are
linked to part records, and you can instantly see all the places where
a part is used. And you can pull up a BOM and see the total parts
cost, instantly. Searching is so fast you usually can't see any delay
at all.
John
Frithiof Jensen
"Nico Coesel" <ni...@puntnl.niks> skrev i meddelelsen
news:49948dda....@news.planet.nl...
>
> Any ideas where to start looking?
Maybe there is water inside the assembly: They cleaned the boards thoroughly
before potting but for some reason did not get them entirely dry before
potting.
Maybe there is some mechanical problem like a dubious solder joint that
normally sticks together but the potting just ever-so-slightly pulls it
apart.
They could, on this occasion - even though the potters are expert e.t.c.,
have got the mixture wrong so the board is chrushed/cooked (or the potting
goo is conductive). Or the potting compund shrink whatever they say!
Maybe the design is marginal anyway and the few pF extra in stray
capacitance from the potting compound moves the timing just enough to hit a
race condition that was always there. Is it worth dunking a board in
paraffin oil and measure?
From previous experience with potting I would guess on theory '3' - yah "the
potting compund doesn't shrink" but it did ANYWAY so before potting we had
to dip the boards in a silicone-based goo so that the rubbery coating would
take the strain from the potting. (It was a quite funny fault: Pulse
transformers that lost all inductance, like the core disappeared, then
recovered when opened for testing).
Wimpie
> A customer of mine has a problem with a PCB potted in polyurethane.
> The circuit consists of some inputs (clamped & filtered), a
> microcontroller with internal RAM / flash and some power electronics
> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> worked fine. 4 of those PCBs where potted and they started to show
> erratic behaviour like RAM corruption, software hang-ups. I already
> milled one of the modules to expose the microcontroller to check
> whether it is hot, but it stays cold. So its not a heat problem.
>
> The polyurethane used is non-conductive and shouldn't shrink, the
> potting was done by a company that specializes in potting electronics;
> they know about ESD precautions.
>
>
> Failure does not prove something is impossible, failure simply
> indicates you are not using the right tools...
> "If it doesn't fit, use a bigger hammer!"
> --------------------------------------------------------------
Hi Nico,
Looking to your profile, you are in NL? Did you work with Purtec?
When the PU company did a good job, and you did a good design job with
sufficient margin, I would first look to contaminants on the board
that may influence the curing of the PU or even lead to gas production
(results in high stress). Even residues of adhesive tape can cause
problems.
Cut one PCB and look for the uniformity of the PU. For the other
boards, run a long test at elevated temperature and see what happens
over time. It is possible that other board will fail also.
To check proper functioning (of the clock oscillator circuitry), you
might run a temperature sweep. When you do not have sufficient margin
in the oscillator, it will fail at low or high temperature. You may
also check for slight frequency drift after potting (of course this
requires to measure the frequency before potting).
Though shrinkage may be low, you will always have some mechanical
stress due to shrinkage.
Best regards,
Wim
PA3DJS
www.tetech.nl
when you remove abc, the address is valid
typhon62
wrote:
I have a board that in in the engine compartment of a pickup truck.
It's wet, dirty, and gets salt on it in the winter.
What else besides potting would I use to protect the PCB and
electronics?
I hate potting, but I really don't see another way to keep all of the
road crud
off of my electronics...
Thanks!
Dirk Bruere at NeoPax
Hey! It's Phil Allison, the guy with the girl's name!
krw
jjla...@highNOTlandTHIStechnologyPART.com says...>
So does DimBulb.
> > The potting materials discussed here have elasticity, not hard epoxy
> >crush capacity, like 'stycast' has.
> >
> > Properly chosen media, and properly utilized, potting has a VERY
> >SPECIFIC place in the electronics industry.
>
> Sometimes. But more often it's a bad idea, often driven by paranoia.
The *one* time I've had my designs potted it was for security
reasons. It was an *expen$ive* PITA.
> > Sad that you are disconnected from the facts to the point that your
> >opinions make you a really bad 'engineer'.
>
> I'm a pretty good, degreed engineer, without quote marks. I've posted
> lots of links to, and pics of, my stuff, so people can judge for
> themselves.
>
> You never post any. You don't even have a name. I think you're a
> frustrated tech.
DimBulb is a tech, certainly not a degreed engineer (you've noted
the phobia of all things math). Nothing wrong with being a tech,
though not nearly as fun. Notice how Dimmie always coaches his
words. "We make" or "our product", never "my design".
> > That's OK. You only recently got out of plain cardboard stock bins.
> >Sad. Do you think there is no use for ESD precautions as well?
> Why sad? We made a ton of money out of the parts that were stored in
> those bins. Several tons, actually.
>
> And our ESD controls are better than they need to be. No problems at
> all.
Why does Dimmie think cardboard can't be ESD treated? Many parts
come in card bard containers.
> The reshuffle of the stock room was driven by new inventory/BOM
> software and a total renumbering of all the parts in stock. It all
> worked great... 4800 parts, about 1.6 million pieces, relabeled and
> rearranged on the shelves, about 600 BOMs converted to the new number
> system, all the computers changed to use the new system, all done in
> two days. We love it.
Oohh. Our purchasing people would croak.
> All the parts are now described in a standardized way, and all part
> numbers are logical. For example, all resistors are now in order by
> both part number and resistance. Datasheets and notes/pictures are
> linked to part records, and you can instantly see all the places where
> a part is used. And you can pull up a BOM and see the total parts
> cost, instantly. Searching is so fast you usually can't see any delay
> at all.
Ours *did* have a standardized part numbering system with the value
as the last digits. From what I'm told, they went to a new release
system that wanted to sequentially assign part numbers, though any
string *could* be assigned. Whoever put the software in set it up
so the old P/Ns were preserved but new ones were sequential. What
a PITA. They *HAD* it right, then blew it. I'm sure the owner has
no idea how much that decision cost him.
MooseFET
A sealed housing works too. Circuit boards I have designed operate in
the ocean. Although it isn't hot like an engine, there are a few
thousand pounds of pressure pushing the salt water in. With good
design it stays on the outside and the PCB keeps working. Aluminum and
O-rings work well.
