LCR meters, how much is too much?

[rx8pilot]

I need an LCR meter for basic low frequency measurements, component ID, etc. The smart tweezers seem great for that and maybe an bench model that goes to 100Khz for more detail.

Thinking about the future as I have no intention of slowing down my education and design ambitions - I was looking at the higher-end options to see what is out there. The most obvious observation is how fast the price climbs along with test frequency (no surprise considering the challenge). This is a budgeting exercise trying to understand the amount of money I will need to take advantage of the educational effort I am putting into EE. The paid projects in the next year are mainly power electronics, but I am getting into high-speed data processing with FPGAs, and if my brain has any space left-over would like to begin a path into RF design (a few years off though).

The question is this: at what point do you suppose I would feel a need for something better than a 100Khz LCR? Are the manufacturer data sheets enough? Who spends $15k, $30k, and more on LCR meters. This is not a 'which one do I buy' question, but more generally, trying to understand the reasoning for putting big $$ LCR instruments on the bench.

[nctnico]

Beyond $1000 I'd look for a (used) network analyser for measuring components and other things (like impedances of transmission lines).

Maybe an expensive LCR meter can measure component values with more precision but since capacitors and inductors have very wide tolerances (not to mention parasitics in a circuit board) I wonder why you need to measure them with high precision for a production circuit.

[rx8pilot]

Beyond $1000 I'd look for a (used) network analyser for measuring components and other things (like impedances of transmission lines).

Maybe an expensive LCR meter can measure component values with more precision but since capacitors and inductors have very wide tolerances (not to mention parasitics in a circuit board) I wonder why you need to measure them with high precision for a production circuit.

I was wondering the same thing....yet there are a number of crazy expensive models out there.

[HighVoltage]

One thing you get with more expensive LCR meters (besides higher precision and accuracy) is speed of detecting the component correctly in auto mode and therefore throughput in a production environment.

Used LCR meters drop in price almost as fast as used scopes and you can find great deals if you have a little time.

[Solder_Junkie]

It's easy to get carried away with test equipment and an LC (and perhaps R) meter certainly falls into that category. I have a cheap and cheerful LC200A that costs little on eBay, they run at 500 KHz and claim an accuracy of 1%.

Measuring capacitors is fairly easy, as are inductors which are air cored, however the problem becomes significant when you try to measure dust and ferrite cored inductors intended to operate above the operating frequency of the meter. Keysight will be happy to part you with tens of thousands of dollars in order to test up to 3GHz and plot the results, their 4287A is $17K (used) on eBay and still only has a basic accuracy of 1%

In a nutshell, you can pay anywhere from about $20 to $20,000 or more.

[DimitriP]

Who spends $15k, $30k, and more on LCR meters.

I'll guess agencies and industries for which a mix of the below are true
A) budget
B) have to follow someone elses speficications for measuring tolerances
and
C) the difference in the measurement  between a low and high tolerance will cause the circuit to absolutely fail.

(potentially D  for "cool factor" for completeness  as in  showing off the meter when making youtube videos    but that is not enough to keep the high end LCR instrument industry "fed" )

[R005T3r]

You need something better than 100Khz when ever you are going up in frequency because capacitance and inductance do not remain constant if you go up in frequency. Think about RF connectors  and relays... If you are a manufacturer, you surely want to see how your product behaves at higher frequencies before putting them in the market.

[Fgrir]

The day you need a >$10K LCR I think two things will be true:

• You will know why you need it without having to ask anyone else for ideas.
• You will still need your <$1K LCR for everyday usage.

So buy a cheap one now and wait until you realize it won't do something you need it to.  As you've noticed the prices seem to go up exponentially so that initial investment won't seem like much in comparison to the prices for getting into the good stuff.

One reason I have run into for wanting something better is the steady march of switching power converters to higher and higher frequencies.  The 100KHz specs from the datasheets for inductors and capacitors don't really tell you a lot about how they will work at 500KHz, 1.2MHz, etc.

[rx8pilot]

So it definitely makes sense for component manufacturing and design where a component needs to be fully characterized and documented.

I am under the assumption that as a circuit designer, I can just count on the  data sheet for all but the most exotic of designs.



Sent from my horrible mobile....

[rx8pilot]

I am definitely not in a hurry to spend $10k on one - as you say, I will know if I really need it.

