ESR Quest

[oc disorder]

I have a discontinued Dick Smith K-7204 kit designed by Bob Parker with a custom-programmed Z86E0412 microcontroller chip.

It also doubles as a low ohms meter useful for finding shorts on a pcb , shorted decoupling caps etc.

It is useful on occasion as it can be used without removing the capacitor from circuit although it must be discharged.

Two links , Bobs home page and another on the recent build.

ESR meter page

Capacitor Lab - Bob Parker ESR Low Ohms Meter from EVB Portugal

[nevetslab]

Update on ESR Test Set-up using Square Wave source & scope

First, when I had set my Function Generator for 10V P-P no load, and then you connect across a capacitor, it's nearly a short circuit, so that 10V open circuit becomes the equiv of 5V loaded by 50 ohms, that being the internal source impedance. So the effective current source is 100mA, not 200mA as I had stated.

I also gave thought to rated charge current on capacitors, and not having a data sheet on those 1985 vintage Mallory caps, I looked at what the rated charging current was on Nichion's VX Axial lead 10uF/350V parts. 85mA at LF, and at 1khz, a correction factor of 1.4, so that increases to 119mA. These being older parts, I'd guess 100mA current source might be a bit much. Also, IN-CIRCUIT testing, one would no doubt encounter semiconductors, where caps being checked are across the power supply buses where IC's and other active circuits reside, so a test voltage of 400mV P-P, maybe even 1V P-P could be used, as loaded down by the cap under test, you'd be about half that potential. I tried 400mV, and selected the most linear 10uF/350V part, to see what the difference was, and to see what the measured capacitance and ESR was at both test currents.



Taking the first image, using the two formula's from the Meettechniek paper, C = (I x t)/Vc, where I is 100mA, t is 100uS and Vc is the magnitude of the voltage ramp...in this case, call it 750mV.

C = (100mA x 100uS)/750mV = 13.3uF. I measured 14uF @ 1kHz on my GenRad 1658 DigiBridge, so we're in the ballpark.

Next, ESR = Vr/(2x I), where Vr is the vertical rise magnitude on the step-ramp waveform....I'd call it 200mV.

ESR = 200mV/(2 x 100mA) = 1 ohm. I measured 1.47 ohms on the bridge @ 1kHz. While not close, we're still in the same decade. And, checking multiple parts and seeing variations in ESR & Capacitance change in the step-ramp waveform, it still holds validity for the application.

Checking at 400mV P-P, which turns into 200mV P-P loaded by it's source imp of 50 ohms under short or near short, our constant current is 200mV/50 ohms = 4 mA

Looking at the second scope photo, we find Vr about 7mV, and Vc to be about 32mV.

C = (4mA x 100uS)/32mV = 12.5uF, and ESR = 7mV/(2 x 4mA) = 875 mOhms. Both reasonably similar to the higher current computed values, and the waveform shape about the same, if I had changed the scope sensitivity to 20mV/Div.

So, with very similar results at two vastly different current drive levels, one could just operate at the lower current level. Now, the downside of the lower current level is the typical 5mV/Div max sensitivity of scopes. Some caps would need greater sensitivity to see the waveform, or higher drive current to keep it in range of the scope's sensivity All part of learning how to use the technique.

I also checked the results of offsetting the function generator so the bottom of the applied waveform to 0V, so no negative voltage is being applied to the cap. I got the same results as yesterday.

I still haven't yet checked this IN-CIRCUIT, so that's coming soon.

[trem]

I use my home-made ESR meter all the time. It's my most-used piece of test gear after my DMM.....

Lots of people have good things to say about the Anatek unit, but there are plenty of schematics that can easily be built up and calibrated. Make sure you get a meter that can test charged caps, or circuits under power.

I am NOT very familiar with cap meters.
The only one that I KNOW of that measures a cap, with several hundred volts across it, is The Capacitor Wizard.
Do you (off the top of your head) know the name of any others.?
Thanks.....nice info btw.

