Page 4 of 7

Posted: Mon Dec 29, 2008 3:53 pm
by Robert Reed
No hot glue tried yet.
Nice job Jwax! Will be watching for future tests with much interest. Also wondering if moisture absobtion would be present in the long term - like if the "caps" were outdoors for a month.

Posted: Mon Dec 29, 2008 4:05 pm
by MrAl
jwax wrote:I fabricated a "parallel-plate capacitor" using two 1" square copper plates, and two narrow (.062") strips of .040" thick glass as spacers on opposite ends of the copper plates. I filled the space in between with hot melt glue and let it cool.
Tested high voltage resistance with a Fluke 1550B HiPot Tester. Setup was the applied voltage ramps up at 100 volts/second to 1,000 VDC. There it holds for one minute, and then measures the resistance at 1000 volts applied. This hot-melt package read "Overrange", which is higher than a million megohms (One Teraohm). Very good insulation indeed!
I'm still researching what the manufacturers spec is for EVA bulk resistivity, but it is right up there.
Also put the assembly on a cheap capacitance meter and read 12 pfd. That translated into a dielectric constant of about 2. Not easy getting a good reading on that small of a cap however.
Next on the test table will be Silicone (GE Silicone II, 100% silicone, indoor/outdoor, clear), and some Loctite QuickSet 5 minute "stereo" type of two-component epoxy.
Hi John,

Fantastic, just what i was hoping for.
I'll be back a little later with some more comments.

Posted: Tue Dec 30, 2008 1:26 am
by MrAl
Hi again,

I had a feeling that the insulation properties were pretty good.
I've used it a few times for pseudo temporary electrical insulation
on a few things and never got a shock.

Just wondering...did you try pumping the voltage up to the full 5kv yet?

Posted: Tue Dec 30, 2008 2:28 am
by jwax
Sure, I can do that! Wiki says air breakdown is 33 KV/cm, which is 3400 volts for my .040" spacing before breakdown. Pashen's Law gives 4380 volts breakdown, so we'll see how high I can go in test voltage.

Posted: Wed Dec 31, 2008 2:30 pm
by jwax
Repeated the HiPot test on Silicone II, and Loctite 2-component epoxy, using the same setup of 1" square copper plates, separated by .040", with the test material sandwiched in between:
The epoxy indicated a resistance of 500 Giga-ohms at 1,000 volts, and 300 Gohms at 2,500 volts. (beginning of breakdown?) Arc over from plate-to-plate occurred at 2,700 volts.
The silicone surprised me as being a relatively poor insulator: With 1,000 volts applied, only 6 Gohms. 2,500 volts applied, still 6 Gohms, and arced over plate-to-plate at 3200 volts.
I should note here that silicone has an erroneous reputation as being water impermeable, and high dielectric. Water does indeed migrate into silicone, and probably why its resistance drops considerably.

Back to hot-melt glue (EVA)- applied ever increasing voltage to 3100 volts when arc-over plate-to-plate occurred, but still measured above one Teraohm.

Happy New Year! :grin:

Posted: Wed Dec 31, 2008 3:03 pm
by MrAl
Hi John,

Very interesting...thanks for doing those tests.

Just wondering, did you do Air too?
I was asking because looking at the numbers it looks like
we may be seeing the effect of arc over at the edges.
I cant tell for sure but maybe if you did the test in the dark
you would see the arc location.

Here are the results of your test in terms of kilovolt per inch:

Hot Melt Glue: 77.5kv per inch arc over point
Loctite 2 part Epoxy: 67.5kv per inch arc over point
Silicone II: 80kv per inch arc over point
Air would be: 76.2kv per inch arc over point

From this we can see they are very much the same.

They are all nice and high though.

Thanks again and Happy New Year to you too.

Posted: Wed Dec 31, 2008 3:36 pm
by jwax
Indeed the breakdown voltages limit the test voltage. And that Vbreakdown is determined by many factors, one of which is radius of curvature of the conductors. A tiny burr on a plates edge can dramatically reduce Vbreakdown.
(My plates were not perfect!)
The tests give us a relative insulation resistance at high voltage, but say little about dielectric breakdown of the materials themselves. In these tests, breakdown was all in the setup, not the materials.
Suffice to say that hotmelt glue is the material of choice for a high voltage high resistance potting material.

hot glue

Posted: Wed Dec 31, 2008 6:59 pm
by evahle
Good stuff jwax! I'm still working on 2 experiments. One with hot glue and the other with silicone. The silicone is taking about 48 hours to cure, because it's about 1" x 2" x 1/2" thick in a plastic mold. I decided that it would be better if we tested for several things instead of one.

