Reference for LCR or ESR meters

[BravoV]

How is silver mica performs compared to polystyrene in tempco and long term stability ?

[ivaylo]

I mailed the capacitors today.

Got mine. Thanks again for doing this!

[linux-works]

this sounds like a good idea.

I wonder if taking it further would be reasonable; mailing a decade box for cap or resistors and having -that- be given measured values for each of the steps.

I can see it could be time consuming, so maybe a fee for your time (at that point) would be fair.

its surprising that no one is selling sets of R, C, L components in a nice marked plastic tray, all ready to go.

[BravoV]

Bumping to thank The Electrician, these "reference" caps & inductors are definitely improving my confident on the cheap Cyrustek based LCR. Sorry, just had time to play with them recently, already received them for quite a while.

The Mastech MS5308 is proven quite accurate within it's specification ranging from 0.5% up to 5% , just can't complain too much since it cost me < $100.

Btw, the esr at the polycarbonate (red box) are quite off, others like polypropylene, polystyrene and the electrolytic are spot on, don't have any idea why is that. 

Again, thank you so much. 


Few results while I was testing at approx. 25 C room temp, 10 C hotter when they were profiled.

[The Electrician]

Added in edit:  I wouldn't have expected the inductors to be measured very accurately because the measured value depends on the level of the voltage applied to them and the temperature.

I also see that I made a mistake in the ESR of the little blue 26 mH inductor.  The ESRs should be in k ohms, not ohms.  In other words, the ESR at 1 kHz is not .109 ohms, but rather 109 ohms.     You could add a k on the tape.

The measurement of the capacitance of the red box cap looks to me to be about .1%, not .5% 

I think the problem with the red cap is that I forgot to mention that the ESR numbers are in milliohms, not ohms.   

Looking at your images, I see that at 1 kHz, you got .04 ohms for ESR, which is pretty close to the .057 ohms I got.

At 100 kHz, you got .200 ohms for ESR, whereas I got .0055 ohms.  I don't have a ready explanation for that except that polycarbonate caps have a very low ESR at high frequency for their capacitance.  It's a difficult measurement to make.  Why don't you try making the measurement at 100 kHz with your meter in Rs mode?

Here's an image of a frequency sweep showing Z and ESR for another of those red box caps:



You can see that the ESR gets quite low at high frequencies.

[fpliuzzi]

@The Electrician
Thanks for clarifying where the decimal points should be in your hand written ESR values for the small inductor that you kindly supplied me. Now I know why my ESR meter wouldn't display any measurement reading when I checked the inductor... the value wasn't 0.192 ohms, but actually 192 ohms (@100KHz) and my vintage Bob Parker ESR meter can unfortunately only display up to 99 ohms. Small mystery solved.

I too had a polycarbonate ESR measurement that was pretty far off. My ESR meter measured 60 milliohms (@100KHz) when the side of the part was marked 10 milliohm (I do not expect to see too many other polycarbonate capacitors cross my test bench).

Luckily, with the help of your supplied reference components, my AADE L/C MeterIIb and Tenma 72-960 LCR meter (BK Precision 828) gave me measurement results that I can have a fair bit of faith in. I must add that my original Bob Parker ESR meter also works very well indeed when used for its intended purpose when repairing equipment on my bench.

Regards

[BravoV]

Added in edit:  I wouldn't have expected the inductors to be measured very accurately because the measured value depends on the level of the voltage applied to them and the temperature.

Yeah, I was quite surprised how accurate it is on inductor, remember, it was at 25 C room temp, 10 C hotter.

The measurement of the capacitance of the red box cap looks to me to be about .1%, not .5% 

I meant the LCR meter's specification is rated only with 0.5 - 5% accuracy depends on the test frequency, and the red box cap result is at 0.15% which is pretty good considering it's price. 

At 100 kHz, you got .200 ohms for ESR, whereas I got .0055 ohms.  I don't have a ready explanation for that except that polycarbonate caps have a very low ESR at high frequency for their capacitance.  It's a difficult measurement to make.  Why don't you try making the measurement at 100 kHz with your meter in Rs mode?

At the 3rd and 4th pictures, the LCR meter was not measuring at 100 KHz, its only 100 Hz and the result is 0.2 Ohm vs yours at 0.395 Ohm, while at 1 KHz, its 40 mili Ohm vs yours at 57 mili Ohm.


If I have a chance today, I will post all measured results from all test frequencies for these cap & inductor references, already had all of them written in paper.

[The Electrician]

@The Electrician
Thanks for clarifying where the decimal points should be in your hand written ESR values for the small inductor that you kindly supplied me. Now I know why my ESR meter wouldn't display ANY measurement reading when I checked the inductor... the value wasn't 0.192 ohms, but actually 192 ohms (@100KHz) and my vintage Bob Parker ESR meter can unfortunately only display up to 99 ohms. Mystery solved.

Sorry about that. 

I too had a polycarbonate ESR measurement that was pretty far off. My ESR meter measured 60 milliohms (@100KHz) when the side of the part was marked 10 milliohm (I do not expect to see too many other polycarbonate capacitors cross my test bench).

