B&K 879B LCR meter Problem

[bsgd]

Hello all,

This is my first post here, even though I have watched all Dave videos since the beginning of his blog.

I have recently bought a B&K 879B LCR meter, as it appears to be a good unit, and this is the most affordable LCR available in my country.

Today I decided to do some impedance measurements, but the readings on my B&K are far from what I expected.
For instance, I wanted to measure the primary winding impedance of a transformer I have here. The specs say it should be 5KOhms at 1kHz. So I went ahead, and connected my 879B and it measured 234 Ohms in Z mode, and the exact same thing in R mode.
Just out of curiosity I mesured the resistance with my Fluke 87 and it measured 21.3 Ohms.

To make sure I wasnt using a defective transformer, I measured the impedance by connecting a 5V 1kHz signal and measuring the current in the primary. Using Ohms Law I calculated Z as 5.1KOhms, very close to the expected 5K.

So, I dont know exactly why, but my 879B gives me a completely unexpected reading.

Does anyone know whats going on here?

Thanks!

[olsenn]

You may have a defective LCR meter. Your reading in resistance (R) mode should equal that of your Fluke 87. Check that you're on a frequency of 1 KHz for you impedance check and try switching between parallel and series mode just to make sure that isn't screwing you up, not that it should.

[bsgd]

You may have a defective LCR meter. Your reading in resistance (R) mode should equal that of your Fluke 87. Check that you're on a frequency of 1 KHz for you impedance check and try switching between parallel and series mode just to make sure that isn't screwing you up, not that it should.

Hi olsen, thanks for replying. Yes, I was using 1kHz and I tried both series and parallel measurements. No difference at all.
I cant believe this tihing is defective, I bought it brand new... argh 

[firewalker]

Just for the record. Have you tried to measure other well known parts?

If the meter is within the warranty don't worry. Just take it back and ask for a new one.

Alexander.

[grenert]

My (very) limited understanding of LCR meters is that some meters simply cannot measure big iron-core inductors like transformers at high frequency.  The losses in the core appear as resistance to the meter and screw up the reading (which also goes along with the same reading on both Z and R).  It will probably give you a more reasonable reading if you test at 100/120 Hz.

Note that your resistance measurement with an LCR will be AC resistance, while the Fluke is giving you DC resistance.

[Drirr]

this reading is caused by dissipation factor of transformer core at 1KHz. I'd say this unit is not defective.

[bsgd]

Just for the record. Have you tried to measure other well known parts?

If the meter is within the warranty don't worry. Just take it back and ask for a new one.

Alexander.

Yes, I did try a couple other transformers, cause thats what I mostly wanna measure. All give pretty wrong readings.

My (very) limited understanding of LCR meters is that some meters simply cannot measure big iron-core inductors like transformers at high frequency.  The losses in the core appear as resistance to the meter and screw up the reading (which also goes along with the same reading on both Z and R).  It will probably give you a more reasonable reading if you test at 100/120 Hz.

Note that your resistance measurement with an LCR will be AC resistance, while the Fluke is giving you DC resistance.

Well, that explains it, but that makes me pretty sad cause I thought this unit could do this. Guess I'll have to keep on using Ohms Law or sepnd money on a 4263B.
Regarding the resistance measurement, yes I know this meter does no use DC so the readings might get screwed. But I was mostly worried about Z measurements.

I will test again at 100Hz and see what I get. I remember trying this measurement at lower frequencies but the error was still too big.

[vk6zgo]

As you have already determined the impedance with your 5kHz signal,make a guess that it is mostly inductive reactance,then calculate the inductance,switch your  LCR meter to read inductance,& see what you get!

Back in the day, (1980s),I used a digital LCR  tester quite a lot, to check transformers,by simply reading inductive X & losses.
The normal way was to compare with a known good unit,then any significant variation indicated a faulty unit.
Very good for switchmode supplies!

VK6ZGO

[bsgd]

My (very) limited understanding of LCR meters is that some meters simply cannot measure big iron-core inductors like transformers at high frequency.  The losses in the core appear as resistance to the meter and screw up the reading (which also goes along with the same reading on both Z and R).  It will probably give you a more reasonable reading if you test at 100/120 Hz.

Note that your resistance measurement with an LCR will be AC resistance, while the Fluke is giving you DC resistance.

I just tried changing frequency but readings vary 10% at most. Far from what I wanted.
There should be a warning somewhere on the manual stating these limitations. I read it and theres nothing about this... makes me feel quite uncomfortable using a measurement equipment that just lies to me without warning.

Too bad the Agilent 4263B cost a fortune here, or I would invest in a higher end unit.