Most of my SMPS designs are between 500khz and 1mhz where you see very fast edges - certainly above 100khz. In those  cases, I just look at the data sheet for guidance when the part is expected to perform at many 10's or 100's of mhz.

Sent from my horrible mobile....

[nctnico]

One reason I have run into for wanting something better is the steady march of switching power converters to higher and higher frequencies.  The 100KHz specs from the datasheets for inductors and capacitors don't really tell you a lot about how they will work at 500KHz, 1.2MHz, etc.

True. What is important is the kind of core material and losses in the windings at higher frequencies. Wurth (for example) has an online tool in which you can specify a switching topology, currents, frequency, etc to get an inductor which suits your requirements (small size, low losses, etc). IMHO this kind of stuff should be in the datasheet but somehow coil winders have not arrived at this point yet. Maybe that is because they use LCR meters where they should use network analysers.

[G0HZU]

I've said this several times already but I have a lot of nice test gear here but I've never bought/owned an LCR meter or an ESR meter and that is across a >30 year time frame. This is mainly because the lower cost ones are usually unrealiable in terms of accuracy although the modern meters like the DE-5000 look to be very good these days.

If I have ever needed to test a (uF range) cap for value/ESR I've done it the old school way with a clean sinewave source (sig/function generator) and a sense resistor and either a dual channel scope or a decent AC voltmeter. With care, I think it is possible to get results that are accurate enough for most things. Not as good as the $$$ lab meters but probably better than most of the sub $150 hobby ESR meters. The cost is effectively zero if you already have the sinewave source and the meter/scope The downside is that the procedure is slow because it involves entering the readings into a spreadsheet or computer program to get the results.

The scope method gets you up into the MHz region if this is needed.

[nctnico]

This is mainly because the lower cost ones are usually unrealiable in terms of accuracy

Maybe that used to be the case in the old days but with a simple microcontroller and some straightforward analog circuitry (synchronous rectifier) you can make a pretty decent (better than 1%) LCR meter very cheap nowadays. The cheap Chinese one I got has proven it's worth and accuracy in a way which made the old HP boat anchor come and go.

[rx8pilot]

I looked at some of the usual manufacturers of MLCC's and found that frequency response data is hard to come by without some digging.

TDK has a nice characteristics PDF page that is helpful
https://product.tdk.com/info/en/documents/chara_sheet/C0603C0G1E101K030BA.pdf

Other manufacturers have similar data, but only for the RF specific products. I am sure the data exists somewhere for the less special products if I looked hard enough. Like nctico said earlier, if I really wanted to learn the characteristics of the capacitor - a VNA seems more appropriate. Can't you measure capacitance ESR, and ESL with a VNA out to the frequency limit of the VNA? I have only read about VNA's - never had an opportunity to use one.

So far, most of the T&M gear that I have read about make sense from a practical and economic perspective in a circuit design and validation role. The really fancy high-end LCR meters, I am not sure what people need them for other than designing or manufacturing components. This question is about to lead me down a rabbit hole of capacitors and how to measure them. Not because I have any need at all - but because I asked myself question and won't stop until I have a reasonable understanding. Personality flaw.

[Someone]

One reason I have run into for wanting something better is the steady march of switching power converters to higher and higher frequencies.  The 100KHz specs from the datasheets for inductors and capacitors don't really tell you a lot about how they will work at 500KHz, 1.2MHz, etc.

True. What is important is the kind of core material and losses in the windings at higher frequencies. Wurth (for example) has an online tool in which you can specify a switching topology, currents, frequency, etc to get an inductor which suits your requirements (small size, low losses, etc). IMHO this kind of stuff should be in the datasheet but somehow coil winders have not arrived at this point yet. Maybe that is because they use LCR meters where they should use network analysers.

It all gets further complicated with temperature and current waveforms, similarly without a HF sweep it can be hard to see the interwiniding capacitance effect.

I am under the assumption that as a circuit designer, I can just count on the  data sheet for all but the most exotic of designs.

Getting maximum performance out of switching regulators you start pushing the limits of the parts, and parameters of the diodes and ferrite core often have substantial thermal dependence. The "solution" to the complexity of all these parameters is often to buy a few eval boards and a pile of candidate parts to torture under the conditions you expect to use them in. A thermal camera is often a better investment than a fancy impedance analyser here.