>
> Thanks!
Vladimir Vassilevsky
John Larkin wrote:
> The reshuffle of the stock room was driven by new inventory/BOM
> software and a total renumbering of all the parts in stock. It all
> worked great... 4800 parts, about 1.6 million pieces, relabeled and
> rearranged on the shelves, about 600 BOMs converted to the new number
> system, all the computers changed to use the new system, all done in
> two days. We love it.
>
> All the parts are now described in a standardized way, and all part
> numbers are logical. For example, all resistors are now in order by
> both part number and resistance. Datasheets and notes/pictures are
> linked to part records, and you can instantly see all the places where
> a part is used. And you can pull up a BOM and see the total parts
> cost, instantly. Searching is so fast you usually can't see any delay
> at all.
This is a big accomplishment. Can you share your experience in more
details; i.e. what softwares did you use; how do you link the sch/pcb
databases with the inventory; how do you link the BOMs to the specific
vendors, who is responsible for keeping the system and how many persons
are involved. Do you have the mechanical parts in the same database
with the electronic components?
My biggest frustration with the systems like that in the past was
because of the hell of bureaucracy, inconvenience and associated
overhead. So people tend to fool the system and do the things in their
own old ways, which makes the system worseless...
Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com
John Larkin
<antispa...@hotmail.com> wrote:
>
>
>John Larkin wrote:
>
>
>> The reshuffle of the stock room was driven by new inventory/BOM
>> software and a total renumbering of all the parts in stock. It all
>> worked great... 4800 parts, about 1.6 million pieces, relabeled and
>> rearranged on the shelves, about 600 BOMs converted to the new number
>> system, all the computers changed to use the new system, all done in
>> two days. We love it.
>>
>> All the parts are now described in a standardized way, and all part
>> numbers are logical. For example, all resistors are now in order by
>> both part number and resistance. Datasheets and notes/pictures are
>> linked to part records, and you can instantly see all the places where
>> a part is used. And you can pull up a BOM and see the total parts
>> cost, instantly. Searching is so fast you usually can't see any delay
>> at all.
>
>This is a big accomplishment. Can you share your experience in more
>details; i.e. what softwares did you use; how do you link the sch/pcb
>databases with the inventory; how do you link the BOMs to the specific
>vendors, who is responsible for keeping the system and how many persons
> are involved. Do you have the mechanical parts in the same database
>with the electronic components?
It's our own thing, a single PowerBasic program, 17k lines including
comments and whitespace. There's a master parts file of fixed-length
binary records, a folder full of text-mode BOMs, and a folder that
contains folders of miscellaneous data, like datasheets, photos,
notes, one sub-folder per part.
At startup, it reads the entire parts file into memory arrays, and
indexes the BOMs and data folders, which takes a few seconds. After
that, it's blindingly fast. Most commercial packages seem to build on
top of some hairy general-purpose database manager, and then do
everything over the network, and get slow.
A part number is of the form nnn-mmmm, where nnn is a category, like
RES 0805 or some such. The mmmm is encoded somehow, like by resistance
in this case.
>
>My biggest frustration with the systems like that in the past was
>because of the hell of bureaucracy, inconvenience and associated
>overhead. So people tend to fool the system and do the things in their
>own old ways, which makes the system worseless...
This system was designed by engineering and manufacturing people to do
exactly what we want. It works, like its predecessor DOS version
(which used 5-digit numbers that eventually got to be sort of random)
because the people make it work, and follow the rules, which are all
available in writing.
We may decide to sell it. People wouldn't be buying just the software,
but the whole set of procedures and concepts, which might be very
healthy for some companies without that discipline. Some people really
don't understand the whole thing: dash numbers, assemblies and
subassemblies, revision control, ECOs, BOMs, parts numbering, all
that.
John
John Larkin
Diecast plastic or aluminum box with o-rings, sealed connectors,
conformal coating maybe.
John
Jim Thompson
In other words, a typical automotive housing.
...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine Sometimes I even put it in the food
typhon62
>
> - Show quoted text -
I'd love to have a nice die cast box, but we only make a 100 or so of
these and a die cast box would break the bank. Plus I have to have
to use connectors that are not easily (impossible actually) sealed
in a box. We tried conformal coating but had corrosion problems
because for some reason the coating did not want to stick to certain
components no matter how clean they were.
Conformal coating is a pain in the ass but for our small quantities
it seems to be the only option...
WangoTango
says...
> Nico Coesel wrote:
> > Dirk Bruere at NeoPax <dirk....@gmail.com> wrote:
> >
> >> WangoTango wrote:
> >>> In article <49948dda....@news.planet.nl>, ni...@puntnl.niks says...
> >>>> A customer of mine has a problem with a PCB potted in polyurethane.
> >>>> The circuit consists of some inputs (clamped & filtered), a
> >>>> microcontroller with internal RAM / flash and some power electronics
> >>>> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> >>>> worked fine. 4 of those PCBs where potted and they started to show
> >>>> erratic behaviour like RAM corruption, software hang-ups. I already
> >>>> milled one of the modules to expose the microcontroller to check
> >>>> whether it is hot, but it stays cold. So its not a heat problem.
> >>>>
> >>>> The polyurethane used is non-conductive and shouldn't shrink, the
> >>>> potting was done by a company that specializes in potting electronics;
> >>>> they know about ESD precautions.
> >>>>
> >>>> Any ideas where to start looking?
> >>>>
> >>>>
> >
> > I only milled away a small part of the compound (just enough to reach
> > the 48 pin package and get to the pins) and covered the hole while
> > testing. The MCU is not consuming a lot of power.
> >
> >>> You milled away the compound and it stays cool 'now'.
> >>> Does the unit still lock up that you cleared the compound away from the
> >>> MCU?
> >
> > Yes, the unit still bahaves odd.
> >
> >>> Could the compound have flowed somewhere it wasn't suppose to and it is
> >>> now mechanically causing an intermittent connection?
> >
> > That could be the case, but that wouldn't explain odd behaviour on
> > multiple modules.
I would if there was a gap in the same place in all the units.
Stranger things have happened.
> >
> >> Also, how is the clock generated?
