Random tools and EV related projects

General Discussion about large electric scooters and motorcycles and other things with no pedals.
User avatar
ElectricGod
100 kW
100 kW
Posts: 1434
Joined: Nov 01, 2015 5:24 pm

Milliohm meter testing

Post by ElectricGod » Jan 09, 2018 1:52 pm

Some time back on my Power Velocity controller thread I stated that I was seeing inconsistent results in AOT290 mosfets. People criticized me and basically said I didn't know what I was talking about. I attempted to get people to refute my tests with testing on their part and none of the critics would in any way present actual test results.

In the process of that discussion, I bought 3 devices. An inexpensive component tester that measures Rds, a 2 wire milli ohm meter and a 4 wire milli ohm meter. I'll state right up front that of course the 4 wire meter is going to do the best job of measuring very low resistances. That was never in question for me despite the many critics claiming I thought otherwise. I just wanted to prove if the 2 wire meter could do the job or not.

So then here's my results...

This is my 2 most recent component testers and a bunch of components that I had that were unknown. They both cost less than $30 and do a great job of identifying loads of 3 legged or 2 legged parts. I highly recommend getting one if you want to quickly determine if your part works or have no idea what something is and want to determine quickly what some component is. I like the tester on the left better than the one on the right. It has far more functionality and the screen is easier to read. The right unit however is capable of measuring Rds...down to about 5 milli ohms. This is not useful for measuring Rds in mosfets used in controllers. Either meter will tell you what kind of component you have and lots of other specs on that part. I use these testers when ever I get a new batch of mosfets. In seconds I know which mosfets are matched and which ones are not. I also find that commonly mosfets are out of spec. Both of these testers on the same component are within 1-2% accuracy of each other.

Image

I haven't gotten around to further mosfet tests yet, but several critics refused to accept that I could test new mosfets on my component testers and see results that were well outside of specs. I pointed out that I was testing on my component testers and that after 3 of these testers, they were quite consistent with their test results despite being 3 different designs and made in 2015, 2016 and 2017. None of that mattered. I "had" to be doing something wrong, frying them with static...whatever, but the parts were sacred and holy and always 100% in spec. LOL! OK...believe what you like. I've tested mosfets so I'm not confused on this subject.

I bought a $40 2 wire milli ohm meter. I wanted to see if it would do the job of measuring down to 1 milli ohm or not. It's spec's say it should...so why not? The meter I purchased is really for testing capacitor resistance, but I already knew it would measure the resistance of anything and it does. Of course the critics leapt on the meter like it was a plague to be extinguished. LOL! I knew it wasn't going to do as well as a 4 wire meter, but that wasn't the point. Was it good enough to measure mosfet Rds or a shunt? I have to say that no it's not, but if all you need is 10 milli ohm resolution then it's good enough for you. The original meter came with cheap alligator clips that would not consistently "grab" a component so they didn't measure low resistances consistently. I replaced the factory alligators with Kelvin clips and then connected both jaws together with some solder braid. The meter is much more consistent now, but it still doesn't reliably measure down into the sub 10 milli ohm range.

This is the meter testing a .005 ohm shunt. As you can see, the factory leads measure it as .012 ohms...just about 250% too high. LOL! If I giggle the leads, the meter would sometimes actually measure .005 ohms, but most of the time it would display something much higher than this .012 ohms.

Image

The kelvin clips arrived a few days ago so I upgraded the meter leads and now I don't have to wiggle and giggle to get consistent readings. Of course it still doesn't do sub 10 milli ohms very well, but it can. What do you want for a $40 2 wire meter?

Image
Image

I paid about $100 for this 4 wire resistance meter. They get lots more expensive than this! All it does is measure very low resistances. It too uses Kelvin clips, but each side of the jaws is a separate test lead. Testing a precision .005 ohm shunt gets me pretty close to what the part should be. No surprises there.