As you have already determined the impedance with your 5kHz signal,make a guess that it is mostly inductive reactance,then calculate the inductance,switch your  LCR meter to read inductance,& see what you get!

Back in the day, (1980s),I used a digital LCR  tester quite a lot, to check transformers,by simply reading inductive X & losses.
The normal way was to compare with a known good unit,then any significant variation indicated a faulty unit.
Very good for switchmode supplies!

VK6ZGO

Doesnt work, not sure why. The primary impedance is 5.1K, resistance is 21 ohms.
If my math is right, I should get around 0.8H inductance. My B&K reads 6 mH 
This is a big Hammond 100W output transformer for tube amplifier, just out of curiosity.

Now I would like to have a know value inductor to try, but I dont have one around.

Im wondering if this is just my unit messing up things or a common problem. Looks like Dave did like his unit. But I feel like I cant trust mine.

For capacitors and resistors it works ok.

[ejeffrey]

Yeah, it definitely should work OK.  In series inductance mode you should get ~ 0.8 henry + some small resistance.  The resistive component will be more than the 21 ohm DCR due to core losses, but still much less than 5 k.  If that isn't what you get, it definitely sounds like some problem with the meter.  Alternately, in Z/theta mode you should get 5.1 kOhm total impedance with a phase angle near 90 degrees.

[grenert]

I just tried changing frequency but readings vary 10% at most.

Is that 10% from the transformer specs (which you say are for 1 kHz), or 10% from the value that you calculate when you retest the transformer with a 100 Hz signal and measure current.  Make sure you are comparing oranges and oranges.

There should be a warning somewhere on the manual stating these limitations. I read it and theres nothing about this...

That is the risk of buying something that was not designed by the company who is selling it.  Unfortunately, as you say, meters from companies that do their own design tend to be costly.

[saturation]

Just FYI, HH LCR meters may not be able to read inductance or resistance properly [using AC versus DC on basic DMM ]  on transformers because of the core and way the windings are wound compared to inductors alone.  They were designed to measure components, not like true impedance meters, which often costs 10-100x more than HH.  You can actually do more measurements with just a scope and function generator than a HH LCR meter, but you need to do calculations; thus LCR meters have limited usefulness compared to a true impedance meter.

http://cp.literature.agilent.com/litweb/pdf/5967-5377E.pdf

Even if you successfully did it on one type transformer, if its all the same manufacturer and you want to simply sort good from bad items, that measurement may not be applicable to other transformers with different build qualities.

Also, when comparing measurements to spec sheets, the frequencies and output voltages for transformers are fairly specific and may not be supported by your LCR meter.

http://www.motech-europe.com/html/mt4080.htm

[bsgd]

Yeah, it definitely should work OK.  In series inductance mode you should get ~ 0.8 henry + some small resistance.  The resistive component will be more than the 21 ohm DCR due to core losses, but still much less than 5 k.  If that isn't what you get, it definitely sounds like some problem with the meter.  Alternately, in Z/theta mode you should get 5.1 kOhm total impedance with a phase angle near 90 degrees.

My B&K reads an angle of 2 degrees. No sense at all.

I just tried changing frequency but readings vary 10% at most.

Is that 10% from the transformer specs (which you say are for 1 kHz), or 10% from the value that you calculate when you retest the transformer with a 100 Hz signal and measure current.  Make sure you are comparing oranges and oranges.

There should be a warning somewhere on the manual stating these limitations. I read it and theres nothing about this...

That is the risk of buying something that was not designed by the company who is selling it.  Unfortunately, as you say, meters from companies that do their own design tend to be costly.

I guess I wasnt very clear here. Measuring a 5K impedance reads 240 Ohms at 100Hz and 220 at 10kHz. Not much difference (~10%) using the exact same transformer.

An Agilent 4263B here costs around USD 5K. I dont wanna spend all that, but might end up doing so. I thought this B&K was up to the job...

Just FYI, HH LCR meters may not be able to read inductance or resistance properly [using AC versus DC on basic DMM ]  on transformers because of the core and way the windings are wound compared to inductors alone.  They were designed to measure components, not like true impedance meters, which often costs 10-100x more than HH.  You can actually do more measurements with just a scope and function generator than a HH LCR meter, but you need to do calculations; thus LCR meters have limited usefulness compared to a true impedance meter.

http://cp.literature.agilent.com/litweb/pdf/5967-5377E.pdf

Even if you successfully did it on one type transformer, if its all the same manufacturer and you want to simply sort good from bad items, that measurement may not be applicable to other transformers with different build qualities.