> >
> > A 10MHz crystal in a HC49 SMD metal case with the usual capacitors
> > (around 20pf). I also measured the crystal's frequency on a potted and
> > non-potted PCB, the readings where correct. I didn't measure the
> > amplitude yet.
> >
>
> Just wondering if the potting compound might change the waveform.
Good call.
Check wave form shape AND amplitude.
I had a similar problem with a product 15 years ago.
Changed units over to an oscillator module and that was that.
The can of the module became the equivalent of a bubble around the
crystal and caps.
WangoTango
> Did you ever stop to think that this design may require potting?
>
> Did you ever stop to think... ever?
>
Ah, another student of Phil Allison.
Off to the killfile for you dipshit.
Nico Coesel
>
>
>Nico Coesel wrote:
>
>> A customer of mine has a problem with a PCB potted in polyurethane.
>> The circuit consists of some inputs (clamped & filtered), a
>> microcontroller with internal RAM / flash and some power electronics
>> to switch loads (up to 40V). 10 PCBs where produced and tested. All
>> worked fine. 4 of those PCBs where potted and they started to show
>> erratic behaviour like RAM corruption, software hang-ups. I already
>> milled one of the modules to expose the microcontroller to check
>> whether it is hot, but it stays cold. So its not a heat problem.
>>
>> The polyurethane used is non-conductive and shouldn't shrink, the
>> potting was done by a company that specializes in potting electronics;
>> they know about ESD precautions.
>>
>> Any ideas where to start looking?
>
>I suggest they stop potting. What an ANCIENT idea.
Actually potting has evolved into a very useful technique and fills a
gap between a standard housing and custom moulded case. In this case
the potting compound is also a very slick looking housing.
Nico Coesel
>On Feb 12, 1:10=A0pm, n...@puntnl.niks (Nico Coesel) wrote:
>> A customer of mine has a problem with a PCB potted in polyurethane.
>> The circuit consists of some inputs (clamped & filtered), a
>> microcontroller with internal RAM / flash and some power electronics
>> to switch loads (up to 40V). 10 PCBs where produced and tested. All
>> worked fine. 4 of those PCBs where potted and they started to show
>> erratic behaviour like RAM corruption, software hang-ups. I already
>> milled one of the modules to expose the microcontroller to check
>> whether it is hot, but it stays cold. So its not a heat problem.
>>
>> The polyurethane used is non-conductive and shouldn't shrink, the
>> potting was done by a company that specializes in potting electronics;
>> they know about ESD precautions.
>>
>> Any ideas where to start looking?
>
>Was this a two part polyurethane that has big trouble if there is any
>water about? I assume it is.
Good question. I'll check that with the company that did the potting.
>Never apply power to the board until the reaction is totally finished.
>
>Did someone make the mistake of using "no clean" flux? If they did
>abandon all hope of potting the boards in question. Make some more
>with washed flux and wash them. Beware of the water from washing
>them.
Sounds reasonable. What is the problem with no-clean flux and potting?
>Is this polyurethane colored or is it the sort of clear stuff? The
>sort of clear works better than the stuff with pigments.
Its black.
>Is the failure heat related?
I doubt it. I already checked that with the one I milled open.
>What silly fool decided to pot them? Perhaps you can delete the
>potting and et on with production.
Potting it is an affordable way to get a slick looking custom housing.
qrk
wrote:
>
>A customer of mine has a problem with a PCB potted in polyurethane.
>The circuit consists of some inputs (clamped & filtered), a
>microcontroller with internal RAM / flash and some power electronics
>to switch loads (up to 40V). 10 PCBs where produced and tested. All
>worked fine. 4 of those PCBs where potted and they started to show
>erratic behaviour like RAM corruption, software hang-ups. I already
>milled one of the modules to expose the microcontroller to check
>whether it is hot, but it stays cold. So its not a heat problem.
>
>The polyurethane used is non-conductive and shouldn't shrink, the
>potting was done by a company that specializes in potting electronics;
>they know about ESD precautions.
>
>Any ideas where to start looking?
Potting is better than air for heat conduction, so I doubt that's it.
Potting can tear parts from the board as it cures. Probably not
likely, though.
You mentioned a crystal. The potting compound will change the
dielectric constant which could alter the capacitance around the
crystal. This can also affect high-speed lines. If you have
fluorinert, you can try dunking the board and see if you can reproduce
the problem. Mineral oil will work, but is messy. If it is repeatable,
you can partially dunk the board to see which area is sensitive. I
have come across this sort of problem when operating electronics in
castor oil. We had a high-speed diff line that was marginal.
---
Mark
cs_po...@hotmail.com
> A sealed housing works too. Circuit boards I have designed operate in
> the ocean. Although it isn't hot like an engine, there are a few
> thousand pounds of pressure pushing the salt water in. With good
> design it stays on the outside and the PCB keeps working. Aluminum and
> O-rings work well.
Interesting. Do you get any leakage through wires?
A friend who was working on one of the university robot sub challenge
projects said that after they sealed around the insulation of wires
piercing the hull, they were still getting water leaking in around the
conductors inside the insulation though the length of the wires (I
believe they had a wetted drive motor outside the hull). There was
some talk of pumping the hull up with positive air pressure to
counteract this (depths involved were not too high) but that was also
undesirable - have to remember to do it, leave a trail of bubbles,
etc...
Robert Baer
> A customer of mine has a problem with a PCB potted in polyurethane.
> The circuit consists of some inputs (clamped & filtered), a
> microcontroller with internal RAM / flash and some power electronics
> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> worked fine. 4 of those PCBs where potted and they started to show
> erratic behaviour like RAM corruption, software hang-ups. I already
> milled one of the modules to expose the microcontroller to check
> whether it is hot, but it stays cold. So its not a heat problem.
>
> The polyurethane used is non-conductive and shouldn't shrink, the
> potting was done by a company that specializes in potting electronics;
> they know about ESD precautions.
>
> Any ideas where to start looking?
>
potting material.
Many polyurethane potting compounds "foam up" as they harden; that
foaming can cause a *lot* of static and so zap a lot of parts.
