Measuring high resistance, major difference using DE-5000 LCR meter vs multimeter.

Think electrolytic effects, unless the water is absolutely pure.

Joe Gwinn

Tap water typically has a conductivity of around 200uS/cm

You've got figure in the area of the electrolytes and their spacing. Two inches by five inches is 64.5 cm^2. If the resin wasn't there you'd see about 200R.

1.4MΩ sounds like holes in your resin, or perhaps an ionic hardening agent.

The multimeter may not measure up to 1.4MΩ., or might use a lower voltage to drive the current than the LCR meter does. Conductivity through water often isn't ohmic. and the resin could be even more non-linear.

The capacitances sound reasonable enough. Water is a frequency dependent dielectric - tap-water has enough ions in it that you see a Warburg impedance

and the frequency of measurement matters.

So only the epoxy dipped portions of the plates were in water? Could the DE-5000 have been autoranging and picked capacitance instead of resistance? You could try sanity checking the DE-5000 with a 1.4meg real resistor or a real 2.2nF capacitor? Try the DE-5000 with test leads polarity reversed?

piglet

I did think about that, but the measurements are immediate. There is no metal contact in the water. I just measured the meter drives, the DE-5000 has a 400mv DC output and my ohmmeter has 60mv DC output, FWIW.

Mikek

I doubt there are holes, if so my multimeter show see resistance it goes to 60MΩ. I'll try a 1.5V battery and a 10meg resistor in series with the plates.

Yes, I measured meter ohmmeter drives as in above post.

The capacitance didn't change much from 100 hz to 100,000 hz.

Could try measuring with the Brymen again but with a 1.5v cell in series

- see if the added test voltage makes a difference? Try the cell both ways round?

piglet

That seems reasonable to me. Tapwater has all sorts of dissolved salts in it that allow current to flow 10M would be the most that you could hope for and in hard water regions it could easily be 100k or less.

Try it again with decent quality deionised or distilled water of the sort sold for using in steam irons or topping up old style vented car batteries. A figure of merit for water purity is exactly how low you can make the conductivity - something that is limited by dissolution of CO2 from the air unless you use sub boiling point distillation under N2.

Distilled water in ordinary labs is often around pH5 which is quite conductive in its own right due to CO2 dissolving in it.

Yes, I made sure the connections did not touch the water. I even dipped the leads to see it go much lower. I did as suggested, tested a 1.5MΩ at 1.4917MΩ and a 2.7nF measures 2.7nF. After several tests with the numbers listed above in the morning, now I get 5.8MΩ on the DE-5000 and 17.6MΩ on the multimeter. I setup 1.5v battery to 10meg, to the plates back to the battery, I get 1V across the 10MΩ, giving mt 5MΩ across the plates. Mikek

I agree with all you said, but I have a double dipped layer of fiberglass resin epoxy on each plate. Even after I diped them, I continued pouring on more epoxy until it was thickening. I doubt I have any holes and If there was, I doubt even lass it would have hols in both plates. Mikek

I'm back to the original Resistance measurement with the DE-5000, I have a problem with the kelvin leads, again. They are a weak point in the meter.

More back ground, The original tests were made with, a wire (as one electrode) centered in a test tube and the test tube inserted in a aluminum tube as the other electrode. It ran fine, then my son built a stainless steel, circuitous flow vessel with every other plate connected and opposite polarity. When it didn't work he brought it home for my look. I found the uninsulated plates of his circuitous flow vessel with just a little water had low resistance, the power supply has 18MΩ series resistance, which worked fine with the test tube. But his circuitous flow vessel had only 19.4kΩ resistance and he wondered why the power supply smelled, the 1MΩ resistors in the power supply got hot! So, this is a test to see if we can epoxy coat the electrodes to save the circuitous flow vessel . So far, not looking good. While my son was working on his circuitous flow vessel, I built a glass vessel, 2" x 2" x 34" tall. I used aluminum tape on two opposite sides and made connection points to attach the High Voltage. When full of water it measures ~400pf and the resistance is above 200MΩ. i.e. I can't measure it. It is shown here; Hopefully this will at least work enough for testing the idea.

BTW, previously in my battery experiment, I measured 1 V across a 10MΩ resistor, I failed to take into account the meters own 10meg internal resistance.