Luckily, with the help of your supplied reference components, my AADE L/C MeterIIb and Tenma 72-960 LCR meter (BK Precision 828) gave me measurement results that I can have a fair bit of faith in. I must add that my original Bob Parker ESR meter also works very well indeed when used for its intended purpose when repairing equipment on my bench.

Regards

I'm going to check the red polycarbonate on a couple of other instruments to verify the result I got on the Wayne-Kerr.

BravoV, that's really great that the MS5308 is so accurate on capacitance.  I have gotten the impression that the Cyrustek chip set is quite capable.  It's amazing that you can buy a true LCR meter that goes up to 100 kHz for well under US$200 with that kind of performance.

[BravoV]

Here are the measurement results ...

The items :






At the Sanyo OSCON electrolytic cap, it failed at 100 KHz with OL (Over Load) error message, don't know why is that ?  <Edit: Fixed with resistance series mode>

Also at the capacitance error results is bigger than other plastic caps, I guess its mostly caused by the larger temp co ?

About these reference caps & inductors, apart from the electrolytic, I guess if they never used, and stored in dry & temperature stable storage, they must be good enough for next few 5 or even 10 years ?

Over all, looks like this el-cheapo LCR meter performs above its rated accuracy, again, many thanks. 

[Conrad Hoffman]

Good stuff! That confirms my suspicion that Oscon caps are a pretty good standard if you need a high value. IMO, it's also a good idea to add a precision resistor to a very low esr cap to check higher readings. For my bridges I do dissipation factor standards consisting of a silver-mica in series with a small metal film to get dial calibration points like .01, .05, .1 and such. Naturally such combinations are good only at the single frequency they're designed for.

Polystyrene is said to have a predictable temperature coefficient. Silver-mica maybe not as much, but I think GR took advantage of the combination to create near zero tempco reference caps by making up a total value with a certain percentage of polystyrene and the balance from mica.

There's a calculation utility on my site that will give you most of the different forms of impedance for whatever input value you might have, or at least many of 'em.

[dr.diesel]

I have an in-cal HM8118, 0.05% base, if it helps any US/Midwest folks.

[The Electrician]

Good stuff! That confirms my suspicion that Oscon caps are a pretty good standard if you need a high value.

Hi, Conrad,

    I put the film caps and the OSCON in the refrigerator, which I keep just a degree or so above 0 C, and them measured them quickly to get at idea of the actual tempco of the caps I sent out.  I was astounded by the low tempco of the OSCON.  Ordinary aluminum electrolytics are terrible.

   I have several of the classic GR standard caps that I got from Boeing surplus.  They still have the cal stickers on them from 40 years ago; they're well aged now.  I measured them about 10 years ago, and just recently they have hardly drifted at all.

[BravoV]

   I have several of the classic GR standard caps that I got from Boeing surplus.  They still have the cal stickers on them from 40 years ago; they're well aged now.  I measured them about 10 years ago, and just recently they have hardly drifted at all.

Wonder what type of dielectric cap they used inside ? Have you ever take a peek ?


Btw, fixed the measurement results and updated the previous spreadsheet, for oscon and other caps. Used series resistance mode which has better resolution at 1 mili Ohm than in cap mode, and measured all without problem. Below OSCON cap measured in Series Resistance Mode @ 100 KHz.

[BravoV]

While observing The Electrician's sweep chart above, I noticed at the circled area is not smooth, and has very rough results and full with crazy fluctuations. As my cheap LCR doesn't have the freq sweep feature and only with certain fixed frequency points like 100,120,1K,10K,100KHz.

Just want to ask you guys if this area has something to do with my experience, that it was so difficult to get a stable reading if it happened that the test frequency was in that region ?

I imagine a tiny drift or instability at the test frequency will change the measured value erratically. 

Correct me if I'm wrong.

[Jay_Diddy_B]

BravoV,

The challenge with this measurement is that the resistive portion is 4 orders of magnitude 10,000 times smaller the impedance at these frequencies.

Jay_Diddy_B

[Vgkid]

   I have several of the classic GR standard caps that I got from Boeing surplus.  They still have the cal stickers on them from 40 years ago; they're well aged now.  I measured them about 10 years ago, and just recently they have hardly drifted at all.

Wonder what type of dielectric cap they used inside ? Have you ever take a peek ?


Btw, fixed the measurement results and updated the previous spreadsheet, for oscon and other caps. Used series resistance mode which has better resolution at 1 mili Ohm than in cap mode, and measured all without problem. Below OSCON cap measured in Series Resistance Mode @ 100 KHz.

Most likely mica, as the really old 505 series used them, along with the 1409 series(though the larger values are polyprop).

[Conrad Hoffman]

Yes, the typical GR cap is mica. It's actually an assembled stack of big mica plates and (lead?) electrode foils crammed together by a big mechanical clamp. The cover of the housing is potted, but the cap is in air, not potted itself. A desiccant pack is included next to the cap. I believe the combo mica and polystyrene cap was used in the 1608 bridge. Not sure if they did that with any others, as the 1608 was the most accurate traditional bridge they (or anybody) made. They were bettered by the 1615 type that used a ratio transformer and nitrogen filled invar plate capacitors.