Robert Baer
> On Thu, 12 Feb 2009 21:10:20 GMT, ni...@puntnl.niks (Nico Coesel)
> wrote:
>
>
>>A customer of mine has a problem with a PCB potted in polyurethane.
>>The circuit consists of some inputs (clamped & filtered), a
>>microcontroller with internal RAM / flash and some power electronics
>>to switch loads (up to 40V). 10 PCBs where produced and tested. All
>>worked fine. 4 of those PCBs where potted and they started to show
>>erratic behaviour like RAM corruption, software hang-ups. I already
>>milled one of the modules to expose the microcontroller to check
>>whether it is hot, but it stays cold. So its not a heat problem.
>>
>>The polyurethane used is non-conductive and shouldn't shrink, the
>>potting was done by a company that specializes in potting electronics;
>>they know about ESD precautions.
>>
>>Any ideas where to start looking?
>
>
> Potting is always a horror. It's expensive, messy, often thermally
> horrible, can wreck parts, makes debugging almost impossible. Some
> potting compounds exotherm big time, or shrink and crush parts. Even
> when it seems like a good idea, it's usually not.
>
> John
>
Check, double check and even triple check!
MooseFET
> On Feb 13, 9:46 am, MooseFET <kensm...@rahul.net> wrote:
>
> > A sealed housing works too. Circuit boards I have designed operate in
> > the ocean. Although it isn't hot like an engine, there are a few
> > thousand pounds of pressure pushing the salt water in. With good
> > design it stays on the outside and the PCB keeps working. Aluminum and
> > O-rings work well.
>
> Interesting. Do you get any leakage through wires?
.. not if you use underwater connectors.
You can also use solid copper wire. Kink it a few times and solder a
small copper washer onto it and then pot the combination into a
roughened hole in the housing.
> A friend who was working on one of the university robot sub challenge
> projects said that after they sealed around the insulation of wires
> piercing the hull,
That doesn't work. Water comes down between the copper in the
insulation because it is a smooth interface.
MooseFET
> MooseFET <kensm...@rahul.net> wrote:
> >On Feb 12, 1:10=A0pm, n...@puntnl.niks (Nico Coesel) wrote:
> >> A customer of mine has a problem with a PCB potted in polyurethane.
> >> The circuit consists of some inputs (clamped & filtered), a
> >> microcontroller with internal RAM / flash and some power electronics
> >> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> >> worked fine. 4 of those PCBs where potted and they started to show
> >> erratic behaviour like RAM corruption, software hang-ups. I already
> >> milled one of the modules to expose the microcontroller to check
> >> whether it is hot, but it stays cold. So its not a heat problem.
>
> >> The polyurethane used is non-conductive and shouldn't shrink, the
> >> potting was done by a company that specializes in potting electronics;
> >> they know about ESD precautions.
>
> >> Any ideas where to start looking?
>
> >Was this a two part polyurethane that has big trouble if there is any
> >water about? I assume it is.
>
> Good question. I'll check that with the company that did the potting.
>
> >Never apply power to the board until the reaction is totally finished.
>
> >Did someone make the mistake of using "no clean" flux? If they did
> >abandon all hope of potting the boards in question. Make some more
> >with washed flux and wash them. Beware of the water from washing
> >them.
>
> Sounds reasonable. What is the problem with no-clean flux and potting?
The "no clean" stuff leaves a sort of plastic stuff that is supposed
to trap the salts etc from the actual flux. The potting compound eats
into this plastic and frees up the salts.
> >Is this polyurethane colored or is it the sort of clear stuff? The
> >sort of clear works better than the stuff with pigments.
>
> Its black. The stuff that makes it black is likely a very fine powder of a solid material. I am not sure about the details of how this leads to trouble but I think that it provides a surface that has uncured material on it long after most of the chemical reaction is done.
> >Is the failure heat related?
>
> I doubt it. I already checked that with the one I milled open.
>
> >What silly fool decided to pot them? Perhaps you can delete the
> >potting and et on with production.
>
> Potting it is an affordable way to get a slick looking custom housing.
"affordable" may not turn out to be true in this case.
WW
"Nico Coesel" <ni...@puntnl.niks> wrote in message
news:49948dda....@news.planet.nl...
>
> The circuit consists of some inputs (clamped & filtered), a
> microcontroller with internal RAM / flash and some power electronics
> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> worked fine. 4 of those PCBs where potted and they started to show
> erratic behaviour like RAM corruption, software hang-ups. I already
> milled one of the modules to expose the microcontroller to check
> whether it is hot, but it stays cold. So its not a heat problem.
>
> The polyurethane used is non-conductive and shouldn't shrink, the
> potting was done by a company that specializes in potting electronics;
> they know about ESD precautions.
>
> Any ideas where to start looking?
>
> Failure does not prove something is impossible, failure simply
> indicates you are not using the right tools...
> "If it doesn't fit, use a bigger hammer!"
> --Nico How about a sealed "bubble" like a lot of retail items are encased
> in. The ones that are a bitch to open. WW
------------------------------------------------------------
John Larkin
He has lots of names. Archimedes'Boytoy. MiniProng. AlwaysWrong.
Scatman.
John
Martin Riddle
"John Larkin" <jjla...@highNOTlandTHIStechnologyPART.com> wrote in
message news:140ap4d623ufp029k...@4ax.com...
> On Thu, 12 Feb 2009 22:40:47 -0500, "Martin Riddle"
> <marti...@verizon.net> wrote:
>
>>
>>
>>"Nico Coesel" <ni...@puntnl.niks> wrote in message
>>news:49948dda....@news.planet.nl...
>>>
>>> A customer of mine has a problem with a PCB potted in polyurethane.
>>> The circuit consists of some inputs (clamped & filtered), a
>>> microcontroller with internal RAM / flash and some power electronics
>>> to switch loads (up to 40V). 10 PCBs where produced and tested. All
>>> worked fine. 4 of those PCBs where potted and they started to show
>>> erratic behaviour like RAM corruption, software hang-ups. I already
>>> milled one of the modules to expose the microcontroller to check
>>> whether it is hot, but it stays cold. So its not a heat problem.