[nevetslab]

General Radio 1617 Capacitance Bridge for one. You can source up to 600VDC from it to check caps for leakage, as well as see if there's any change in caps being checked with bias applied. The company that took over building GR bridges (IET Labs) may still build it, though it would be thousands $$ for a new one. You can download the manual from them....good general info within for a knowledge base on cap bridges and techniques. Four terminal input, so when needed you can use Kelvin leads. Measures up to 1 Farad, down to a few picofarads. ESI made/makes a number of bridges that will tolerate external bias, as does my GenRad 1658. I don't recall how much external bias I can apply to my ESI 296 or my GenRad 1658 without checking the manuals. Th common GR 1650A or B Impedance bridge will accept external bias up to 600VDC....plenty of them on the used market, a little limited in max capacitance range (1100uF), but a solid work horse.

[Mick Bailey]

I am NOT very familiar with cap meters.
The only one that I KNOW of that measures a cap, with several hundred volts across it, is The Capacitor Wizard.
Do you (off the top of your head) know the name of any others.?
Thanks.....nice info btw.

The circuit I built AC-couples the ESR test voltage (100khz) through a low-value cap to block DC into the circuit. I chose to use a moving-coil meter, mainly so I could calibrate the meter in pencil and then transfer the markings to a printed scale. I also added a meter protection circuit.

I'll take a look at my old PC to see if the schematic is on there. Just did a quick search and couldn't find the particular circuit I used on the web.

Some ESR testers (such as the Peak capacitor analyzer) discharge caps before testing. I thought of adding a discharge function to my meter but never really needed it or got round to it.

[nickb]

I have one of those Anatek Blue ESR meters. It was the first time I'd built an electronic kit since I was kid. The last kit was a Sinclair pocket FM radio.

But I digress. It works very well for it's intended purpose of finding bad caps, especially if you have a few similar ones to compare to and it's worthwhile for that.

OTOH as a low R meter it's not very good as it's a two point probe so contact resistance is a big variable and dominates measurements.

[nevetslab]

IN-CIRCUIT Capacitor / ESR Testing....Dip Tant buss caps, HT buss caps

I still haven't yet checked this IN-CIRCUIT, so that's coming soon.

For IN-CIRCUIT testing, I set up the Square Wave source for 800mV P-P & level-shifted the waveform so the bottom of the wave is at 0V, just so there's no reverse-bias voltage applied. Easy enough to do, even though thus far, I haven't seen any difference in the test waveforms on caps tested. 800mV P-P, which when loaded by any good caps, the test voltage across them is quite low, after adjusting the Test Frequency to display a nice linear step-ramp waveform. AC-coupled, since the source IS Level Shifted, and the vertical sensitivity is 10 or 20mV/Div. So, no semiconductors are turned on (unless there's open-circuit caps or ceramic caps only).

I selected a PCB assembly out of a Wandel & Goltermann TMS VF-1 Telecommunications Test Set of mine that hasn't yet been fully restored. This particular board has both good, failing and failed 22uF Dip Tantalum caps on it, with 22 ohm decoupling resistors feed from the +/- 15V supply rails, many of which have burned, as I discovered while checking this ESR Test Method out. The burnt resistors are, of course, a dead giveaway that either shorted Dip Tantalums are present, or an IC has failed. But, the status of the remaining Dip Tantalums??? Let's have a look:



The first image is PCB assembly I pulled out of the instrument, and clipped the test leads onto one of the color-coded Dip Tantalum caps. I swept the Frequency a bit before settling in on 5khz as the test frequency, yielding a reasonably linear appearing waveform. The first waveform is of that cap clipped on, which looks nominal. Using the two formulas C = It/Vc, where I is the current of 8mA and t is the duration of the ramp portion of the step-ramp waveform, here 100uS, and Vc is the magnitude of that ramp, here roughly 36mV, so C = (8mA x 100uS)/36mV = 22.2uF, and ESR = Vr/(2 x I), where Vr looks to be 5mV, so ESR = 5mV/16mA = 313 mOhm.

The next waveform is of a different 22uF cap, it having much greater ESR, seeing it has 26mV vertical step, so ESR = 26mV/16mA = 1.63 ohms, more than 5 times higher. The last waveform is one that has shorted, or nearly shorted. The last image is the Scope & Function Generator, with a Frequency Counter attached. The generator signal has an in-line T adapter, sending a sample of the test signal to the scope input, and the test signal down a short cable with a BNC to EZ-Hook adapter for connecting across the cap under test.

So, this does lend some credibility to screening parts like these Dip Tantalums, spotting some parts that may be on their way to failure, and certainly verifying those that have already failed, causing the decoupling resistors to overheat and burn. as seen in one of the images on that PCB assembly.