Also, when comparing measurements to spec sheets, the frequencies and output voltages for transformers are fairly specific and may not be supported by your LCR meter.

http://www.motech-europe.com/html/mt4080.htm

Well, my intention here was to be able to do quick impedance measurements of output transformers (which are wound in a very specific and complicated way, and use GO cores). I wasnt aware that HH meters could give me false readings like the ones Im getting.

Right now I dont know if my B&K is defective or not. Maybe this is normal?
What I do know is that from the calculations, this specific transformer that I randomly chose to do these tests, should meaure 0.8H... my B&K says 6mH (wtf??) in series mode and 12H in parallel mode.
My cheap $20 Minipa DMM reads 0.87H, quite acceptable.

I dont know whats going on here. Some say these HH LCR meters are not up to the task, others say I just got a defective unit. Does any of you have a known value output transformer and could try to read the impedance/inductance with a B&K 879B? That would be good...

[vk6zgo]

You should be able to wind up a small inductor,using one of the formulas which appear in most textbooks,then check it with your sinewave generator as before---it should be fairly close to the expected value.
Now,test that inductor with your B & K.
If it reads nonsense,then you probably have a faulty unit.
Does it measure C & R correctly?
VK6ZGO

[bsgd]

OK, so here are some results...

I made a 0.5H inductor, and the final value measured using other instruments was 0.57H (not too bad).

Using the B&K I get:

At 100Hz - 1.3H theta=76.6 degrees
At 120Hz - 1.24H theta=76.35 degrees
At 1kHz - 0.63H theta=77 degrees
At 10kHz - 0.24H theta=44 degrees

Is such huge variation normal? Im not sure inductors vary that much in frequency but anyway, it is possible.

I also found an old 8H inductor here and decided to try it... my other instruments say 8.1H.
Using the B&K at 1kHZ gives me 8.6H but at 10kHz it says -2.1H (yes, minus). What does it mean exactly?

[bsgd]

Sorry vk6zgo, forgot to answer your questions... the B&K reads resistors perfectly, and also seems to work fine with capacitance. There is a big difference between the read capacitance values when you select different measurement frequencies, but I suspect this is normal...

[vk6zgo]

That is a serious inductor to wind up  by hand!
I was thinking more along the lines of a few mH!
 You will notice with both your handwound & "junkbox" inductors,the figure using 1KHz is quite close to what it should be,with the  low & high frequencies being way out.
The inductance reading should hardly change,but the reactance  will normally increase with frequency.The instrument should be calibrated at the factory to allow for this.
The capacitance measurement should also not vary radically at different measurement frequencies.

Most of the comments re core loss etc,on the original transformer seem more aimed at power transformers,rather than the wideband audio transformer you were testing.
I would suggest that your LCR meter is faulty,& you should return it under warranty.

VK6ZGO

[saturation]

It should not vary much from the specified 'basic accuracy'.  b&k should specify the accuracy at each frequency in the manual.  Check the output frequency of the LCR meter with a scope, it might be off or unstable.  The unit appears defective.

[bsgd]

I would like to thank you all for the help.

I have decided to send this B&K back for a replacement (since here I just cant get my money back), and I'll be looking for a new (bench type) LCR meter soon.

Thanks again guys.

[The Electrician]

Well, my intention here was to be able to do quick impedance measurements of output transformers (which are wound in a very specific and complicated way, and use GO cores). I wasnt aware that HH meters could give me false readings like the ones Im getting.

The impedances associated with the windings of an audio transformer are sometimes a source of confusion.  Consider an output transformer with a 5k ohm primary and an 8 ohm secondary.  What that means is that with an 8 ohm load connected to the secondary, an impedance of 5k ohms will be measured at the primary.

But you are apparently measuring the primary impedance with nothing connected to the secondary.  That's not the proper way to measure the rated impedances of the windings.

See this thread: http://www.electro-tech-online.com/general-electronics-chat/124261-audio-transformers-two-ports.html

OK, so here are some results...

I made a 0.5H inductor, and the final value measured using other instruments was 0.57H (not too bad).

Using the B&K I get:

At 100Hz - 1.3H theta=76.6 degrees
At 120Hz - 1.24H theta=76.35 degrees
At 1kHz - 0.63H theta=77 degrees
At 10kHz - 0.24H theta=44 degrees

Is such huge variation normal? Im not sure inductors vary that much in frequency but anyway, it is possible.

I also found an old 8H inductor here and decided to try it... my other instruments say 8.1H.
Using the B&K at 1kHZ gives me 8.6H but at 10kHz it says -2.1H (yes, minus). What does it mean exactly?