>>>
>>> The polyurethane used is non-conductive and shouldn't shrink, the
>>> potting was done by a company that specializes in potting
>>> electronics;
>>> they know about ESD precautions.
>>>
>>> Any ideas where to start looking?
>>>
>>> --
>>
>>Boards were dirty or contaminated. The potting probably never fully
>>cured.
>>That’s really the first place to look.
>>
>>Cheers
>>
>>
>
> That's a good suggestion. We've received boards from outside
> assemblers that would drip water if you shake them.
>
> It's hard to find assemblers nowadays who use solvent wash. And
> sometimes they use dirty water, or don't completely dry the boards,
> especially under parts. Potting will lock in any wet crud, and
> electrolytic corrosion will eventually mess things up.
>
> John
>
It would be Interesting if a malfunctioning board could be x-rayed.
Possibly by a board house that is setup for BGA stuff. I'm not sure if
it would show anything.
Un cured potting may be the same density as the cured stuff.
Cutting and digging may be the only solution to find the problem.
Cheers
MooseFET
The foaming action is often the result of humidity or water. You are
best to work in a dry environment. If you use the bipack method of
mixing, and start with fairly pure materials and keep the water out,
foaming in not going to be an issue.
Jim Thompson
<jjla...@highNOTlandTHIStechnologyPART.com> wrote:
Mixed in with other "turd-rated"...
Author: makenoattempttoadjustyourset | itsasecretdummy | higgsfield |
archimedes' lever | onebi...@infiniteseries.org | valleygirl |
luvye...@likeiwouldgiveit.comeon | ultimatepatriot |
itissoeasyacavemancandoit | fatbytestard |
stickthatinyourpipeandsmokeit | the great attractor | massiveprong |
the last mimsy | bungalow bill | klipstein | tutamongus | spurious
response | falk willberg | dorothy with the red shoes | do i really
need to say
Robert Baer
> On Fri, 13 Feb 2009 13:05:16 -0800, Robert Baer <rober...@localnet.com>
> wrote:
>
>
>>Many polyurethane potting compounds "foam up" as they harden; that
>>foaming can cause a *lot* of static and so zap a lot of parts.
>
>
>
>
> If you experienced one that did, that was your experience with ONE
> material. Your blanket remark points to a lack of experience, since it
> is not true.
>
> Potting compounds are specifically made to be potted into locations
> using very low pressure to remove all air and chemical gasification
> bubbles. Even a full (near full) vacuum.
>
> It is not a good compound if it exhibits gas while curing. If it
> foams, it is a media that was specifically meant to foam... pretty much.
OK; i stand corrected; all 3 i tried were foam and since they were
for insulation purposes it would seem to be safe to say they were
specifically meant to foam.
Thems alls wuz i uzedd.
Robert Baer
Ultrasonics *might* show something, but i would guess that would not
be dependable.
John Larkin
<marti...@verizon.net> wrote:
Even uncured potting is not going to be conductive. Trapped water
will.
If he were to un-pot the thing (lots of x-acto knifing) and rip parts
off the board, he may be able to spot voids and green electrolytic
corrosion underneath.
Potting makes all this detective work much more difficult.
John
MooseFET
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
[....]
> Even uncured potting is not going to be conductive.
It often has ionic stuff in it that goes away when it hardens. I have
had potted assemblies made go bad by applying voltage before the
reaction is done.
z
> A customer of mine has a problem with a PCB potted in polyurethane.
> The circuit consists of some inputs (clamped & filtered), a
> microcontroller with internal RAM / flash and some power electronics
> to switch loads (up to 40V). 10 PCBs where produced and tested. All
> worked fine. 4 of those PCBs where potted and they started to show
> erratic behaviour like RAM corruption, software hang-ups. I already
> milled one of the modules to expose the microcontroller to check
> whether it is hot, but it stays cold. So its not a heat problem.
>
> The polyurethane used is non-conductive and shouldn't shrink, the
> potting was done by a company that specializes in potting electronics;
> they know about ESD precautions.
>
> Any ideas where to start looking?
>
> --
> indicates you are not using the right tools...
> "If it doesn't fit, use a bigger hammer!"
back in germanium days, i potted some circuits that got cooked from
the heat of the resin setting.......
Jim Thompson
<TheSlimeFr...@oozingacrossyourlivingroomfloor.org> wrote:
>On Fri, 13 Feb 2009 19:19:49 -0700, Jim Thompson
><To-Email-Use-Th...@My-Web-Site.com> wrote:
>
>> klipstein || falk willberg |
>> ...Jim Thompson
>
>
>
> klipstein wrong.
>
> falk willberg wrong.
>
> ThompsonTard wrong, as usual.
AlwaysWrong can't even read.
...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
Teacher's Unions Cause Global Warming
Nico Coesel
>On Feb 13, 9:38=A0am, n...@puntnl.niks (Nico Coesel) wrote:
>> MooseFET <kensm...@rahul.net> wrote:
>> >On Feb 12, 1:10=3DA0pm, n...@puntnl.niks (Nico Coesel) wrote:
>> >> A customer of mine has a problem with a PCB potted in polyurethane.
>> >> The circuit consists of some inputs (clamped & filtered), a
>> >> microcontroller with internal RAM / flash and some power electronics
>> >> to switch loads (up to 40V). 10 PCBs where produced and tested. All
>> >> worked fine. 4 of those PCBs where potted and they started to show
>> >> erratic behaviour like RAM corruption, software hang-ups. I already
>> >> milled one of the modules to expose the microcontroller to check
>> >> whether it is hot, but it stays cold. So its not a heat problem.
>>
>> >> The polyurethane used is non-conductive and shouldn't shrink, the
>> >> potting was done by a company that specializes in potting electronics;
>> >> they know about ESD precautions.
>>
>> >> Any ideas where to start looking?
>>
>> >Was this a two part polyurethane that has big trouble if there is any
>> >water about? =A0I assume it is.
>>
>> Good question. I'll check that with the company that did the potting.
>>
>> >Never apply power to the board until the reaction is totally finished.
>>
>> >Did someone make the mistake of using "no clean" flux? =A0If they did
>> >abandon all hope of potting the boards in question. =A0Make some more
>> >with washed flux and wash them. =A0Beware of the water from washing
>> >them.
