Author Topic: Bad cap or bad measurement?  (Read 12741 times)

0 Members and 1 Guest are viewing this topic.

Offline SpidersWebTopic starter

  • Supporter
  • ****
  • Posts: 62
  • Country: nz
Bad cap or bad measurement?
« on: July 23, 2014, 11:08:09 pm »
So, I bought two sets of some 10uF 50V high current high frequency capacitors from two suppliers.

Xicom +/- 20% 10uF 50V NP - looks cheap <--- both read 10.4uF on my multimeter in both polarities.
NICHICON +/10% 10uF 50V NP - looks like a quality part <-- all FOUR of them read between 20.5 and 21uF (datasheet, part number, advert, and cap indicate 10uF)

So before I ask for a refund and make a  :o face - is it possible my meter is reading these caps wrong, or are they actually that far out of range?
I'm just wondering if in certain circumstances (e.g. testing at say 1V on a 50V cap, or low current on something designed for high current etc etc) if it's possible to get incorrect readings??

The Xicom went in circuit and is working 100%. I didn't bother trying the Nichicon's.
 

Offline jamesglanville

  • Contributor
  • Posts: 47
Re: Bad cap or bad measurement?
« Reply #1 on: July 23, 2014, 11:43:33 pm »
I may be entirely wrong on this, but I know some capacitors have a much lower capacitance at high voltages. Is it possible the nichicons are higher capacitance at low voltages to ensure they meet spec at higher voltages?
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5022
  • Country: ro
  • .
Re: Bad cap or bad measurement?
« Reply #2 on: July 23, 2014, 11:45:19 pm »
NP means non-polar, they don't have polarity.  Shouldn't matter how you connect the leads (and wouldn't matter normally anyway)

High current and high frequency  AND 10uF 50v don't go together. And NP capacitors are not really suited for high frequency...

If you're not sure of your multimeter, you can always double check with a resistor and a power supply.

Get a 5v or 12v or something below your capacitor's rating, put a 100k or something like that in series with the power supply and connect the capacitor, with the multimeter set on voltage and the leads connected on multimeter.. measure the time it takes to hit 2-3v on the capacitor. Then compare with other capacitors.

 

Offline SpidersWebTopic starter

  • Supporter
  • ****
  • Posts: 62
  • Country: nz
Re: Bad cap or bad measurement?
« Reply #3 on: July 24, 2014, 12:42:58 am »
NP means non-polar, they don't have polarity.  Shouldn't matter how you connect the leads (and wouldn't matter normally anyway)

High current and high frequency  AND 10uF 50v don't go together. And NP capacitors are not really suited for high frequency...
Yeah it's for an almost 30 year old CRT. Not a part you'll find in stock at  digikey/mouser/jaycar etc.
The 10uF Xicom caps (which are working fine in the circuit and measured 10.4uF) came from an Arcade Electronics repair company in the US.
The 10uF Nichion's that read 21uF came from a supplier on ebay.

Datasheet for Nichion: http://www.mouser.com/ds/2/293/e-ha-5630.pdf (the ones which measured 21uF but label says 10uF)

Quote
If you're not sure of your multimeter, you can always double check with a resistor and a power supply.

Get a 5v or 12v or something below your capacitor's rating, put a 100k or something like that in series with the power supply and connect the capacitor, with the multimeter set on voltage and the leads connected on multimeter.. measure the time it takes to hit 2-3v on the capacitor. Then compare with other capacitors.
Would this be different to using the multimeter? Because I'm confident in the meter, I'm just not sure if the test conditions suit this particular capacitor.
As in, was it common for some special use capacitors to show 2 x capacitance at low voltages - but act like a normal part in the conditions they were designed for?

I may be entirely wrong on this, but I know some capacitors have a much lower capacitance at high voltages. Is it possible the nichicons are higher capacitance at low voltages to ensure they meet spec at higher voltages?
That's what I was wondering - maybe the Nichion's were specially designed for this type of circuit and this is a side effect?? I'd hate to return them and find out that's the case.
 