[The Electrician]

While observing The Electrician's sweep chart above, I noticed at the circled area is not smooth, and has very rough results and full with crazy fluctuations. As my cheap LCR doesn't have the freq sweep feature and only with certain fixed frequency points like 100,120,1K,10K,100KHz.

Just want to ask you guys if this area has something to do with my experience, that it was so difficult to get a stable reading if it happened that the test frequency was in that region ?

I imagine a tiny drift or instability at the test frequency will change the measured value erratically. 

Correct me if I'm wrong.

The noise in the curve you noted is due to the high quality (Q) of the polycarbonate capacitor.  The ESR is so much smaller than the reactance of the capacitor that the meter has difficulty making the measurement.  Increasing the applied voltage from .5 volt to 1.0 volt and changing the sweep speed to a slower setting allows a better measurement as can be seen in the following image:



I suspect this high Q of the capacitor is the reason your ESR measurements of this capacitor at low frequencies are so much different from mine; this is undoubtedly the reason for the instability of your measurement in that region.  The Wayne-Kerr is designed to be able to measure high Q components, whereas your meter is probably not as capable in that regard.

I am impressed by the overall performance of the MS5308; one can't complain about it very much.

[fpliuzzi]

Hello all,
Quite some time ago I sent my measurement results directly to The Electrician for the four reference components that he kindly mailed to me. I had to hold off supplying results in this thread because I was waiting for a new DER EE DE-5000 to be sent from Japan. I wanted to include its results, along with those of my Tenma 72-960 handheld LCR meter that I've been relying on for about 12 years.

I didn't include any measurement results here from my AADE L/C IIB inductance and capacitance meter (which I like) because its applied test frequency of approximately 10KHz to 750KHz isn't user adjustable and the measurement results won't fit neatly into this four test frequency reporting format.

Also, since BravoV had such a nice layout for his test results above, I just went with that.

On a related side note; when I first received the DE-5000, the displayed measurement results seemed to be a bit unstable (unlike the demonstration by robrenz in one of his videos). Opening the meter I found some poor solder wetting on a couple of the joints where the split banana jacks and component insertion slot blades attach to the PCB. There was a visible ridge in the solder. Adding a bit of flux and retouching the joints improved the stability of the meter's readings. I also slipped a piece of thin plastic into the bottom area of the 9V battery to force it a bit more toward the two battery compartment contacts. Those two contacts didn't seem to deflect backwards at all when the battery was installed without the added plastic shim (i.e., a small piece of an old plastic gift card).
 
Regards

[robrenz]

On a related side note; when I first received the DE-5000, the displayed measurement results seemed to be a bit unstable (unlike the demonstration by robrenz in one of his videos).

I am using kelvin connections that are separate all the way to the component leads in the video. They seem to perform a little more consistently than the standard crock clips.

[fpliuzzi]

On a related side note; when I first received the DE-5000, the displayed measurement results seemed to be a bit unstable (unlike the demonstration by robrenz in one of his videos).

I am using kelvin connections that are separate all the way to the component leads in the video. They seem to perform a little more consistently than the standard crock clips.

Hello robrenz. I was aware that your meter's impressive measurement repeatability was because of your very nicely made custom 4-wire mod to the crock accessory, but my DER DE-5000 was originally making some fairly big swings when I was measuring down to the lower milliohm range. I expected it to be closer to, but certainly not exactly like, the excellent results that you showed in your video.

My DER DE-5000 acts better now, but I'll attempt to do some version of your 4-wire crock accessory mod to gain as much stability and repeatability as I can. Your four mini-grabber version looks doable. Thanks for your thought provoking videos.

Regards

[BravoV]

Frank, I believe these Cyrustek based LCR meters will only give a stable reading when using 4 wires measurement mode.

My MS5308 results were rock solid without even flickering at the least significant digit once the 4 wires alligator clips clamped or bit firmly at the dut pins as shown at my previous photos.

All this time I thought DER DE-5000 came with standard 4 wires alligator clips.

[robrenz]

The standard crock clips are 4 wire but only to one leg of the crock clip. The crock clip is not a Kelvin type clip that has the two halves insulated from each other to keep the force and sense leads separate all the way to the component lead.

[fpliuzzi]

BravoV, robrenz,
I think I'll just substitute a pair of MUELLER BU-75K Kelvin Test Clips in place of the standard crock clips on the TL-21 accessory. This would probably be the quickest path to a more stable and repeatable measurement. As you've both stated, a true 4-wire measurement is the way to go.

Thanks for the feedback

www.ebay.com/itm/2-pcs-MUELLER-Electric-Kelvin-Test-Clip-10-amp-BU-75K-Blakc-NEW-/191109195279?pt=LH_DefaultDomain_0&hash=item2c7efee60f

[robrenz]

I have some of those clips and they are ok but they tend to slip off the leads rather easily. I would look for some that have actual serrations in the jaws that will grip a lead better.