ESR Meter confusion

But it's just soooooooooo much easier to do it in-circuit with an ESR meter. The manufacturer usually supplies a chart of expected values for various combinations of cap voltage and capacitance, but I rarely need to refer to it. If a cap has become clapped out over time, the readings are almost always *miles* out from what you would expect - and the meter I use flags that up immediately anyway.

Here's the one I use:

When testing some circuit that is live, monitor each side of a cap, in some cases you can do differential measurement with two scope probes, if the cap is grounded you do not need that, in fact even if it is not grounded you usually do not need that. I would not in very sensitive circuits use those big alligator clips and cables. Neither do you want to remove all SMDs to test those. Of course depending on what you have .. big swollen up electrolytics in power supplies, or basically most high ripple carrying electrolytics are suspect, but you do not want to solder them all out to measure, just look at the signal across it. Paralleling with a good one to see if the fault clears is a good test too.

Just last week I had to find a bad 100 nF cap in the lnb_reference thing I build, As it is pertinax the board bends, and the SMDs on it get squeezed or pulled on. Scope both ends and see what there, depending on frequency you should have some idea, a lot of RF voltage over a 100 nF cap at 25 MHz means something is wrong. I have no idea what the ESR of those caps is... Replaced it, OK.

Obviously! ;-) But seriously, the thought processes of designers vs. techs are very different, as everyone who's been here a while will know. (I'm neither by the way). Yes, I scope the power rails as a matter of course. I have 13 different varieties of oscilloscope, but a lot of best techs have NONE and don't actually need ANY. Different horses for different courses.

I've found the LCR-T4 to be quite accurate too. What it needs is an additional 0805 sized display which I could see under the microscope while I'm probing a component.

Cheers

A tech has the advantage of knowing that the thing worked once. A problem is usually a part failure, but could be an undiscovered design problem.

Testing the first article of a new design, it's the opposite: mostly design errors, seldom a bad part.

Testing a new design (this is s.e.d.) I don't expect a dried-out electrolytic cap. But if something isn't right, I need to understand it, not just fix it, because we expect to manufacture a lot of them with near zero turn-on problems. [1]

So an oscilloscope and a DVM are basic.

Very rarely indeed do I ever see a *truly* innovative idea here, but you got one right there. Can't fail!

Ah, well I can see you're very well acquainted with the Zen of electronic design. :-)

Install a CMOS camera and an LCD display. With a computer driving the display, you can insert the output of your DVM or whatever and display it in a PiP box on the same LCD screen as is commonly done with security camera displays:

Sigh. Take a binocular microscope. Put one eye into one eyepiece and point the other eyepiece at an LED DVM display. If the LED's are bright enough, you should see something displayed on the PCB. You will probably need to play with the eyepiece type, power, and LED display size. One of the ancient 0.3" high red LED's should be small enough to fit a few digits across the eyepiece.

Maybe take an LED projector and just point it at the PCB.

Or maybe find a small transparent LCD display installed into one eyepiece. Something like a microscope graticule but instead use a projector type LCD display. Making a round LCD display might be tricky, but not impossible.

Good idea, but as an implementation detail, easier to put an earbud-size speaker into the gizmo, and have a pleasant digitized voice reciting the readings...

Watching where the probes are placed in a full-both-eyes-squinting task.

Big difference.....

Fact is though, many electronic items are built by people who have little idea what they are doing or the mistakes they are making ( or naively copying ) - so it takes a damn good tech to spot what needs fixing or leave good enough alone.

Of course, repairs have to be done within time and cost constraints that few designers would ever tolerate.

.... Phil

king.

out power applied and taking no regard to the state of operation of the ite m.

idea what they are doing or the mistakes they are making ( or naively copy ing ) - so it takes a damn good tech to spot what needs fixing or leave goo d enough alone.

few designers would ever tolerate.