Offline The Electrician

  • Frequent Contributor
  • **
  • Posts: 743
  • Country: us
Re: Bad cap or bad measurement?
« Reply #4 on: July 24, 2014, 08:09:45 am »
Try applying rated voltage to a Nichicon for several minutes, discharging, applying rated voltage in the other direction for several minutes, discharging and then measuring.
 

Offline Wh1sper

  • Regular Contributor
  • *
  • Posts: 151
  • Country: de
    • Zockertown
Re: Bad cap or bad measurement?
« Reply #5 on: July 24, 2014, 08:40:51 am »
Your Multimeter is what?
Standard Multimeters do an very simple Check with very low frequency while data sheet gives you +-10% for 120Hz
I think This is why you measure 21 µF I'm pretty sure the nichicon are the better ones to use.
Read the datasheet :"Horizontal Deflection Current Correction Use"
So they are guaranteed for your use case.
 

Offline KedasProbe

  • Frequent Contributor
  • **
  • Posts: 646
  • Country: be
Re: Bad cap or bad measurement?
« Reply #6 on: July 24, 2014, 08:51:22 am »
Depending on your circuit there is a good chance that more than 10µF is better.
Not everything that counts can be measured. Not everything that can be measured counts.
[W. Bruce Cameron]
 

Offline sfiber

  • Regular Contributor
  • *
  • Posts: 107
  • Country: tr
Re: Bad cap or bad measurement?
« Reply #7 on: July 24, 2014, 09:10:55 am »
you should analyze to datasheets.Maybe capacitors that you ve bought from NICHION is specified for difference voltage and frequency values and your multimeter cant supply it.Some capacitors are only specified for only little range of frequency and voltage.At other ranges,capacitance of it can be changed
 

Online wraper

  • Supporter
  • ****
  • Posts: 16860
  • Country: lv
Re: Bad cap or bad measurement?
« Reply #8 on: July 24, 2014, 09:26:10 am »
Who  are those suppliers? Not ebay by chance?
 

Offline KedasProbe

  • Frequent Contributor
  • **
  • Posts: 646
  • Country: be
Re: Bad cap or bad measurement?
« Reply #9 on: July 24, 2014, 09:38:12 am »
I was just thinking the model of a NP is two caps in series this means to get 10µF you need 2 of 20µF so my obvious reasoning is if one is shorted you get 20µF instead of 10µF.
I could be wrong about this I don't know the NP cap design, I just thought I should mention it, so someone with NP cap experience can comment on it.
Not everything that counts can be measured. Not everything that can be measured counts.
[W. Bruce Cameron]
 

Offline SpidersWebTopic starter

  • Supporter
  • ****
  • Posts: 62
  • Country: nz
Re: Bad cap or bad measurement?
« Reply #10 on: July 24, 2014, 08:32:24 pm »
Who  are those suppliers? Not ebay by chance?

For the Nichion - yep, ebay. That's kind of why I'm asking becuase I need to post feedback and/or request refund.
8600 trades, packs of 4, arrived in a sealed static bag etc - looked quite professional - and there is full refund with 'seller pays return shipping'.

I just don't want to be that ass who asks for a refund, says the product is bad, when it's my understanding that's lacking. Just wanted to double check with you guys.


Your Multimeter is what?
Standard Multimeters do an very simple Check with very low frequency while data sheet gives you +-10% for 120Hz
I think This is why you measure 21 µF I'm pretty sure the nichicon are the better ones to use.
Read the datasheet :"Horizontal Deflection Current Correction Use"
So they are guaranteed for your use case.
Yep that's the kind of thing I was wondering.
If being out by +100% at low freq/voltage is normal - I'll post positive feedback and keep them as special purpose spares.

I'm just trying to work out if this is normal, or if they're garbage (but I don't want to re-open that CRT, desolder, and install the part just for testing).
 

Online wraper

  • Supporter
  • ****
  • Posts: 16860
  • Country: lv
Re: Bad cap or bad measurement?
« Reply #11 on: July 24, 2014, 08:46:10 pm »
I would never buy capacitors on ebay, especially Japanese brands. Tear apart one of them to see if there isn't such a joke inside http://studioboytimes.tumblr.com/post/68565865754/a-fake-electrolytic
Anyway very likely it is fake.
 