>>
>> Sounds reasonable. What is the problem with no-clean flux and potting?
>
>The "no clean" stuff leaves a sort of plastic stuff that is supposed
>to trap the salts etc from the actual flux. The potting compound eats
>into this plastic and frees up the salts.
I'll have to ask whether the potters cleaned the PCBs before potting.
>> >Is this polyurethane colored or is it the sort of clear stuff? =A0The
>> >sort of clear works better than the stuff with pigments.
>>
>> Its black. The stuff that makes it black is likely a very fine powder of=
> a solid material. I am not sure about the details of how this leads to tr=
>ouble but I think that it provides a surface that has uncured material on i=
>t long after most of the chemical reaction is done.
I've milled/cracked one module open but the compound seems to be fully
cured.
>> >Is the failure heat related?
>>
>> I doubt it. I already checked that with the one I milled open.
>>
>> >What silly fool decided to pot them? =A0Perhaps you can delete the
>> >potting and et on with production.
>>
>> Potting it is an affordable way to get a slick looking custom housing.
>
>"affordable" may not turn out to be true in this case.
So far 4 modules have been potted. The customer has a different
product of which hundreds have been potted by the same firm without
any problems.
krw
<OneBi...@InfiniteSeries.Org> wrote:
>On Fri, 13 Feb 2009 07:15:49 -0600, krw <k...@att.zzzzzzzzz> wrote:
>
>>
>>DimBulb is a tech, certainly not a degreed engineer (you've noted
>>the phobia of all things math).
>
> You were given numbers, you retarded fucks. And yes, I am a degreed
>engineer.
From where?
> So fuck you too KiethKeithStainTard.
You're Roy's boy toy.
MooseFET
Cleaning the "no-clean" stuff is double trouble. You know that stuff
that kills all the birds in a 100 mile radius and rots the "O-
zone"[1]. Thats the stuff you have to use to get rid of the "no-
clean". If any hydrocarbon touches it, you are forced to use that
stuff.
[1] For some reason is it goes right past the L-zone, M-zone and N-
zone.
[...]
> I've milled/cracked one module open but the compound seems to be fully
> cured.
You can often tell by the smell.
Greegor
> Interesting. Do you get any leakage through wires?
>
> A friend who was working on one of the university robot sub challenge
> projects said that after they sealed around the insulation of wires
> piercing the hull, they were still getting water leaking in around the
> conductors inside the insulation though the length of the wires (I
> believe they had a wetted drive motor outside the hull). There was
> some talk of pumping the hull up with positive air pressure to
> counteract this (depths involved were not too high) but that was also
> undesirable - have to remember to do it, leave a trail of bubbles,
> etc...
Where was the water penetrating into the wires?
You seemed to imply that the plastic coating on the wires was porous
or the other end of the (tube) of insulation was exposed to water.
Greegor
were blatantly done to prevent reverse engineering but
also drove customers away because they were so
horribly unrepairable. Finding a circuit board with
a potted section generally irks me off about like finding
one where chips have all had their makers part numbers
all ground off or dissolved off.
The only potting I ever used was with silicone rubber RTV
compound.
Have you determined yet whether the compound is
altering the time base wave form?
If it's not, then the suggestion to use some kind
of dielectric oil and selectively dipping portions to
find the failure was a good one I think.
I would find some way to hang or mount the board
so various parts are in the dielectric oil and leave it
in that position for a few hours worth of tests before
dipping further in and testing a few more hours.
I don't have an engineering degree.
Could you confirm that the four failed potted
boards all failed in exactly the same way?
Nico Coesel
>Most of the potted circuits I've seen outside of automotive
>were blatantly done to prevent reverse engineering but
>also drove customers away because they were so
>horribly unrepairable. Finding a circuit board with
>a potted section generally irks me off about like finding
>one where chips have all had their makers part numbers
>all ground off or dissolved off.
>
>The only potting I ever used was with silicone rubber RTV
>compound.
>
>Have you determined yet whether the compound is
>altering the time base wave form?
Nope. I have to be at the customer's and they're closed during the
weekend.
>If it's not, then the suggestion to use some kind
>of dielectric oil and selectively dipping portions to
>find the failure was a good one I think.
>
>I would find some way to hang or mount the board
>so various parts are in the dielectric oil and leave it
>in that position for a few hours worth of tests before
>dipping further in and testing a few more hours.
>
>I don't have an engineering degree.
Well, a lot of people that have a degree can't engineer :-) The
answers to the questions 'do you have a soldering iron at home?' and
'do you build electronic devices?' tell a lot more.
>Could you confirm that the four failed potted
>boards all failed in exactly the same way?
It appears its always the microcontroller that behaves erratic.
I'll work on this tomorow.
- I'll check the waveform from the crystal and compare it with the
waveform on a non-potted module
- See if there are any open pins on the MCU (the ESD suggestion
someone made)
- Check the power supplies, reset, etc
- Have two more modules potted, one with all open pins connected to
ground.
- See where the software goes wrong (at one point or random). As
someone suggested it might be a race condition which surfaces due to
changes in timing.
RFI-EMI-GUY
--
Joe Leikhim K4SAT
"The RFI-EMI-GUY"©
"Use only Genuine Interocitor Parts" Tom Servo ;-P
Eeyore
typhon62 wrote:
> On Feb 12, 5:38 pm, Eeyore <rabbitsfriendsandrelati...@hotmail.com>
> wrote:
> >
> > I suggest they stop potting. What an ANCIENT idea.
> >
> > Graham
>
> I have a board that in in the engine compartment of a pickup truck.
> It's wet, dirty, and gets salt on it in the winter.
> What else besides potting would I use to protect the PCB and
> electronics?
>
> I hate potting, but I really don't see another way to keep all of the
> road crud off of my electronics...
Ever heard of enclosures and IP ratings ?
http://en.wikipedia.org/wiki/IP_Code
Graham
Eeyore
Phil Allison wrote:
> "Eeyore"
> John Larkin wrote:
> >>
> >> Potting is always a horror. It's expensive, messy, often thermally
> >> horrible, can wreck parts, makes debugging almost impossible. Some
> >> potting compounds exotherm big time, or shrink and crush parts. Even
> >> when it seems like a good idea, it's usually not.