Very well said, Phil. :)

Many of the sick electrolytics I was dealing with were in the "hot" (rectif ied 240V mains) side of VCR and other switching power supplies. What's easi er: to power the VCR from a 1:1 isolation transformer and carefully poke ar ound with an oscilloscope looking at waveforms (very carefully so as not to short anything or get zapped and assuming that the power supply is even st ill functioning), or leave it disconnected from power, discharge the main f ilter cap, and safely check the state of the electrolytics with an ESR mete r which identifies the defective ones just as readily as looking at wavefor ms?

That's one of the main reasons I designed my old ESR meter in the first pla ce.

Bob

Not an individual one, and we never measure Hfe. It's is on the data sheet, and Vbe is predictable enough that it doesn't usually matter. Spice models work well. Designs intended for production have to be tolerant of possible component and temperature variations.

(Besides, I mostly use fets.)

There is a lot of bad design around, and electronics is complex enough that even good designers make mistakes. One thing that I have to continuously remind my test techs is that recurring problems are probably design errors, not part or production issues. Stop fixing the damned things and tell engineering about it.

VCR? Do any of them still work? We just took a big bag of old tapes to the dump. Consumer electronics isn't repaired much any more; if it fails, it's tossed. Some old test gear is worth repairing, but more modern stuff can't be repaired by free-lance techs; parts and schematics are unavailable.

You're talking about repair of old gear; I'm designing new gear, and I assume that a part meets its datasheet specs.

ESR is to us usually a loop stability issue in linear or switching regulators, where too much or too little can cause problems. We only measure it to verify the data sheets, and then only on a few parts. Most caps have a typ or max ESR spec, so it's interesting to check a few to see what they really are. I use a function generator and a scope, which shows me ESL as well.

Ceramic caps have radical C-V curves, so are best measured with a small AC signal on top of a variable DC bias. Ramp-type c-meters aren't good for that measurement.

ach.

working.

ithout power applied and taking no regard to the state of operation of the item.

ick.

tle idea what they are doing or the mistakes they are making ( or naively c opying ) - so it takes a damn good tech to spot what needs fixing or leave good enough alone.

at few designers would ever tolerate.

tified 240V mains) side of VCR and other switching power supplies. What's e asier: to power the VCR from a 1:1 isolation transformer and carefully poke around with an oscilloscope looking at waveforms (very carefully so as not to short anything or get zapped and assuming that the power supply is even still functioning), or leave it disconnected from power, discharge the mai n filter cap, and safely check the state of the electrolytics with an ESR m eter which identifies the defective ones just as readily as looking at wave forms?

place.

I've still got a working VHS VCR. It's good for dubbing old tapes onto DVDs .

I'm talking about the domestic electronics situation of the early 1990s whi ch drove me to design an ESR meter to speed up my repairs way back then. Ev erything has changed since then, obviously.

ESR meters are still very useful when dealing with circuitry which is inacc essible, dangerous to poke around in, or simply doesn't work when it's powe red up.

Hfe is a variable with up to a 10:1 range, matched ( ie for diff) pairs are sometimes needed.

Devices from different batches can vary significantly and from different makers vary alarmingly. Always nice to know what you have got.

.... Phil

I still have one, and last time I looked it worked. But all tapes have been converted to mpeg2 or Divx, and are now on DVD-R or blu-ray disks. I will keep the VCR and for $^$$$ convert your precious VCR tape to whatever you have as system then.

But chances are the VCR tapes last longer than some of those optical disks. So I keep those too. Have backups on harddisk too.

We often do that too, after a quick visual check looking for tombstoned parts and other obvious flaws. (You have to do that anyway, because tombstoned bypass caps can lead to flaky failures.)

Finding dry solder joints and so on can be time-consuming, and they can cause overstress to other parts when the board gets powered up. Plus (like you folks) our BOM cost is a smallish fraction of the price.

I have a set of Smart Tweezers that I use for looking at parts in-circuit and for testing e.g. random electrolytics from the drawer that I'm putting in a proto. Not cheap but pretty good. That's about the only time I need to test ESR.

Re: beta

Beta distributions for a given type are sometimes annoyingly tight--I can't get plain BCX71s, for instance, and the suffixes are a moving target--they're al the way up to K now. Design in a BCX71G, and you find you can't get them anymore. Even the Js are getting scarce.

Cheers

Phil Hobbs