Offline The Electrician

  • Frequent Contributor
  • **
  • Posts: 743
  • Country: us
Re: Bad cap or bad measurement?
« Reply #12 on: July 24, 2014, 11:11:50 pm »
For the Nichion - yep, ebay. That's kind of why I'm asking becuase I need to post feedback and/or request refund.
8600 trades, packs of 4, arrived in a sealed static bag etc - looked quite professional - and there is full refund with 'seller pays return shipping'.

Could you provide a link to that particular item on eBay?
 

Offline SpidersWebTopic starter

  • Supporter
  • ****
  • Posts: 62
  • Country: nz
Re: Bad cap or bad measurement?
« Reply #13 on: July 25, 2014, 01:30:07 am »
« Last Edit: July 25, 2014, 01:32:25 am by SpidersWeb »
 

Online wraper

  • Supporter
  • ****
  • Posts: 16860
  • Country: lv
Re: Bad cap or bad measurement?
« Reply #14 on: July 25, 2014, 09:40:39 am »
Can't find any mention of HF series (on ebay picture, NP is not series) in google. I searched by part number UHA1H100KHA which is given by seller, I found LINK
Not to mention that if you look it the datasheet how to decode part number, it shows that it is HA series part number.
Well, actually I can find HF series (on picture) if search for NP, on aliexpress  :palm: http://www.aliexpress.com/store/product/s-original-bags-nichicon-NP-bulky-electrolytic-capacitors-25v10uf-audio-Promise/527691_1618537992.html
That says a lot about their origin.
« Last Edit: July 25, 2014, 09:45:28 am by wraper »
 

Offline The Electrician

  • Frequent Contributor
  • **
  • Posts: 743
  • Country: us
Re: Bad cap or bad measurement?
« Reply #15 on: July 25, 2014, 10:10:36 am »
Check out this item the same guy has for sale:

http://www.ebay.com/itm/2N3525-RCA-TO-66-VINTAGE-5A-SCR-TRANSISTOR-NOS-FREE-SHIPPING-/121381010466?pt=LH_DefaultDomain_0&hash=item1c42df3c22

Seems a little pricey; am I missing something? 
 

Online wraper

  • Supporter
  • ****
  • Posts: 16860
  • Country: lv
Re: Bad cap or bad measurement?
« Reply #16 on: July 25, 2014, 10:17:05 am »
Check out this item the same guy has for sale:

http://www.ebay.com/itm/2N3525-RCA-TO-66-VINTAGE-5A-SCR-TRANSISTOR-NOS-FREE-SHIPPING-/121381010466?pt=LH_DefaultDomain_0&hash=item1c42df3c22

Seems a little pricey; am I missing something?
Likely just out of stock, so rise up price until in stock again.
 

Offline The Electrician

  • Frequent Contributor
  • **
  • Posts: 743
  • Country: us
Re: Bad cap or bad measurement?
« Reply #17 on: July 29, 2014, 06:02:16 am »
I ordered some of the caps from the eBay seller, made some measurements and here's what's going on.

Here is a sweep showing capacitance on a linear vertical scale with zero at the bottom and 20 uF at the top; the green curve is capacitance and yellow is ESR (which is shown on the visible log scale).  At low frequencies the capacitance is approaching 20 uF, but it decreases with increasing frequency, reaching about 10 uF at 100 kHz.  At 15.85 kHz the capacitance is 11.76 uF, and even closer to 10 uF at frequencies approaching 100 kHz.  Considering that this capacitor is designed to carry currents at the horizontal deflection frequency of a CRT video display, we might expect its capacitance to be the rated value at those frequencies and not at 120 Hz:



The reason the capacitance decreases so much as the frequency increases is because this capacitor is made with aluminum foil that has been etched to increase the capacitance for a given can size.

I discussed this here: https://www.eevblog.com/forum/testgear/mastech-ms5308-lcr-meter-with-esr-measurement-on-discount-at-the-moment/msg107489/#msg107489

This article: http://www.kemet.com/Lists/TechnicalArticles/Attachments/191/Why 47 uF capacitor drops to 37 uF- 30 uF- or lower.pdf

discusses the effect at length.

The black curve in Figure 4 of that article shows the effect.

