Measuring very low ESR

[The Electrician]

A few decades ago Hewlett-Packard began selling a fairly high performance impedance analyzer, the 4192A.  This analyzer could measure component impedance over a frequency range of 5 Hz to 13 MHz, with a maximum resolution of 1 mHz--that's millihertz!

This analyzer used the standard technique of applying a sine wave of voltage (or current) to the component being analyzed and measuring the amplitude and phase of the resulting current (or voltage).  From this the real and imaginary parts of the impedance can be determined.  The ESR of a capacitor is just the real part of the impedance.  To get a good measurement it's necessary to use a proper fixture with the analyzer.  The fixture which was most convenient for measuring loose capacitors was the 16047A.  Here's a picture of this fixture:



The 16047C uses the 4 terminal technique.  There are 4 stainless steel spring loaded "blades" that grip a capacitor's leads for the measurement.  The spring pressure exerted on the "blades" is not great, and one of the irritating problems with the 16047A is that dirt, solder from the cap leads, and oxidation products would build up on the "blades", leading to erratic measurements.  The user had to constantly clean the fixture "blades".

Some years later, HP introduced another impedance analzyer, the 4194A.  This analyzer could make measurements up to 40 MHz.  At the same time, HP also made available a new fixture, the 16047C.  This fixture was much improved; it featured thick gold plated copper bars that grip the DUT (device under test).  There were two screw knobs that could be tightened to apply a large pressure to the leads of the capacitor being measured.  This was a great improvement over the 16047A:



The years rolled by and once again HP/Agilent introduced another impedance analyzer, the 4294A.  This analyzer could make measurements up to 110 MHz and also had better sensitivity and accuracy.  The fixtures on all three of these analyzers are connected to the analyzer by 4 BNC connectors.  The mechanics of this connection are not very rigid, and the 16047A and 16047C fixtures could be moved by just inserting a component (and in the case of the 16047C, by the act of loosening and tightening the two knobs).  This movement would result in poor repeatablity of measurements.

When the 4294A came out, the new fixture that it came with (the 16047E) was very similar to the 16047C, but with brackets on each side that screw into the front of the 4294A and hold the fixture tightly to the 4 BNCs.  This gave greatly improved repeatability:



The new specifications for the 4294A included better performance at milliohm impedances.  When using these analyzers, to get the most accurate measurements it's necessary to perform an open/short calibration after the instrument warms up and just before making measurments.  The "short" cal is performed by inserting a "shorting bar" into the fixture.  To get milliohm accuracy, the shorting bar must really be almost zero ohms.  It can't really be zero ohms, but for milliohm accuracy of measurements, the shorting bar should have a resistance substantially less than 1 milliohm.

The 16047E comes with a gold plated "shorting bar"; the bar can be seen in the picture above in the top middle of the fixture; it is held in place by a screw knob.  Here's a view with the shorting bar removed:



  This shorting bar is much smaller than the bar that came with the 16047A and 16047C.  HP realized that to get it to function as a "short" of less that 1 milliohm, it would be necessary for the contact resistance in the fixture to be very low.  Thus a small area of contact with the gold plated measuring contacts in the fixture would be required, and a high pressure must be exerted by tightening the two screw knobs (until it hurts your fingers when you've done dozens of times).  The shorting bar that came with the 4192A and 4194A was much larger than the little gold plated one that came with the 16047E, and it isn't possible to get a contact resistance as low as can be obtained with the 16047E gold plated one.

When I was first using the 4192A and 4194A at work, it was inevitable that the shorting bar that came with the instruments would get lost.  When this happened, we made a DIY replacement that was the same size and configuration as the official one that had come with the instrument.  We no longer have the official bar that came with the 16047A, but here's a picture of the DIY replacement 16047A shorting bar and also the small one that came with the 16047E:

[The Electrician]

Now on to the problem of measuring very low ESR.

I'm using a more recent impedance analyzer, the Hioki IM3570.  This analyzer can not only measure impedances and display them as a single frequency value, but it can also produce a graphic showing the impedance over a frequency range.  It has a high sensitivity mode for measuring modern polypropylene capacitors having very low ESR.