I built 2 circuits on PC boards and placed one in hot glue, the other in silicone. That way we can test for corrosion on the LED leads, also Photocell leads, copper traces, solder joints and flux, to see if the glues will react to any of these materials. Also I'm testing to see how well the parts (trimmer pot, photocell, LED and transistor) will hold up to the heat or glue seeping into the parts.

The hot glue project is done and I have no signs of any corrosion. It will be mounted outside to see how it will hold up out there. It works fine on the work bench so far.

The other project with the silicone is not fairing too well. The corrosion is already evident on the solder joints but nothing else. So I wonder if the lead content, or the flux, is reacting to the silicone while it is curing. Maybe it will stop after it is cured?

Tomorrow, I'll remove the silicone project from the mold and take pictures. You should be able to see the interior somewhat. Then I'll mount them outdoors and apply +5v to them to turn them on.

These small circuits are nothing more than simple night lights, so I can see them operating from indoors. It's cold out there! Haha.

I'll keep everyone posted.
evahle :smile:

hot glue

Posted: Wed Dec 31, 2008 7:21 pm
by evahle
I forgot the schematic of the night light. I modified Forrest Mims fridge alarm to turn on at night, and used an LED instead of a buzzer. ... tGlue1.jpg

evahle :smile:Image

Posted: Wed Dec 31, 2008 10:05 pm
by MrAl
Hello again evahle,

I had suspected that something like that might happen with the
silicone. They make a brand that is made for electrical work but
i dont remember the name of it. RTV something.

BTW jwax, is that Silicone II the type that is paintable, and that's
why they call it Silicone II ??

Posted: Thu Jan 01, 2009 7:14 am
by jwax
Corrosion on leads from silicone is probably due to acetic acid- that sharp vinegar-like odor you get when using silicones.
I don't think this silicone can be painted MrAl- HomeDepot recommends an acrylic latex caulk with silicone if you need to paint it.

Posted: Thu Jan 01, 2009 8:50 am
by MrAl
Hi again John,

You mean it doesnt say "paintable" on the side of the tube?

Posted: Thu Jan 01, 2009 9:22 am
by Robert Reed
"I had suspected that something like that might happen with the silicone. They make a brand that is made for electrical work but i dont remember the name of it. RTV something. "

That brings back memories as some of the guys in research used it quite a bit. It was called RTV and made by General Electric. There was a clear type and a red type (which they preferred) and it had incredible resolution.

Posted: Thu Jan 01, 2009 12:03 pm
by jwax
Nothing about "paintable" on this silicone MrAl.
Here's a pic of the epoxy and the silicone tested:
[img][img] ... potest.jpg[/img][/img]

hot glue

Posted: Fri Jan 02, 2009 8:52 am
by evahle
These are the results of my experiments on "High-Temp Hot Glue" and "GE-I 100% clear Silicone", so far:
I created 2-NightLight circuits and placed each in a mold for each type of glue. The molds were approx. 1-1/2"wide x 2"long x 1/2"deep. The mold was cut from the bottom of an old plastic TUMs bottle for a 1"deep mold.

The silicone was purchased at Home-Depot in a tube and I used a caulk gun to dispense it into the 1st mold. The Hot glue was dispensed into the other mold using a standard High-Temp Glue Gun; however the gun started to cool after the mold was 3/4 full. I had to squeeze the trigger harder to finish filling the mold. The silicone didn't fully cure after 50 hrs. So I removed it anyway, which created a messy, but still functional bottom. A few hours later, it finally cured.

1) The high-temp hot glue, nor the silicone, affected the circuit operation.
2) The hot glue cured to the consistency of a small bar of soap.
3) The GE-I 100% clear silicone cured to the consistency of soft rubber.
4) Corrosion and oxidation occurred only with silicone, and only while curing.
5) No other corrosion was evident on copper traces, LED leads or photocell.
6) Corrosion only occurred to the solder joints, probably due to lead content.
7) If silicone is preferred, only use about 1/16" when covering solder joints.
8 ) Using hot glue on solder joints prior to silicone molding also works.
9) 1/2" of 100% silicone didn't fully cure after 50 hrs. at room temperature.
10) From jwax's tests we found that hot glue makes a good insulator.

My final tests will be to place these projects outdoors, where they will be exposed to ice, snow, rain, sleet, high winds and heat.
I will be able to monitor the night lights at anytime by just watching the LEDs glow. At some point I will bring them back in to examine them for any additional corrosion, oxidation or water penetration.

evahle :smile:

The following pictures show the results and descriptions:

SEE PG. 6-7 to see the photos.