Smart Tweezers - ESR test?

Hand tools, soldering irons, scopes, multimeters. Talk about em HERE!

Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.
Locked
User avatar
hberg32
 
Posts: 52
Joined: Mon Apr 07, 2014 2:38 pm

Smart Tweezers - ESR test?

Post by hberg32 »

Hi - can the smart tweezers you carry for SMD components do an ESR test of a capacitor in circuit?

I done blowed up something on a board with all SMDs. This will be my first attempt at diagnosing and trying to repair such a board. Wish me luck...

Thanks,
Henry

User avatar
adafruit_support_mike
 
Posts: 67391
Joined: Thu Feb 11, 2010 2:51 pm

Re: Smart Tweezers - ESR test?

Post by adafruit_support_mike »

No, it doesn't.

Testing capacitors in-circuit is tricky because you have to know what else is connected to them. Debouncing capacitors are often connected in parallel, for instance, and the low ESR of a good one will mask high ESR in a bad one. A cap that's shorted internally can also look good to some ESR meters.

It's easier to get useful information from an oscilloscope. Do you happen to have access to one?

User avatar
hberg32
 
Posts: 52
Joined: Mon Apr 07, 2014 2:38 pm

Re: Smart Tweezers - ESR test?

Post by hberg32 »

I do indeed, it's the USB one you carry (the name escapes me). Would the test procedure be to set the scope to generate a square wave and measure the waveform to see if the capacitor is rounding the corners?

User avatar
adafruit_support_mike
 
Posts: 67391
Joined: Thu Feb 11, 2010 2:51 pm

Re: Smart Tweezers - ESR test?

Post by adafruit_support_mike »

Not exactly, but you're close.

You want a +/-200mV signal you can switch between 1kHz and 100kHz. The voltage is low enough to avoid forward biasing any semiconductors that might be in the current path, and the two different frequencies will tell you about the cap's impedance.

Connect a 1k resistor between the oscilloscope probe's tip and its GND clip, and attach the signal generator's negative probe to the scope probe's GND clip. Apply the signal to one side of the capacitor, and touch the scope probe to the other side of the cap.

At 1kHz, you should see the square wave turn into spikes that drop back to 0V. At 100kHz you should see a square wave with some slight peaking at the rising and falling edges. If you don't see anything, it means the cap has failed open. If you see a square wave at both frequencies, it means the cap has failed with a short.

Actually measuring ESR is a bit more involved.

Start by getting rid of the 1k resistor, then connect the scope probe's GND clip and the signal generator's negative probe to one side of the cap. Connect one end an inductor to the signal generator's positive lead, and the other end of the inductor to the free side of the capacitor. Touch the scope probe to the point where the inductor meets the cap.

The exact value of the inductor doesn't matter that much, but should be fairly large. Try for 1mH or so. One coil from a junk-box transformer would work.

Set the signal generator down to about 10Hz and feed signal through the inductor to the cap. The scope should see a high-frequency oscillation whose amplitude drops off over time. The dropoff rate will tell you the parallel DC resistance of the scope probe, the inductor's ESR, the cap's ESR, and any resistors connected to the cap.

Exact values aren't all that important. You can test the readings with a known-good cap that isn't connected to anything, then compare that waveform to the one you get for caps in the circuit.

The oscillation is caused by the inductor and capacitor passing energy back and forth. If both of them had zero ESR and infinite DC resistance to GND, the amplitude would never fall.. they'd keep passing the energy back and forth forever. Energy that passes through a resistor gets converted to heat though, and is lost from the oscillation.

The scope probe's DC resistance will be about 1M, and if that was the only resistance in the signal path, it would take 4-5 seconds for the oscillation with a 1uF capacitor to drop to a point where it was too small to measure.

Any other resistors between the scope probe and its GND will also make the oscillation damp out faster. If you see a signal dropping faster than you expect, measure the DC resistance between the scope probe's tip and its GND probe with a multimeter.

ESR in the inductor and capacitor will also convert energy to heat, making the amplitude of oscillation drop off faster. The higher the series resistance through the inductor and cap, the faster the oscillation will die out.

The frequency of oscillation will be controlled by the inductor and capacitor values: f=1/2pi sqrt(LC). For a 1mH inductor and a 1uF capacitor, it will be about 5kHz. You can use a known capacitor value to calculate the value of your inductor, then use that to find the value of unknown capacitors in the circuit.