To start, here's a sweep showing the impedance of a moderate precision 5 milliohm shunt.  The analyzer is sweeping from 100 Hz to 1 MHz; the top of the image is 10 miliohms with 1 milliohm at the bottom.  The green curve is impedance magnitude (|Z|) and the yellow curve is ESR (Rs).  This shunt is made of a rather thick resistance alloy and is in the shape of a wire loop rather than a surface mount device.  The sweep goes to 1 MHz and the impedance of the shunt begins to rise at the higher frequencies.  But it's purpose is to test the open/short calibration of the analyzer.  The open/short calibration of the analyzer was done with the little gold plated shorting bar.  We see that the analyzer shows that the resistance of the shunt is very close to 5 milliohms, as it should be:



Next, having done the open/short calibration of the analyzer with the little gold shorting bar, let's measure the impedance of the DIY shorting bar.  Of course, any shorting bar should be measured as ideally zero ohms, but in reality, some very low resistance.  If we want to measure ESR in the low milliohm range, the shorting bar should look like a resistance of less than 1 milliohm in the fixture.  Here's a sweep of the DIY shorting bar itself.  I have used this bar to open/short calibrate various analyzers, and it is perfectly adequate when the impedances I want to measure are not in the few milliohm range:



Not good!  Apparently the DIY shorting bar has a resistance of about 2.3 milliohms when inserted in the fixture; this effective resistance isn't very repeatable, either.  Why isn't it nearly zero ohms?  It isn't gold plated, and it is considerably larger than the little gold plated one.  Because it's so large, even when I screw the knobs down as much as I can without breaking the fixture, I can't get enough contact pressure for the resistance to be less than 2.3 milliohms.

Now, what would happen if I performed the open/short calibration with the DIY shorting bar rather than the little gold plated one?  I've done the open/short calibration with the DIY shorting bar, and then done a sweep of the 5 milliohm shunt.  Here's the result:



Now the 5 milliohm shunt appears to have a resistance of about 2.7 milliohms.  This is much less than its known resistance of 5 milliohms.  This is just what we would expect.  If we do the open/short calibration for an analyzer using a shorting bar that has a resistance R ohms, then measure various impedances, they will read too low by R ohms.  In this case, the DIY shorting bar has a resistance of several milliohms, so the measured ESR of capacitors will be too low by several milliohms.

If we want to get an accurate measurement of ESRs in the few milliohm range, we must use a fixture that can apply a large contact pressure, and we must use a small, clean, shorting bar which will allow a small contact resistance.  I noticed this problem when several years ago I was trying to measure milliohm impedances with the 16047A fixture.  I tried some kluges to increase contact pressure, but the only viable solution was to use the 16047C fixture with its large contact pressure.

[The Electrician]

Now let's try measuring the ESR of a modern high quality metallized polypropylene capacitor.  I bought 5 ECW-F2475 from Digikey.  I calibrated the analyzer with the little gold plated shorting bar in the 16047C fixture.  Here is a sweep of the capacitor; the impedance is the green curve and the ESR is the yellow curve.  The top of the image is now 1000 ohms, and the bottom is 1 milliohm:



The ESR at 1 kHz is about 10 milliohms, at 10 kHz it's about 2.78 milliohms, and at 100 kHz, it's about 3.02 miliohms:.

Here'a a screen shot showing the parameters at a single frequency of 100 kHz:



Now, I've done the open/short calibration of the analyzer using the DIY shorting bar.  Remember that the DIY bar doesn't look like a resistance of less than 1 milliohm in the fixture.  I have changed the scale so the top of the image is now 100 ohms, and the bottom is 100 microohms.  Here's a sweep of the impedance and ESR.  The left part of the ESR curve is noisy because I forgot to use the slow sweep speed.  We see that the apparent ESR actually goes below 1 milliohm in the 30 to 40 kHz frequency range.



This is a false result due to having done the open/short calibration with the DIY shorting bar.

We see that it's necessary to use appropriate fixtures and shorting techniques to get an accurate result.

[The Electrician]

Next I measured another polypropylene capacitor that has been measured in another thread with a handheld LCR meter.

The capacitor used is a Panasonic ECW-F4105HL.  I bought 5 of these from Digikey.