Image

If you want to test components to see generally if they work or not or to determine what it is, get a Chinese component tester. They work great for most 2 or 3 legged electronic part. Some exceptions should be noted. Quite a few 3 legged parts are actually an integrated circuit and NOT just a single component inside. Voltage regulators and halls are great examples of this. It is not uncommon for these testers to NOT be able to identify them since they are multiple components integrated together into a 3 legged package. Don't assume that because your part is displayed as Unknown or bad that it is bad. Go look up it's data sheet to be sure it's not a 3 legged IC. Things like mosfets, resistors, caps, transistors, diodes, LED's, inductors should all read correctly. If any of them come back as unknown or bad, then that's an actual bad part.

This is a mosfet I tested...looks good!

[img}http://i1055.photobucket.com/albums/s51 ... 20test.jpg[/img]

Why would someone want a meter that can only measure very low resistances and what is a milli ohm? A typical DMM measures down to about 1 ohm. That's enough resolution for checking your car speakers or for general continuity testing. What do you do when you want to measure lower resistances? Lets say you want to check some switch or relay contacts and you are concerned that they might be scorched. The $40 meter will test that. If you get anything above 100 milli ohms, those contacts are probably damaged. So then what is a milli ohm. Milli indicates 1 thousand or in this context 1/1000. So imagine your car speakers are 3 ohms each and you wanted measure the resistance of the speaker cable. Are they really "low ohm" cables? The $40 meter will tell you that close enough that it won't matter.

What if I want to test the resistance of a mosfet or a shunt? Now we are getting down into the very low milli ohm range. Modern mosfets such as the IRF4110, AOT290 and so on are used in motor controllers and BMS and are designed to have almost no resistance from source to drain (Rds). Achieving ever lower resistances across the switched junctions in a mosfet is something every mosfet manufacturer is seeking. For measuring this low of a resistance, you really need a 4 wire milli ohm meter. The 2 wire meters just can't do the job reliably. My tests show that's the case and I was already suspicious it was true, but hey test and now I know for sure. The next question is why do you care about the drain to source junction resistance? Resistance is loss and loss usually means heat in mosfets. The mosfets in your controller or BMS can only handle so much heat and then they die. Quite often they get too hot while still operating well under their current handling limits. Using mosfets that can handle more heat and also have very little Rds means they can run much closer to their current limits for more time than a mosfet that has a much higher Rds and less heat handling. Let's look at the the AOT290 vs the IRF4110. Both have similar Rds at 3.5 and 3.7 milli ohms, but the AOT290 can handle 500 watts of heat while the IRF4110 can handle 370 watts. The AOT290 can run hotter for longer than the IRF4110. Now lets talk about the AOT290 vs an IRF4115. The IRF4115 has 9.3 mill ohms Rds and 380 watts of heat dissipation. This is 2.6X more resistance than the AOT290. Resistance=loss=heat. The IRF4115 running at the same amperage and voltage as the AOT290 will get 2.6X hotter and also can't handle as much heat as the venerable AOT290. So low Rds keeps your mosfets running cool which means they can run harder and longer. The lower the junction resistance just means your controller won't overheat as easily as it would with higher Rds mosfets...and that is good for you!

So then why not just get low Rds mosfets and plug them into your controller. Great idea! Do it. But, I want to know if they are actually within manufacturer specs or not. That is where the component tester and 4 wire meter comes into play. Now I can measure all the mosfets specs and Rds for myself and then pick the mosfets that most closely match up. OR!!! I can buy loads of a specific mosfet and then find the ones with the lowest Rds and use them in my controller. Remember... resistance=loss=heat.
ElectricGod...it's a joke...don't take things so seriously folks!

XB-502 Moped:
https://endless-sphere.com/forums/viewt ... 0#p1222730

Currie scooter:
https://endless-sphere.com/forums/viewt ... 7#p1227407

User avatar
ElectricGod
100 kW
100 kW
Posts: 1434
Joined: Nov 01, 2015 5:24 pm

Mosfet testing

Post by ElectricGod » Jan 09, 2018 3:08 pm

I'm moving this over here and that helps keep the Power Velocity controller thread cleaner and on topic.