Mikek

Try wetting it out with silicone oil (or even olive oil or candlewax).

My suspicion is that you do have enough holes in the epoxy coating for ion diffusion to take place. It is quite difficult to avoid having systematic errors when dealing with resistors in the >10^9 class.

Every fingerprint matters. Degreasing with very pure alcohol may help.

You may be right. Last night I cut two sheets of plexiglass and sandwiched an aluminum sheet between, leaving a connection tab, I sealed the edges with a silicone (Permatex Ultra grey). With this sandwiched plate, I get no current flow testing with the 1.5V battery, neither of my meters can measure the resistance (OL). I do measure about 10pf in air and 90pf in the water. A further test coating the original in silicone oil did reduce the current flow in half, leaving me to believe there is leak in the epoxy. The sandwich is a viable solution, just more work. My son was so disappointed when his first vessel didn't work, I said that is much of research, you're learning by your failures. Thanks for the input guys. Mikek

Your average 4.5 digit voltmeter with a 200-mV range and 10-Mohm shunt resistance is actually a picoammeter in disguise.

The last digit is 10 uV across 10 Mohm, which is 1 pA. So with that and a flashlight battery, you can easily measure hundreds of gigohms.

I suspect that your insulation is fine, but your epoxy is not fully cross-linked, and so has horrible soakage (aka dielectric absorption).

If so, you’ll see the charging current drop off more slowly than an exponential. You can probably fix that by baking at 100C for an hour.

You don’t have to worry about fingerprints at the 1-Gohn level, but you do have to worry about fingers!

Cheers

Phil Hobbs

The epoxied pieces are in the oven.

Yes, If there aren't failures and you don't learn from them you are not actually doing research. I've been through that loop many times. It can be very painful, but when things finally work out it is incredibly rewarding.

John

Yes, I think some of it is he put work into it, now he has to do more work, just being lazy.

I have baked the original fiberglass epoxy, it still had a an easily measured resistance in the water. Then my son find continuity in a few places on the edge of the aluminum sheets. I put 6 coats of "Gorgeous Pink" fingernail polish over the edges, it made no difference. The tests I made sandwiching the metal plate in plexiglass sealed with Permatex Ultra has worked, but the Permatex does not adhere very well to plexiglass, it's much better on glass. I also used heat to seal the edges of the plexiglass and that worked, but over a few hours I found it also had a leak. The next order will include an acrylic solvent, to see how well that works. So we have three, 12" x 12" by 0.035" stainless steel plates that need to be insulated, the insulation needs to standup to many types of oils and 90*C temperatures. Any thoughts on an epoxy or other ways to insulate the plates. Thanks, Mikek

I have used a paint product called POR15 that should work for you. See porproducts.com for their full line, and click on sales tools and look at the flyer for Rust Preventive

. This is an epoxy paint that adheres to rust and seals it in to stop further corrosion. It's water thin so easy to brush on in thin coats and is amazingly durable on car frames and suspension parts. They call it hammer tough and I've tested that and it just doesn't chip off :-). It's only weakness is UV light so it won't stand up to direct sunlight forever. They say it's heat resistant to 450F/232C. I know it's impervious to water, oil, gasoline, and brake fluid. They have another product, Top Coat, which is supposed to be similar but will stand up to UV/sunlight, but it's more expensive and I've never used it. That data sheet says nothing about temperature limits.

I've never tested the electrical conductivity beyond having to grind off a bare spot to get a ground connection on car wiring with the gloss black. They have clear but in pints or larger which gets pricey. They have the gloss black in 4 oz. cans but as a sixpack so double the price of a pint They do have a Stop Rust Kit that's half the cost of the sixpack with one 4 oz. can and some other things you really don't need. It's on Amazon, and local auto parts stores like autozone usually carry at least one size. Eastwood.com has the pints and larger but not the sixpack of 4 oz. cans, sigh.

Email me if you want more info, just delete the obvious from my From address.

Hi Carl, I received your message, I sent a couple of messages yesterday, I just got the bounces back today. I guess I'm not smart enough to un-munge your address. That's OK. I found the product, I gave my son the info and will let him make any decisions. Thanks, Mikek