After doing the open/short calibration with the little gold shorting bar, screwing down the two knobs in the 16047C fixture so hard that it hurts my fingers, here is a sweep of the capacitor.  Impedance in green, ESR in yellow, 1000 ohms at the top of the image and 1 milliohm at the bottom.  This is a 1 uF capacitor, whereas the previous one was 4.7 uF, so we should expect the ESR of this capacitor to be higher:



At 100 kHz, the ESR is about 4.4 milliohms.  Here are the parameters at a single frequency of 100 kHz.  The measured Q is 365:



I measured this capacitor with my DE-5000 and tried to use the trick of measuring the Q and calculating the ESR from that.  The DE-5000 couldn't measure the Q; the display showed OL. 

The measurement of the ESR or Q of this capacitor is not possible with the low-cost LCR meters on the market.

Metallized polypropylene capacitors in the few microfarad range of the sort made by Panasonic, Rifa, Wima, Vishay, etc., just don't have ESR as low as 1 milliohm.  Using the low cost LCR meters (of which I have some myself) may give an apparent reading of 1 milliohm, but this is bogus.  It takes careful procedure and bench instruments and fixtures to get a real result.

Here's what it takes to get a capacitor with 1 milliohm ESR: http://www.mouser.com/ds/2/427/mkp1848dcl-35623.pdf

Look at the first parts listed; you have to go to the 400 uF part to get a 1.5 milliohm ESR.

[The Electrician]

To show the quality of the Panasonic ECW-F4105HL, I did a sweep of all 5 that I bought and superimposed the sweeps.  Notice how the 5 sweeps are almost on top of each other:

[tigr]

Good job. For complete happiness, there is a lack of measurement of high-capacity ceramic capacitors. This is also the weak point of cheap LCR meters.
Ceramic capacitor 75uF.ESR<0,001Ohm.

[precaud]

Nice writeup. The IM3570 is a nice tool, indeed, though out of my budget. I have the same sort of functionality (at <=300kHz) using a Wayne Kerr 6425 and some software to supervise the data collection and plot the results. To reach into the MHz, I'm currently working on other methods of low-Z measurement using a network analyzer. The techniques are similar to those used to measure loop gain in DC-DC converters.

Your points about contact resistance and fixture pressure are spot on. The problem, of course, is repeatability. Good measurements can be made with the 16047A but require some fidgeting with each component insertion. And regular cleaning, as you noted. But that's true with any fixture, really.

Also worth mentioning is the usefulness of making a few milli-Ohm "standards" to use as a quick reality check. Chip-type current sense resistors are great for this. To lower their residual inductance, parallel a few of them up to get the desired value. That's what the HP four-terminal resistance standards use.

[Marco]

Won't DIY "standards" suffer the same problem as his shorting bar?

AFAICS the problem being that you work harden the copper and it stops conforming, and in the case of tin coating you won't crack the oxidation without deformation.

[Jay_Diddy_B]

Hi Electrician and the group,

Great write up !! Your data shows the difficulty of measuring low impedance devices.

Let me throw in a contribution.

HP16047A Test Fixture

I have an HP16047A that looks like it has seen very little use. It was new or nearly new when I got it a decade ago and only sees occasional use.

I made a copy of the Keithley 8610 low thermal 4W short.

First we will measure the 4W short with a Keithley 2001:



The Keithley 0.26m . This must be subtracted from all the following readings.

Short

My DIY short consists of a piece of 12 awg bare copper wire, bent into horseshoe:



I don't have the official short from HP.

I made a set of cables with BNC sockets and banana plugs so that I connect the 16047A fixture to the Keithley 2001:



I place my short in the fixture. This is the reading from the Keithley:




The meter reads 0.57m   When subtract 0.26m I get 0.31m . This is not too bad.

ESI Kelvin Clips

I have made a set of leads for the 4274A using ESI (Electro Scientific Industries) Kelvin clips, RG174 and BNC plugs.



If I test these, in a similar fashion to the 16047A fixture, I get:



This is 0.8m subtract 0.26m  I get 0.54m


If I measure a 5m 2 wire resistor, I get (apology for the poor pornography):



This is after doing a short calibration, with both clips on one side of the resistor.

All this supports the concept that you have to be aware of the fixture limitations when making measurements below 10m

Thanks again to the Electrician.