> >
> > Couldn't agree more.
>
> ** Ever see inside a H-H " Dual 500 " power amp from the mid 70s ???
>
> The two drive PCBs ( 6 x 3 inches) were fitted into a plastic tray filled
> with hard black resin leaving only a bunch of pins poking out.
>
> Madness.
When it all turned sour and Harrison formed another company ( HIT ), I recall a
similarly potted PCB with the epoxy scraped off strategically to as to add a
small cap. Doubtless someone screwed up the compensation.
Graham
Nico Coesel
>Greegor <Gree...@gmail.com> wrote:
>
>>Most of the potted circuits I've seen outside of automotive
>>were blatantly done to prevent reverse engineering but
>>also drove customers away because they were so
>>horribly unrepairable. Finding a circuit board with
>>a potted section generally irks me off about like finding
>>one where chips have all had their makers part numbers
>>all ground off or dissolved off.
>>
>>The only potting I ever used was with silicone rubber RTV
>>compound.
>>
>>Have you determined yet whether the compound is
>>Could you confirm that the four failed potted
>>boards all failed in exactly the same way?
>
>It appears its always the microcontroller that behaves erratic.
>
>I'll work on this tomorow.
>- I'll check the waveform from the crystal and compare it with the
>waveform on a non-potted module
>- See if there are any open pins on the MCU (the ESD suggestion
>someone made)
>- Check the power supplies, reset, etc
>- Have two more modules potted, one with all open pins connected to
>ground.
>- See where the software goes wrong (at one point or random). As
>someone suggested it might be a race condition which surfaces due to
>changes in timing.
Waveform and powersupplies look okay (at least, I get the same
readings from a non-potted working unit). Tried to alter the software
slightly to add more debug but this resulted in a change of behaviour.
That sounds like the internal flash may be 'broken' at some point.
Depending on which codes sits there, the behaviour will change of
ofcourse. Two more potted modules are on their way. Meanwhile I'll try
to check the internal flash.
Nico Coesel
I loaded several test patterns into the flash without any errors (the
flash has ECC so actual errors may have hidden). After that I tried to
load the firmware again. The module has been running fine since then.
No more weird behaviour.
Frithiof Jensen
"John Larkin" <jjla...@highNOTlandTHIStechnologyPART.com> skrev i
meddelelsen news:ef4bp4h03kjnstgbu...@4ax.com...
> On Fri, 13 Feb 2009 09:24:35 -0600, Vladimir Vassilevsky
> <antispa...@hotmail.com> wrote:
> At startup, it reads the entire parts file into memory arrays, and
> indexes the BOMs and data folders, which takes a few seconds.
Ackkhhh - pfuii! Caching is Evil; roll-your-own complex memory structures
are worse still ;-)
I very much doubt that anything that needs just the most basic search
functionality will beat SQlite3 in performance and ease-of-programming -
SQLite databases reside in in memory too but gives you SQL queries, indexes,
e.t.c. on top of that. http://www.sqlite.org/
Anyway, it's just my opinion based on very little information on what you
really need to do so pardon for busting in like that.
> After
> that, it's blindingly fast. Most commercial packages seem to build on
> top of some hairy general-purpose database manager, and then do
> everything over the network, and get slow.
Most/Many commercial packages use an XML file as database, then cache
everything in memory via a DOM parser; then have to traverse the tree and
iterate over *every* sub node to do a parameter search, then it's programmed
in Java too to get that Glacial Speed Feel.
I often get the feeling that Java developers think that data structures
shall look like the SUN jdk class library ....
>
> A part number is of the form nnn-mmmm, where nnn is a category, like
> RES 0805 or some such. The mmmm is encoded somehow, like by resistance
> in this case.
>
>>
>>My biggest frustration with the systems like that in the past was
>>because of the hell of bureaucracy, inconvenience and associated
>>overhead. So people tend to fool the system and do the things in their
>>own old ways, which makes the system worseless...
>
> This system was designed by engineering and manufacturing people to do
> exactly what we want. It works, like its predecessor DOS version
> (which used 5-digit numbers that eventually got to be sort of random)
> because the people make it work, and follow the rules, which are all
> available in writing.
>
> We may decide to sell it. People wouldn't be buying just the software,
> but the whole set of procedures and concepts, which might be very
> healthy for some companies without that discipline. Some people really
> don't understand the whole thing: dash numbers, assemblies and
> subassemblies, revision control, ECOs, BOMs, parts numbering, all
> that.
>
> John
>
>
John Larkin
<frithio...@diespammerdie.jensen.tdcadsl.dk> wrote:
>
>"John Larkin" <jjla...@highNOTlandTHIStechnologyPART.com> skrev i
>meddelelsen news:ef4bp4h03kjnstgbu...@4ax.com...
>> On Fri, 13 Feb 2009 09:24:35 -0600, Vladimir Vassilevsky
>> <antispa...@hotmail.com> wrote:
>
>> At startup, it reads the entire parts file into memory arrays, and
>> indexes the BOMs and data folders, which takes a few seconds.
>
>Ackkhhh - pfuii! Caching is Evil; roll-your-own complex memory structures
>are worse still ;-)
It's not complex; just arrays of fixed-length records.
>
>I very much doubt that anything that needs just the most basic search
>functionality will beat SQlite3 in performance and ease-of-programming -
>SQLite databases reside in in memory too but gives you SQL queries, indexes,
>e.t.c. on top of that. http://www.sqlite.org/
Indexes need to be built and maintained, and can get snarled. We do
brute-force linear searches. CPUs ar so fast these days, it's
essentially instantaneous.
>
>Anyway, it's just my opinion based on very little information on what you
>really need to do so pardon for busting in like that.
Again, CPUs are so fast, and gigabytes of RAM so cheap, that brute
force is perfectly viable. 10,000 part records, at 640 bytes each,
fits not only in ram, it fits in cache!
John
Archimedes' Lever
<jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>It's not complex; just arrays of fixed-length records.