The measured inductance and impedance of a winding on an iron core can vary with freqency and with the excitation voltage applied by the LCR meter making the measurement.

The permeability of iron core material varies with flux density; it decreases with decreasing flux density.  For a constant applied voltage, the flux density decreases with frequency so the measured inductance will decrease with increasing frequency.  Some core materials such as those with a high nickel content don't exhibit this effect to such a large extent, but regular grain oriented silicon steel does exhibit the effect to a noticeable degree.

Also, the high impedance winding of an audio transformer is probably wound with a lot of turns of small diameter wire.  Such a winding will have a parallel resonance due to the distributed capacitance of the winding.  As the inductance measurement is made at a frequency approaching the resonance frequency, the apparent inductance may decrease and eventually become negative at frequencies above the resonance.

I've attached a couple of images showing a frequency sweep of the primary impedance and inductance of a 5k winding on a Bogen audio transformer.  The green curve is the impedance and the yellow curve is the inductance.  The impedance is shown logarithmically with 100 ohms at the bottom and 1 megohm at the top.  The inductance is shown on a linear scale with -8 henries at the bottom and +8 henries at the top.  The frequency sweep goes from 100 Hz at the left to 100 kHz at the right.

One image shows the measurements with an excitation voltage of .1 volt and the other with 5 volts.

Parallel resonance occurs at about 3 kHz and the effects I described can be seen in these images.  Notice how the measured inductance becomes negative above 3 kHz.

To test your B&K without these effects confusing the results, use an inductor without an iron core.

[saturation]

Thanks electrician, your electro-tech post is very good and welcome to eevblog.

Which, slightly off topic, brings up the question of the value of handheld LCR meters, what's your take on them, like this B&K.  I think folks erroneously equate them with an impedance analyzer and ask too much of them.  Its fine for checking specs on a spec sheet, say for sorting or doing QC checks, but the actual performance of reactive components in application requires a bode plot.  Given the costs of such a device, one can do a similar task with just a scope and a function generator, so HH LCR meters have limited usefulness.

[The Electrician]

Circumstances will, of course, determine whether a hand-held LCR meter can be justified.  For a hobbyist who only occasionally needs to measure a capacitor or inductor, certainly an audio generator and scope can do the job.  In an engineering and manufacturing environment, a couple hundred bucks is nothing, compared to the $10k to $45k price for a high end impedance analyzer, and can be quite useable if the limited capabilities of the HH unit are sufficient; it keeps the technicians from having to wait for access to the more powerful instrument for simple QC tests.

The OP seemed to feel that for him the HH was worth the convenience and speed and a measurement at 1 kHz or 10 kHz is apparently enough.

For engineering design work, I agree with you that a full frequency sweep is vital.

[bsgd]

The Electrician,
First I would like to thank you for your post. Really nice and well explained.
Even though I did not mention, yes, I was using the correct secondary load. Is my case, I was using 8 Ohms connected to the secondary.

saturation,
I now understand that I was expecting too much from such a simple meter. I wasnt aware these devices were not adequate for my needs... in my head, an 'impedance meter' should always be able to measure impedance since the manual did not state these limitations we have been discussing. Yes, I know I was a little wrong in my assumption!
IMO the manufacturer should make this clear in the manual. There should be a note somewhere stating the instrument limitations. The B&K manual is simply awful... I am very unsatisfied and do not intend to buy any other B&K equipment.


Once again, I did learn a lot with this discussion and experience

[saturation]

Hi bsgd

You're welcome.  Electrician's post said it best, but sorry its slightly off your initial request.

The HH are great for spot checking components.

I think in your OP you used a function generator ?? or similar to spot check the HH.  You can do this for most all measurements too and do what both a HH LCR meter and an expensive impedance meter can do without spending more.  The problem is the set up and the cables to make it convenient; on the Rigol 1052e you can set up a pass-fail waveform to make it easier.  In high frequencies, the clamps, cables, and other fixing hardware used have intrinsic capacitance and inductance which now matter more, and have to be taken into account.

[bsgd]

I think in your OP you used a function generator ?? or similar to spot check the HH.  You can do this for most all measurements too and do what both a HH LCR meter and an expensive impedance meter can do without spending more.  The problem is the set up and the cables to make it convenient; on the Rigol 1052e you can set up a pass-fail waveform to make it easier.  In high frequencies, the clamps, cables, and other fixing hardware used have intrinsic capacitance and inductance which now matter more, and have to be taken into account.

I guess I need some more info on how to set this up. Can you help me?
I do have a DS1052E and a DG1022, and also an HP 8903 so this is more than I need to make these plots I suppose. But I am not sure how to do it.