This question is about to lead me down a rabbit hole of capacitors and how to measure them. Not because I have any need at all - but because I asked myself question and won't stop until I have a reasonable understanding.

Capacitor data from manufacturers is often excellent, murata have sweeps of most of their products and capacitance vs voltage etc. Its improved a lot in recent years to the point where I don't bother measuring capacitors anymore.

[rx8pilot]

Capacitor data from manufacturers is often excellent, murata have sweeps of most of their products and capacitance vs voltage etc. Its improved a lot in recent years to the point where I don't bother measuring capacitors anymore.

Murata was one that I did not easily find frequency sweeps for - clearly looking in the wrong place.

[PA4TIM]

Your work dictates the level and type of LCR meter you need. Caps in RF filters ask for different gear as high power inductors.  Companies that build calibration gear, or LCR meters  have other demands as a company that makes simple cheap power supplies.

You  want  the value at the frequency the component operates.  But there is  DC bias, shielding your DUT, guarding, more wire measurements, precision, repeat-ability, stability (f.i like over time and temperature)  Output frequency, voltage and current.

Then speed, ease of use (you can measure aF with a GR-1620 set but that can give you 30 minutes of fun and frustration, that is, if you know how to use it correct. )

I was building a high resistance standard for a customer to be sure the leakage through the cabinet was as small as possible I used my old
HP-4329A ( http://www.pa4tim.nl/?p=2269 )  While doing that I could not get stable readings, after cleaning it over and over and a lot frustration I found the problem. I was standing about a meter or so from my dut. The movement of my belly while breading caused the fluctuations. 

Yesterday I had to match/measure some Peltiers for a drywell design. You need to measure the resistance with AC but they produce a current so you need an LCR meter that does not care about DC. Not about LCR meters, I once repaired an old analog simple function generator. The reason for repair instead of replacing it with a modern one was the conditions it was used in. A modern digital would not survive that.

So there are a lot of factors and  if you need something almost nobody needs it becomes expensive, that is, if enough people need it and a manufacture makes what you need. The price drops if something can be mass produced with "normal" parts.


And there are many more special cases.

[Someone]

Capacitor data from manufacturers is often excellent, murata have sweeps of most of their products and capacitance vs voltage etc. Its improved a lot in recent years to the point where I don't bother measuring capacitors anymore.

Murata was one that I did not easily find frequency sweeps for - clearly looking in the wrong place.

You need to search the part on their website and then it brings up plots of the measured parameters with the specifications. It also lets you pick different parts to overlay and compare. Lots of quality data and you can figure out which parts are just binnings of each other

[Bud]

I am for one that a VNA would be much more useful. Plus you get a VNA 
I still use a handheld Agilent LCR meter for casual measurements though.

[Richard Head]

I use a spectrum analyser and tracking generator to measure the self resonant frequency of capacitors and inductors. It's essential for RF work to know how high in freq they can be used. Once I characterize the components I archive the info for later use.
For power electronics you can also measure the self resonant freq of transformers and inductors. Inductive components with a high number of turns can have a surprisingly low self resonant freq due to the inter-winding capacitance. But actually with a bit of ingenuity, a scope and a sig gen you can measure it also, no need for expensive gear.

[rx8pilot]

Did you buy or build a test fixture for that method?

Sent from my horrible mobile....

[Richard Head]

Very basic really. I used the two MIL-17 teflon coaxes from the SA and TG and soldered them to the component under test (mounted on a scrap PC board with a via to the ground plane. I connected the coaxes directly to the capacitor/inductor with as little shield and centre lead length possible. I repeated the test with many different values of 0805 capacitors and trimmers. The first dip in the response indicated the SRF. Knowing the self resonant freq you can work out the parasitic inductance which seems to be package dependent. I ended up with a value of about 1.4nH for 0805 NPO chip caps below 1nF.
All it really gives you is an idea of the upper limit of the capacitor or inductor. Essential for knowing if a particular value of bypass cap will do the job properly.
The proper way to do it is to use a VNA to characterize the component and then export the model into an RF simulator. You can also use the method to measure the stiffness of power rails to see if the bypass caps are correctly selected. 
For inductor testing you look for the first peak in the response. You'll notice foil wound inductors tend to have a lower SRF compared with wire windings. Useful to know for PSU design.