>
Solid State Hard drive. No memory arrays required.
It IS a memory array.
Archimedes' Lever
<jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>
>Again, CPUs are so fast, and gigabytes of RAM so cheap, that brute
>force is perfectly viable. 10,000 part records, at 640 bytes each,
>fits not only in ram, it fits in cache!
>
Your application is that time critical?
John Larkin
It's just a parts/inventory database thing. Doing a search (like,
"opamp and rrio") takes way under a second, which is nice. If it took
10 seconds, it wouldn't matter much. But the speed is free nowadays;
you have to really work at it (as Microsoft does) to burn up billions
of cpu cycles per second.
John
krw
Jees, are you really this stupid, DimBulb?
Archimedes' Lever
<jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>On Tue, 03 Mar 2009 17:29:43 -0800, Archimedes' Lever
><OneBi...@InfiniteSeries.Org> wrote:
>
>>On Tue, 03 Mar 2009 06:50:10 -0800, John Larkin
>><jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>>
>>>
>>>Again, CPUs are so fast, and gigabytes of RAM so cheap, that brute
>>>force is perfectly viable. 10,000 part records, at 640 bytes each,
>>>fits not only in ram, it fits in cache!
>>>
>> Your application is that time critical?
>
>It's just a parts/inventory database thing. Doing a search (like,
>"opamp and rrio") takes way under a second, which is nice. If it took
>10 seconds, it wouldn't matter much. But the speed is free nowadays;
>you have to really work at it (as Microsoft does) to burn up billions
>of cpu cycles per second.
>
>John
Solid State Hard Drive.
Archimedes' Lever
Instead of beating on your chest like a big jock pussy with no dick,
why don't you try declaring what it is that you are refuting, asswipe?
krw
<OneBi...@InfiniteSeries.Org> wrote:
>On Tue, 03 Mar 2009 19:48:47 -0600, krw <k...@att.bizzzzzzzzzzz> wrote:
>
>>On Tue, 03 Mar 2009 17:28:59 -0800, Archimedes' Lever
>><OneBi...@InfiniteSeries.Org> wrote:
>>
>>>On Tue, 03 Mar 2009 06:50:10 -0800, John Larkin
>>><jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>>>
>>>>It's not complex; just arrays of fixed-length records.
>>>>
>>> Solid State Hard drive. No memory arrays required.
>>>It IS a memory array.
>>
>>Jees, are you really this stupid, DimBulb?
>
>
> Instead of beating on your chest like a big jock pussy with no dick,
>why don't you try declaring what it is that you are refuting, asswipe?
Why don't you read the absolute nonsense you wrote and tell us,
DimBulb?
John Larkin
<OneBi...@InfiniteSeries.Org> wrote:
>On Tue, 03 Mar 2009 17:45:27 -0800, John Larkin
><jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>
>>On Tue, 03 Mar 2009 17:29:43 -0800, Archimedes' Lever
>><OneBi...@InfiniteSeries.Org> wrote:
>>
>>>On Tue, 03 Mar 2009 06:50:10 -0800, John Larkin
>>><jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>>>
>>>>
>>>>Again, CPUs are so fast, and gigabytes of RAM so cheap, that brute
>>>>force is perfectly viable. 10,000 part records, at 640 bytes each,
>>>>fits not only in ram, it fits in cache!
>>>>
>>> Your application is that time critical?
>>
>>It's just a parts/inventory database thing. Doing a search (like,
>>"opamp and rrio") takes way under a second, which is nice. If it took
>>10 seconds, it wouldn't matter much. But the speed is free nowadays;
>>you have to really work at it (as Microsoft does) to burn up billions
>>of cpu cycles per second.
>>
>>John
>
>
> Solid State Hard Drive.
Why? The problem with flash memory is write endurance: it wears out.
For a camera or a memory stick, it's OK. For a general purpose
computer, pounding the disk constantly, it's not such a good idea.
Besides, this application is small enough to run entirely in ram. It
only needs a drive at startup, for a few seconds.
John
Robert Baer
For real fun, have variable length records..
MooseFET
> John Larkin wrote:
> > On Tue, 3 Mar 2009 08:27:09 +0100, "Frithiof Jensen"
>
> >>"John Larkin" <jjlar...@highNOTlandTHIStechnologyPART.com> skrev i
> >>meddelelsennews:ef4bp4h03kjnstgbu...@4ax.com...
>
> >>>On Fri, 13 Feb 2009 09:24:35 -0600, Vladimir Vassilevsky
> >>><antispam_bo...@hotmail.com> wrote:
>
> >>>At startup, it reads the entire parts file into memory arrays, and
> >>>indexes the BOMs and data folders, which takes a few seconds.
>
> >>Ackkhhh - pfuii! Caching is Evil; roll-your-own complex memory structures
> >>are worse still ;-)
>
> > It's not complex; just arrays of fixed-length records.
>
> >>I very much doubt that anything that needs just the most basic search
> >>functionality will beat SQlite3 in performance and ease-of-programming -
> >>SQLite databases reside in in memory too but gives you SQL queries, indexes,
>
> > Indexes need to be built and maintained, and can get snarled. We do
> > brute-force linear searches. CPUs ar so fast these days, it's
> > essentially instantaneous.
>
> >>Anyway, it's just my opinion based on very little information on what you
> >>really need to do so pardon for busting in like that.
>
> > Again, CPUs are so fast, and gigabytes of RAM so cheap, that brute
> > force is perfectly viable. 10,000 part records, at 640 bytes each,
> > fits not only in ram, it fits in cache!
>
> > John
>
> For real fun, have variable length records..
Using arrays of pointers to the actual records can have a lot of
advantages when you need to do something like sort them.
If the database is not changing a lot, the records can also be mostly
pointers to the strings that describe the part and give the buying
info etc. Those strings can then be stored one right after another in
the "string space". You can then use the fast string search methods
to find the strings that match.
John Larkin
Programming shouldn't be fun. That's the big problem nowadays,
programmers having fun.
John
Jeroen Belleman
That one goes into the file of snappy sayings. I heartily agree.
Cheers,
Jeroen Belleman