My Fluke 187 DVM gives a capacitance measurement of 20.3 uF, and I would expect measurements made with DVMs to be similar.

Making the measurement with an LCR meter at higher frequencies will give a much lower value.  My DE-5000 at 100 kHz gives a value of 11.35 uF.

So, there's nothing wrong with the cap; its capacitance when used at the horizontal frequency of a CRT display is much closer to 10 uF than 20 uF.

« Last Edit: July 29, 2014, 06:16:45 am by The Electrician »
 

Online wraper

  • Supporter
  • ****
  • Posts: 16860
  • Country: lv
Re: Bad cap or bad measurement?
« Reply #18 on: July 29, 2014, 03:59:02 pm »
10 uf electrolitic capacitor rated at 100kHz? Give me a break, even this 1uF ceramic cap is rated at 1khz http://www.samsungsem.com/servlet/FileDownload?type=spec&file=CL10B105KO8NNNC.pdf
 

Offline The Electrician

  • Frequent Contributor
  • **
  • Posts: 743
  • Country: us
Re: Bad cap or bad measurement?
« Reply #19 on: July 29, 2014, 07:23:37 pm »
10 uf electrolitic capacitor rated at 100kHz? Give me a break, even this 1uF ceramic cap is rated at 1khz http://www.samsungsem.com/servlet/FileDownload?type=spec&file=CL10B105KO8NNNC.pdf

The spec sheet referenced in reply #3 says: "Designed specifically for horizontal deflection current correction of TV's where high
frequencies and high ripple currents are applied. "

Horizontal frequencies of CRT based monitors are most likely in the range 10 kHz to 100 kHz.
 

Online wraper

  • Supporter
  • ****
  • Posts: 16860
  • Country: lv
Re: Bad cap or bad measurement?
« Reply #20 on: July 29, 2014, 08:39:01 pm »
10 uf electrolitic capacitor rated at 100kHz? Give me a break, even this 1uF ceramic cap is rated at 1khz http://www.samsungsem.com/servlet/FileDownload?type=spec&file=CL10B105KO8NNNC.pdf

The spec sheet referenced in reply #3 says: "Designed specifically for horizontal deflection current correction of TV's where high
frequencies and high ripple currents are applied. "

Horizontal frequencies of CRT based monitors are most likely in the range 10 kHz to 100 kHz.
There is no datasheet in post #3 but is datasheet in post #4 for HA which is different series, HF (OP's one) just don't exist in the nature (except aliexpress). But even if we are taking HA series, yes it says:
Quote
Designed specifically for horizontal deflection current correction of TV's where high
frequencies and high ripple currents are applied.
However if you would read datasheet a little bit more carefully:
Quote
Capacitance Tolerance ±10% at 120Hz, 20°C
 

Offline The Electrician

  • Frequent Contributor
  • **
  • Posts: 743
  • Country: us
Re: Bad cap or bad measurement?
« Reply #21 on: July 29, 2014, 10:07:23 pm »
There is no datasheet in post #3 but is datasheet in post #4 for HA which is different series, HF (OP's one) just don't exist in the nature (except aliexpress). But even if we are taking HA series, yes it says: "Designed specifically for horizontal deflection current correction of TV's where high frequencies and high ripple currents are applied. "

I didn't say that the spec sheet was referenced in post #3; I said it was referenced in reply #3.  Look just below the subject line of each post after the first--the reply # is there.

The OP seemed to think that the linked spec sheet might be relevant to the application for this cap even though it isn't the spec for the HF series.  He says the cap goes in a 30 year old CRT.  Apparently the HF series is obsolete, and probably was used to carry currents at the horizontal frequency of a CRT; the designation HF might have stood for "high frequency".  I searched for a spec for a Nichicon HF series, but couldn't find one.

Other evidence that this cap is intended to carry currents at frequencies substantially higher than 120 Hz is its physical size and its low ESR.

A typical 10 uF, 50V electrolytic is nowhere near this big, and the ESR for this cap is around .1 ohm, more or less, in the vicinity of 10s of kilohertz.  If the cap is carrying 3 amps, for example, the dissipation in the cap would be around 1 watt.  A more typical 10 uF electrolytic has an ESR of several ohms.