The attenuation is controlled by the resistance and the inductance: a=R/2L, where the units are the time it takes for the amplitude to drop by 63% (to 1/e of its original value). You can use a collection of known-value caps to estimate the attenuation of the probe and inductor, then compare that to the attenuation you see when measuring caps in the circuit.

User avatar
hberg32
 
Posts: 52
Joined: Mon Apr 07, 2014 2:38 pm

Re: Smart Tweezers - ESR test?

Post by hberg32 »

Thank you for that detailed response! That would make a good HowTo post for others that need to rule out capacitor damage. I think mine are ok, so the damage to the circuit must lie somewhere else. Most of the rest of circuit consists of AND gates and comparators (lmv339) which look like they are easy to test (I'm printing one of these to help place pogo pins: https://www.thingiverse.com/thing:2318886).

One other question for you - I've damaged one of the through-hole pads and wondered if you knew of a decent one-off tool/procedure for putting in an eyelet or rivet? I can replace the trace from a nearby via with 30 gauge wire and it would be nice to put something solid in the hole for neatness, if nothing else. It doesn't really make sense to lay out for a rivet gun or punch tool to repair 1 hole. Do you think a rivet can be placed with a nail and a tack hammer?

Henry

User avatar
adafruit_support_mike
 
Posts: 67391
Joined: Thu Feb 11, 2010 2:51 pm

Re: Smart Tweezers - ESR test?

Post by adafruit_support_mike »

Rivets are pretty easy to do as one-off items.

Find a piece of copper wire that fits snugly in the hole and anneal it: hold it in a flame until it starts to become red hot, then dunk it in water. Copper is exactly the opposite of steel in that respect: quenching from high temperature makes it softer.

Find a piece of rod you can use below the PCB as an anvil, cut a chunk of annealed wire about 1/8" longer than the thickness of the PCB, and chamfer the edges of the hole so the copper has somewhere to go when it expands. Put the wire in the hole, set it on the anvil, and tap it down from above. Annealed copper is pretty soft, so it won't take much force.

The copper will work-harden as it changes shape, and should give you a good rivet.

If you can't find wire the right size, start with a larger diameter and draw it down. Start with a piece a few inches long, grip both ends with pliers, and pull. The wire will get longer, narrower, and almost perfectly straight. It will also work harden, so if it isn't narrow enough to fit the hole, anneal it and draw it out again. As long as you keep annealing the wire between pulls, you can take it down to almost any size.

User avatar
hberg32
 
Posts: 52
Joined: Mon Apr 07, 2014 2:38 pm

Re: Smart Tweezers - ESR test?

Post by hberg32 »

Ok, I never would have thought of that. Just to be sure I'm clear - I'd be filling the entire hole in the PCB with a solid piece of copper then drilling a new hole in it? Hmmm, I need to replace a bit of the trace and was dubious of how to secure it to the rivet in a way that wouldn't just fall apart while soldering. But with this technique if I put both the rivet wire and the trace wire through the hole before tapping it down I'm thinking I'd get a good physically secure attachment that solder can flow into as well.

Thanks!

User avatar
adafruit_support_mike
 
Posts: 67391
Joined: Thu Feb 11, 2010 2:51 pm

Re: Smart Tweezers - ESR test?

Post by adafruit_support_mike »

Do you need a through-hole for a component?

If not, don't bother trying to drill the rivet after you swage it. Just use 30-gauge wire to make connections from the top and bottom of the rivet to the nearest traces. Solder all the necessary connections, then put a small dot of hot-melt in the center of each wire to keep it from getting caught on anything.

If you do need a through-hole, don't bother with a rivet.. there's a better option. To avoid repetition I'll call the component's lead 'the lead' and a piece of 30-gauge wire running beside it 'the runner'.

Expand the hole if necessary until you can fit the lead and the runner through the hole at the same time. On the back of the board, fold both the lead and the runner over and solder them to the stub trace on that side. On the component side, loop the runner under the component to get one tight wrap of runner around the lead, then solder the runner to the stub trace on that side. Then go back to the hole and solder the runner and lead together on both sides of the board. Finish off with small dots of hot-melt between the solder joints on both sides of the board for strain relief.

The component will have good support on the back of the board from its own lead, and on the front of the board from the runner.

Locked
Please be positive and constructive with your questions and comments.

Return to “Tools Tools Tools”