waterproof test 1

[scda]

Hi,

We have a project to design an electronic meter device which needs to be waterproof IP67. The mechanical engineers have design the casing but we need to ensure the finished product is IP67. The device has a LCD display and the dimensions are about 220x95x35 mm .

We intend to have a waterproof test for each meter. Can anyone help with an idea of how can we check that after the test there is no water inside the meter?

If there is a lot of water we can see on the lens that is in front of the LCD display but how can we know if there are only a couple of water drops? We cannot open the meter!

Someone came with the idea to centrifuge the meter so that the water is forced to go on the lens. Another was to heat the meter until the lens become flushing (hope flushing is the right word ).

Have anyone done this?
Thanks,
Cristi

 

[btrueblood]

We test a certain number of units by immersing them in water and waiting for them to fail. Crude, but gives a good idea if the design is working.

If you are looking for a QC test to conduct on every unit, you may want to consider a helium leak test. Doing a leak test on a sealed unit can be problematic, if no provision for a test port is made, but there are some manufacturers who seal the test gas (helium) inside the unit and then "sniff" the exterior some time later with a He detector.

Finally, one way to concentrate moisture vapor in a closed space is to cycle the enclosure above and below the freezing point of water. If you can make one part of the case more thermally conductive, it will tend to reach the freezing point first, collecting any moisture vapor in the case as frost on its surface. As the air in the container dries out due to condensation of the moisture, trapped moisture elsewhere (on hopefully still warm surfaces) vaporizes and is transported to the cold plate. Cycling multiple times (hundreds if you can get away with it) helps to "pump" moisture from crevices. Scraping/collecting the frost from the cold plate periodically and weighing it can give you an estimate of the moisture vapor transmission rate through seals, if the test is continued for a long enough period to see an asymptotic behavior in the data suggesting equilibrium.

 

[drawoh]

scda,

I have not read the full IP standard. The NEMA standards allow a little water to get in. Why can't you just take the thing out of the water, dry it off, take off the cover and inspect it?

There are people who formally test these things and issue certificates.

--
JHG

 

[IRstuff]

Generally, doing a test that can fail, at that stage of the process, is bad economic design. If you run the test and never fail, then what does the test tell you? If you run the test and it fails a lot, you've now consumed gobs of manufacturing cost that is now down a sinkhole.

Balance all of that against the cost of running the test on each unit.

If you must do this test, then perhaps you might want to install a humidity sensor

http://www.ti.com/lit/ds/symlink/hdc1000.pdf

inside your box.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

Of course I can. I can do anything. I can do absolutely anything. I'm an expert!

 

[btrueblood]

"Generally, doing a test that can fail, at that stage of the process, is bad economic design."

At what stage are you referring to IRstuff? Passing a failure along at any point in the engineering process is potentially a game-over event.

I.e. not doing the test on production units could be even more harmful, especially if the unit goes on to be installed in a sensitive facility where failure can cost the customer 100x the cost of the meter or more.

Most certainly, testing in the development phase is most cost effective, but testing on a sample basis of the production line output is far better than no testing at all.

 

[1gibson]

Resistive humidity sensors are readily available, perhaps your board should incorporate one that is accessible in a diagnostic mode. Would provide confirmation of the factory test, and add value for the user. Can you support the additional cost?

 

[IRstuff]

"We intend to have a waterproof test for each meter."

Which implies a completed product. Finding and fixing failures at this point is only less expensive than having them fail in the field. Note that the OP stipulates "EACH" meter.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

Of course I can. I can do anything. I can do absolutely anything. I'm an expert!

 

[TheTick]

An alcohol rinse can get rid of a lot of moisture. Be careful about chemical compatibility of seals, etc.

Maybe include desiccant in the meter packaging?

Maybe just keep the meters in a controlled dry space for a period of time after testing?

 

[zeusfaber]

How about:

Attach a scrap of cobalt chloride paper inside the enclosure, so it will be visible through the meter window if you look from the right direction.

Flush the enclosure with dry gas immediately prior to sealing (always a good plan if you want the display not to mist up on the first cold day)

Pressure test as planned

Warm the case up a little to turn any leaked water to vapour that can be picked up by the indicator paper.

Peer anxiously through the window at the colour of the paper.

Might want a bit of experimentation - but cheap enough that you can probably afford to do that.

A.

 

[77JQX]

Can you weigh the meter before and after the test?

 

[3DDave]

Add a small dessicant pouch and a humidity indicator plug to the design. This will tell you if it has failed the test and let the customer know if there is a failure after installation.

Otherwise, get a precision balance and weigh the box before the test, then rinse with alcohol and dry and reweigh. Water is roughly a gram per cc.

 

[scda]

Hi Guys,
Thank you for your answers. IRstuff is right. But I am new in this company which is an owned company and the owner ( this may look odd ) is involving in things he does not have knowledge/abilities.
The design of these meters is bad, rejecting up to 1% of the production. In my former jobs I was used that manufacturing is counting parts/million fails and when we were getting close to 30-50 it was considered an issue. Here I am the new R&D manager and the plan is to change the actual situation, but this will take me more than 3 years. So I am in position of trying to make the things to go ahead until we start improving. I am not mechanical although I have a good understanding of it, but the mechanical lead that I have at the moment is weak. So not much space to play around.

All I can do at this moment is to ensure the product that goes out of the companies gates achieves it's specification.

<<
"IRstuff says:
Generally, doing a test that can fail, at that stage of the process, is bad economic design. If you run the test and never fail, then what does the test tell you? If you run the test and it fails a lot, you've now consumed gobs of manufacturing cost that is now down a sinkhole.
Balance all of that against the cost of running the test on each unit.
If you must do this test, then perhaps you might want to install a humidity sensor

http://www.ti.com/lit/ds/symlink/hdc1000.pdf

inside your box.
>>

 

[IRstuff]

I feel your pain. There was a company that used to have to weigh their missiles immediately after assembly, so that they could tell if the coolant gas leaked out, prior to a test. They eventually fixed the problem with the coolant tanks, but it was surreal for about a year.

Has anyone done the failure analysis to identify the root cause of the failures? Given what you've described, there's really no way to determine leakage without:
> disassembly
> some internal sensor
> some visible sensor, as described above

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

Of course I can. I can do anything. I can do absolutely anything. I'm an expert!

 

[asimpson]

Fogging on screen is always a dead give away of moisture inside instrument. If you chill the device from 20 C to 10 C fogging should be fairly obvoius and last for a good period of time so long as temperature dosnt rise. Not ideal but it may improve your confidance level

 

[dbill74]

How are you testing the meter? (Sorry not familiar with IP67 or it's requirements.)
If in the testing you are submerging the meter, why not simply look for bubbles?
No bubbles = no water getting in. Bubbles = water getting in and unit failure.

 

[Pud]

....and make sure the water being used is warmer than the equipment to ensure gas inside expands.

H

www.tynevalleyplastics.co.uk

It's ok to soar like an eagle, but weasels don't get sucked into jet engines.

 

[scda]

Hi Guys,
Thanks for your answers. My goal with this issue was to go back to the design table and work it out.

It seems that I have been able to convince the top management that is better to work for an improved design.

Actually there is a micro USB connector where we found leakage. The mechanical guys are now looking for some solutions.

Thanks,
Cristi

 

[3DDave]

check L-com:

http://www.l-com.com/usb-waterproof...e-assemblies-revision-20-compliant-ip67-rated