Next, I had the main PCB assembly out of an Ampeg SVT4-Pro on the bench for service, and took a look at the 100uF/450V caps that are part of the power supply assembly mounted to one of the Power MosFET heat sinks.



Here, I first swept the frequency, selecting a lower test frequency of 1kHz, and had to increase the vertical sensitivity of the scope to get a suitable waveform. Both caps look the same, and computing their capacitance and ESR....the Vc magnatude looks to be 43mV, t = 500uS, so capacitance C = (8mA x 500uS)/43mV = 93uF. Reasonable for a 100uF marked cap. Vr looked to be about 7mV, so ESR = 7mV/(2 x 8mA) = 438 mOhms.

So, this does look like a useful & simple test set-up to scrutinize electrolytic and tantalum capacitors. I had also looked at some large value film caps....2.2uF & 10uF. There was no visible vertical step in the waveform..looked like a pure triangle.....which, if you ever look at the schematic of a function generator, you'll find a good quality film capacitor(s) in the integrator section turning the 50% duty cycle square wave into a triangle wave, and then with diode shaping networks, shape it into a sine wave, though with noticeable 'tits' at the peaks of the waveform (very visible at higher frequencies).

[nsubulysses]

Nevetslab I like how you are a scientist. Great posts.

[Mick Bailey]

Probe resistance can be a nuisance. With many analogue circuits the null control is on the front panel and its just a case of shorting the probes and zeroing the meter.

I spent a long time searching for my schematic but it's not on my old machine. In general though, an ESR meter is easy to build and in my view one of the most useful measuring instruments. Wish I'd got one 30 years ago. Can't understand why in this day of cheap 'do everything' DMMs that this function is missing.

As a low R meter the greatest drawback is temp stability, but I guess a lab-quality unit would be stable.

It shouldn't be too difficult to design an add-on circuit for a DMM and read off mV as ohms ESR.

[bsco]

Aside from the "working" ESR discussion going on here, it's important to note that an ESR meter is a good tool for finding caps that have begun to dry out or leak when leakage is not visible. Like the average tube tester can't tell you for certain a tube is good, an ESR meter can't tell you for certain a cap is good. It's still a useful tool for finding caps that are definitely drying out without having to remove them.

I am considering getting one but for now, I use a scope and a signal generator set at 100KHz....with a 50mv sine wave.......I find this works good.....it just takes a minute to set it up....but it probably would be nice to have one.....

[trem]

nsubulysses -
You raise a very good question. I believe there is a meter called The Capacitor Wizard that can test caps in circuit, and at High Voltage. I THINK it is about 250-300 bux.

But maybe Alan0354 (and others) raise the most relevant point.....in a Tube Guitar amp, how important is one of these meters really.?
You can parallel a filter cap with another one of like value, and see if your troubles go away...or just replace all the inexpensive by-pass caps, etc etc.
I guess that is why I have never bought one...guys with Much More knowledge/experience than myself usually put a pretty low priority on an ESR Meter.
I Dunno.........

I am NOT very familiar with cap meters.
The only one that I KNOW of that measures a cap, with several hundred volts across it, is The Capacitor Wizard.
Do you (off the top of your head) know the name of any others.?
Thanks.....nice info btw.

No edit feature at this late date (or is edit available at all.?)
Anyway.....I just wanted to revisit this discussion.
1. The Cap Wizard CANNOT handle high voltage. I was told they could back when this post went up, but I was just looking at their manual, and it says there CANNOT be any power applied to the cap. Sorry for my error, do not want to mislead anybody, especially an "innocent" person doing an Internet Search.
My Apologies.

2. How important is high voltage when measuring ESR.?
That is to say...a cap in a power supply, that normally sees 400 VDC in circuit...is it possible for it to test "good" with the power off, but them leak badly with real voltage applied.?
How much worse (if any) is a cap likely to be when it has rated voltage applied, compared to the same cap with no voltage across it.?
Thank You

[bsco]

actually, I picked up the Blue ESR meter back a couple of weeks ago...before that, I was using a scope and signal generator set to 100K....so far I think the meter is great....it sure beats desoldering caps that look good to check them.......at least this way I can check them all and if anything looks odd then I can desolder it to check it.....the scope set-up also worked good but was not very portable......I can throw the ESR meter in my tool kit and go....For the money, I think it is a great buy......