Supposedly A $30 component tester can't test a mosfet and get real results. Supposedly only the manufacturer of that part can test accurately and you can't. LOL...OK...believe what ever you want, but I know better. I have purchased several pieces of test equipment for this test. I'm going to test a bunch of mosfets and let the chips fall as they may.

I am testing 10 new of each mosfet: All mosfets are purchased from digikey, mouser or arrow except the ebay IRF4110's. Values in parenthesis are for identifying each mosfet of each type. They are in order of cost where number one is the cheapest mosfet.

If you know of another mosfet that has a low Rds, low on time, 100 volt and high current, high wattage, please post them. I'm willing to test other mosfets. If your suggestion is less than 375 watts, I'm going to ignore it if it doesn't also have a less than 3 milli ohm Rds.

1. Chinese IRF4110 purchased on ebay (C1 to C10)
2. IRF4110 (I1 to I10)
3. AOT290 (A1 to A10)
4. IXFP180N10T2 (X1 to X10)
5. CSD19536KCS (S1 to S10)

The tests consist of several things.
1. Put each mosfet labeled 1 to 10 in 2 component testers, record it's measured specs.
2. Set my bench PSU to Vt, put a .3 ohm resistance in series with the mosfet source and 15 volts on the drain. Measure the voltage drop across the resistor. Do I get 250uA across the resistor? If this matches what the component testers measure, that's proof the testers work as expected.
3. Set my PSU to Vgs(th) min/max and test Rds. This is just for curiosity. Do all mosfets of the same kind turn on the same?
4. Set the PSU to 10 volts at the mosfet gate and measure Rds.

A note about Vgs(th)...
I have seen this value vary widely in AOT290's. In my testing with component testers, this value can be way over specs. In testing the AOT290, I have seen a few mosfets at the minimum value of 2.9 volts, quite a few at or near the typical spec of 3.5 volts and many at or above the maximum of 4.1 volts. I've been told by the critics how this is not possible, but my testing shows what it shows.

For anyone who cares, here's a great PDF on mosfet testing.
https://www.vishay.com/docs/90715/an957.pdf

A few specs from the spec sheets. Sometimes they don't list all the values so this is what I have to work with.
IRF4110:
Rds: typical=3.7 milli ohms, max=4.5 milli ohms
Vgs(th): minimum=2.0v, maximum=4.0v

AOT290:
Rds: typical=2.5 milli ohms, max=3.2 milli ohms
Vgs(th): minimum=2.9v, typical=3.5v, maximum=4.1v

IXFP180N10T2:
Rds: max=6 milli ohms
Vgs(th): minimum=2.0v, maximum=4.0v

The results will be tabulated and presented here. As posted many times before, the results will be the results!
ElectricGod...it's a joke...don't take things so seriously folks!

XB-502 Moped:
https://endless-sphere.com/forums/viewt ... 0#p1222730

Currie scooter:
https://endless-sphere.com/forums/viewt ... 7#p1227407

User avatar
ElectricGod
100 kW
100 kW
Posts: 1434
Joined: Nov 01, 2015 5:24 pm

Re: Random tools and EV related projects

Post by ElectricGod » Jan 14, 2018 4:01 am

More LED work done...

I got these little adjustable BUCK boards on ebay for 50 cents each. They work perfectly for driving XM-L2's. I'll be adding them to the XM-L2 lights soon. They take up to 24 volts in and convert it down to as low as 1 volt.

https://www.ebay.com/itm/10pcs-Mini360- ... 2749.l2649

Image

Image

Image

This XM-L2 is driving from 3.5 volts. It's running on the 6S pack it's sitting on with the BUCK dropping that to 3.5 volts.

Image
ElectricGod...it's a joke...don't take things so seriously folks!

XB-502 Moped:
https://endless-sphere.com/forums/viewt ... 0#p1222730

Currie scooter:
https://endless-sphere.com/forums/viewt ... 7#p1227407

Post Reply