Regards,

Jay_Diddy_B

[The Electrician]

Jay_Diddy_B, having carefully done the open/short cal as shown above, what do you now get for the polypropylene 4.7 uF capacitor that previously gave you a result of 1 milliohm?

What is the nature of the 5 milliohm resistor?  Is it similar to one of these shunts: https://www.digikey.com/product-detail/en/bourns-inc/PWR4412-2SDR0050F/PWR4412-2SDR0050F-ND/2564443

[The Electrician]

Won't DIY "standards" suffer the same problem as his shorting bar?

AFAICS the problem being that you work harden the copper and it stops conforming, and in the case of tin coating you won't crack the oxidation without deformation.

The nice thing about gold is that it doesn't work harden or oxidize.

A person could make a really nice shorting bar out of one of these: http://www.ebay.com/itm/1-10-Gold-Wafer-Bar-24K-Scotiabank-Valcambi-Sealed-Assay-Low-AA-Serial-Canada-/272863873528?hash=item3f87f449f8:g:YjgAAOSwJoNZytrz

I think that if I were to improve the DIY shorting bar, I would make it smaller, mill both sides flat, polish and gold plate.

The 16047C and 16047E have heavy gold plated contact bars in the fixture, and the little shorting bar that came with the 16047E is gold plated.

[tigr]

My calibration gold shunt.

[nctnico]

IMHO a shortening bar is not the right way to do a calibration (or better put: establish a base line for the test fixture) when it comes to low resistances. Most of the errors come from offsets and noise and you just can't calibrate those away. I used to own the 4274A LCR meter as well and it just sucked at measuring below 50m Ohm. The cheap Chinese meter I have does a much better job although it doesn't have a very wide frequency range.

I think somewhere the measurement method HP is using is prone to errors hence the need for clamping the DUT very tight. In theory you measure the current through the DUT and the voltage across it. Poor contact shouldn't affect the sense wires because no current flows through them. The same is true for the exitation wires because they are (ideally) driven from a current source.

[tigr]

IMHO a shortening bar is not the right way to do a calibration (or better put: establish a base line for the test fixture) when it comes to low resistances.

Yes, I absolutely agree with you. It was necessary to make a special adaptation and replace the Chinese probes with high-quality Japanese ones. For radioamateurish tasks (calibration of ESR meters), it is enough.

[precaud]

Yeah, it would be nice to have gold-plated contacts. And gold-plate component leads, too 

7 or so years ago I experimented with various size/shape shorting "bars" in the 16047A and Wayne Kerr 1005 component fixture, using radial and axial components. I got the most consistent and repeatable results doing short cal with this .05" thick copper bar. It contacts a large percentage of the contact leafs on both fixtures. I do have to move the short around in both of them while watching the display for the lowest Z reading. Fidgety, but it works.

One of the things I really like about the 16047A is that it accepts parts with really short radial leads, very useful when cannibalizing parts from pcbs.

Worth noting is that the Wayne Kerr uses a different grounding scheme for the shields of the 4-term leads than HP does. They specifically state the Hpot and Lcur ones be left unconnected at the fixture. While HP insists all four be connected there.

All of these devices are operating in constant current mode on the lowest range(s). Not all of them state what the drive current is. The WK defaults to a beefy 100mA and I get more consistent results with it than the 427x HP's which use lower current drive.

I'm thinking something like a ZIF socket repurposed for components might be worth playing with. I don't see any 2-pin ZIF's being sold, though...   

[tigr]

and replace the Chinese probes with high-quality Japanese ones.

[Machlee]

I made a copy of the Keithley 8610 low thermal 4W short.

Could you explain how the jumper is made? What type of connector did you use and how are they connected?

[Jay_Diddy_B]

I made a copy of the Keithley 8610 low thermal 4W short.

Could you explain how the jumper is made? What type of connector did you use and how are they connected?

The construction is described in this message:

https://www.eevblog.com/forum/testgear/restoration-glory-of-keithley-2001-dmm/msg999334/#msg999334

I had the gold plated banana plugs in my junk box, but you buy similar parts from Digikey.

Think of the short as the 4 wire resistance measurement of track 20mm wide and 0.3mm long, and you will get the right idea.

Regards,

Jay_Diddy_B