The large physical size of this cap allows it to dissipate 1 watt without any problem.


However if you would read datasheet a little bit more carefully:
"Capacitance Tolerance ±10% at 120Hz, 20°C"

I read it carefully enough to see that the spec is, strictly speaking, for the HA series, not the HF series.

If we could find an actual spec sheet for the HF series, it might specify the capacitance at 15.75 kHz, or even higher.  That would make sense, since they obviously used heavily etched foil to make this cap.  Perhaps the HA series didn't use etched foil, and its capacitance doesn't vary 2 to 1 as the measurement frequency increases toward the SRF.

As I said above, even though the spec sheet is for the HA series, the description of a capacitor designed to carry horizontal frequency currents would seem to be applicable, even if perhaps not exact in every respect.

You seem to be missing the point of my post.  The point I was making is that if the Nichicon cap is intended to operate at frequencies above 10 kHz, the capacitance at those frequencies is what we care about, and it's nearer to 10 uF than it is to 20 uF.  The capacitance measured by a DVM (they effectively measure at a very low frequency), or by an LCR meter at 120 Hz is not really relevant to use of the capacitor at horizontal deflection frequencies.  The capacitance marked on the cap is correct if the measurement is made at a much higher frequency than 120 Hz.

The OP wondered why his meter gives a value of about 20 uF when the value marked on the cap itself is 10 uF.  I gave him an explanation.
 

Online wraper

  • Supporter
  • ****
  • Posts: 16860
  • Country: lv
Re: Bad cap or bad measurement?
« Reply #22 on: July 29, 2014, 10:30:03 pm »
Explanation is that it is just a Chinese fake of non existing series. One should know that it is much easier to buy fake capacitor than genuine one on ebay/aliexpress (and any mention of it exists only there). Especially if it is some rare and expensive one, and non polar ones are of such kind. Show me a single datasheet for electrolitic capacitor where capacitance is specified in kHz range. If you would think a little bit, then you would understand that it is barely possible to measure such capacitance at 100kHz frequency. Capacitive reactance for 10uF capacitor is 0.16 Ohm at 100 kHz and 1.6 ohm at 10 khz. For such capacitors ESR usually is >1 Ohm. How do you expect to reliably measure capacitance in conditions when ESR prevails over capacitive reactance by order of magnitude?
 

Offline The Electrician

  • Frequent Contributor
  • **
  • Posts: 743
  • Country: us
Re: Bad cap or bad measurement?
« Reply #23 on: July 29, 2014, 10:52:26 pm »
If you would think a little bit, then you would understand that it is barely possible to measure such capacitance at 100kHz frequency.

I've thought a lot about it, and it's not just barely possible.  A DE-5000 can easily do it.

Reply #17 shows the result of such a measurement over a 100 Hz to 1 MHz span.

Capacitive reactance for 10uF capacitor is 0.16 Ohm at 100 kHz and 1.6 ohm at 10 khz. For such capacitors ESR usually is >1 Ohm. How do you expect to reliably measure capacitance in conditions when ESR prevails over capacitive reactance by order of magnitude?

I expect to use an instrument that can measure capacitance and ESR when ESR and capacitive reactance are more than 3 orders of magnitude different, such as these:

http://www.waynekerrtest.com/global/html/products/LCR/6430%206440.htm

http://www.home.agilent.com/en/pd-715495-pn-E4980A/precision-lcr-meter-20-hz-to-2-mhz?cc=US&lc=eng

http://www.hioki.com/products/lcr_resistance_signal/lcr_resistance_meters/654
 

Offline The Electrician

  • Frequent Contributor
  • **
  • Posts: 743
  • Country: us
Re: Bad cap or bad measurement?
« Reply #24 on: July 29, 2014, 11:05:48 pm »
Explanation is that it is just a Chinese fake of non existing series.

It's not a fake of the sort you linked to:
http://studioboytimes.tumblr.com/post/68565865754/a-fake-electrolytic

I opened one up, and it doesn't appear to be a fake at all.  It's well built and reminds me of a photoflash capacitor.  The internal foil pak fits tightly into the can, to give good thermal connection to the can:

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf