High-precision L/C Inductance Capacitance meter Design

[Mechatrommer]

as requested by OP of https://www.eevblog.com/forum/index.php?topic=1855.0;topicseen ... i have to move on here. So i hope some feedbacks...

[Mechatrommer]

You can make your own even simpler since the relationship in high pass filters are:
http://www.play-hookey.com/ac_theory/hi_pass_filters.html
So you can build this very quickly and measure either capacitance or inductance; getting a meter really matters if you don't have the time to calculate it as listed here:
https://www.eevblog.com/forum/index.php?topic=533.msg7756#msg7756

ok mr saturation! i think its time and place to discuss this more in depth. this looks nasty'ly simple (theoritically) and i've been astonished with it since the first time i opened the ee book and saw it. my basic rule is that when i cannot understand a simple system, then i should go even simpler. now, the RC and RL hi/lo pass filter is the most basic thing on earth, theoritically to build an LCR meter is great with the super simple 2 passive components circuit... but just how to practicality it is, is where i'm still stumbled at. so let me point out of what i understand...

1) all we need is a setup/schematic shown in your link (either RC or RL), and... Vin supposedly be connected to some Vac generator... mcu PWM pin maybe?
2) all on earth the basic equation is the famous w = 1 / RC. is this means, we need to setup a known w and R in order to get the C value?
3) since we need to feed alternating V, is this means Vcc to -Vcc peak to peak instead of Vcc to 0V? so how to get the -Vcc voltage? can we put the R or L ground (as in the link's picture) to some dummy reference at Vcc/2 so we have Vcc/2 and -Vcc/2 seen by the filter circuit?
4) and the rest of the math is done in mcu programming right? ok leave this for now, just assume i'm too good to program that "software'tically" ... as i said, i always stuck with analog side .

i'm looking forward/expecting/hope to see this basic questions evolve to a bigger thing like the many DIY circuits provided by dear fellow members here.

[Mechatrommer]

maybe i got it wrong for the equation...
continuation from http://www.play-hookey.com/ac_theory/hi_pass_filters.html
here goes the ugly sketching. hope i got clearer solution.

[Mechatrommer]

That's good math work, safri, but the equation to use is different and far simpler.

Notice how all LCR meters use a specific freuquency for a range of C or L, that's roughly how this works.









Resistor here is a constant, use any thin film or carbon composite.  Don't use wirewound as its a inductor too.

Now, put your unknown capacitor or inductor in the circuit above, as necessary.

Get a function or audio generator generator as Vin, and measure its output Voltage with a scope for say 500kHz or 500 Hz, just to copy the frequencies the device in the original post used.  Say adjust it for 1 Vp-p  so its easy to read.

If your cap is as big as your pinky or your inductor is fairly large, use the 500 Hz.

Now connect the generator to Vin and the scope to Vout.  Adjust the output frequency precisely until the output voltage is 0.707Vp-p.  That's the fc.

Now that you have fc, substitute it algebraically into the equation and that's your L or C.

Units:

fc= Hz
R = ohms
C = farads
L = henry

For example, you have an unknown small inductor.
 
Use the inductor high pass circuit above.  You have some resistor lying around that is 500,000 ohms.

You pass 500kHz into it and it comes out 1V p-p.  Adjusting frequency down [ since its a high pass filter its already passing everything ABOVE the cutoff] , at 300kHz the Vout is 0.707 p-p.

Solving:

L = R / 2 x pi x fc

L = 500,000 / 2 x pi x 300,000 Hz

L = 0.265 H or 265 mH.



As alm and I suggested in a old post, most hobbyist only need a good scope, like the Rigol 1052E, a function generator, a good DMM and PSU.  An LCR meter is nice to have, but if you don't measure C or L often, and a good DMM comes with C meter already too, you can rig one up fairly quickly and get very accurate results with it.

Some modifications of the same rig can calculate ESR, DF, and Q too.  But they can be more tedious to do on-the-fly, so it would be worth it to buy or build an ESR meter than a LCR meter without ESR built in.

http://octopus.freeyellow.com/esr.html

[jimmc]

How about doing it this way?
http://www.sillanumsoft.org/ZRLC.htm

Jim

[TechGuy]

Get a function or audio generator generator as Vin

ok, the explanation and implementation seem simple. but the above quote... is a "houston we have a problem" thing. i dont have any. is there a way generating controlled Fc/Vac through a mcu?

Sure. A MCU with a built in PWM output or timer with output. You just need the MCU to output a square-wave signal to feed in to the LCR circuit. Alternatively you can use a VCO/PLL with a DAC to control the VCO frequency. Ideally you want to use several data points at different frequencies to get a more accurate inductance value. The more data points you use the better precision you will get. You may want to consider using some logic to determine the optimal starting frequencies to begin data sampling. You would probable start out with a low frequency (1 Khz) and work you way up so the the ADC value in near the center of the ADC range, than start collecting data points by increasing the frequency in 5 to 10% increments. than use a math formula to calculate the inductance value. You will also need to include support to calibrate the meter by running tests with an open and short connection (where the inductor will go) so you include them into your calculations. For instance your ADC (used to measure the output voltage will may probable not be zero in a open circuit configuation. Also consider using a ADC with 12 to 14 bit resolution as the built in ADCs (usually 8 to 10-bit resolution) may not provide sufficient accuracy. Perhaps using low-side MOSFET driver would be better than driving the LCR circuit from the mcu PMW output. the MOSFET driver would provide some isolation\protection from the DUT and also permit you to feed in a higher voltage (12V to 18V for the typical MOSFET low side driver). You can also use a a pair of BJTs (NPN - PNP) in a totem pole circuit, but the beware of the Vce drop with BJTs.

FWIW: I purchased a sencore LC-103 meter. the secore meter provides capacitor ESR and current leakage and can test caps up to a 1000V for leakage (many HV caps don't show high ESR or leakage at low voltage that most LCR meters are limited to).

If your looking for a low cost LCR meter take a look at this model from Peak (atlas)
http://www.peakelec.co.uk/acatalog/jz_lcr40.html
http://www.anatekcorp.com/testequipment/atlaslcr.htm (US distributor)

I have the Peak Semiconductor tester which works pretty good. Great for checking for busted transistors and diodes.
http://www.anatekcorp.com/testequipment/atlas.htm

[alm]

I think using a square wave makes things needlessly complex, since square waves are composed of many harmonics, and the attenuation of the filter is different for the various harmonics. I would stick to sines. You could make an RC oscillator , or use some sort of audio source (don't expect great stability).

I consider a function generator a basic piece of electronic equipment, together with stuff like soldering iron, DMM and scope, so you may consider getting one at some point in time, low-spec used ones can be pretty cheap.

[TechGuy]

I think using a square wave makes things needlessly complex, since square waves are composed of many harmonics, and the attenuation of the filter is different for the various harmonics. I would stick to sines. You could make an RC oscillator , or use some sort of audio source (don't expect great stability).

I consider a function generator a basic piece of electronic equipment, together with stuff like soldering iron, DMM and scope, so you may consider getting one at some point in time, low-spec used ones can be pretty cheap.

Suit yourself:
http://focus.ti.com/lit/ds/symlink/cdc3s04.pdf (5 bucks for a SPI controlled sinewave generator)
Or the ever popular XR2206:
http://www.exar.com/Common/Content/Document.ashx?id=16&LanguageId-1033 (VCO function generator)

[Mechatrommer]

http://www.anatekcorp.com/testequipment/atlaslcr.htm (US distributor)

low cost? $117? WOW! thanx for the info though

i've seen a discussion here somewhere about generating sine with mcu, might to have a look/search back on it and see how i can combine it in rc/rl filter.
for now the http://electronics-diy.com/lc_meter.php is on my left, the RLC pass filter is on my right, and in the middle with me? are the stars, moon and angel

[saturation]

Oops!  Safri you should make it your mission to get a function generator when you can.  They are all over the place, its one reason I assumed you owned one.  As alm suggests, I too think every lab should have basic measurement instruments: a scope, a DMM, a function generator.  Since these are present in most any labs, you can easily rig up an LCR measurement tool in a snap.

If you have an MCU, some come with built in oscillators, you can output that.  But to be able to measure reliably you'll need a wide range of programmable frequencies, otherwise you'll need a wide range of resistance that isn't wire wound. it may be too much work.

Alm is right, square waves just adds more problems, but it can be done with it.  The waveform will distort as the lower harmonics will be filtered, you need to watch the fundamental and you should know how to read a distorted square wave as a representation of frequency response. 

Get a function or audio generator generator as Vin

ok, the explanation and implementation seem simple. but the above quote... is a "houston we have a problem" thing. i dont have any. is there a way generating controlled Fc/Vac through a mcu?

I think using a square wave makes things needlessly complex, since square waves are composed of many harmonics, and the attenuation of the filter is different for the various harmonics. I would stick to sines. You could make an RC oscillator , or use some sort of audio source (don't expect great stability).

I consider a function generator a basic piece of electronic equipment, together with stuff like soldering iron, DMM and scope, so you may consider getting one at some point in time, low-spec used ones can be pretty cheap.

[saturation]

You can certainly build an FG yourself.  But getting a good one should be on a builder's bench.

The cost of building a quality FG versus what you can buy with the same money isn't what it used to be, its just not cost effective to build a quality FG as a lab tool, just as with making your own DMM or a good scope.

Its good to get the best tool you can afford to obtain, as all devices you build thereafter, and thus its quality, are measured against it and your skill.  A skilled craftsman can only build so well with mediocre tools, but with good tools, the best possible can be had.

That said, there are a lot of XR2206 style FG out there that cost as much as Instek's DDS version, which is a world away in stability, accuracy and distortion.  There are many models, but I find Instek's cheapest can beat many costly but old designs easily.



I got a model 1003 for $US150 delivered, from tequipment.net.

Suit yourself:
http://focus.ti.com/lit/ds/symlink/cdc3s04.pdf (5 bucks for a SPI controlled sinewave generator)
Or the ever popular XR2206:
http://www.exar.com/Common/Content/Document.ashx?id=16&LanguageId-1033 (VCO function generator)

[saturation]

Briefly, I used the circuit and equation setup on this thread, and used channel 1 for Vin and 2 for Vout.

I took a bunch of random capacitors from a box, not knowing their values, but estimating them in the uF and nF range.

I used a 9880 ohm resistor as checked with a Fluke 87V, it just happened to be around.

I used sine waves, and began at 1 kHz, using a Instek 1003 DDS function generator. 

I first measured the FG output alone as Vin,then loaded with the circuit and the unknown capacitance.  Generally, with the resistance used [ effectively impedance] Vin ~= Vout passed fc, but it was a good way to monitor both channels.

Accuracy: 0.5% to 1% compared against capacitance measurement by a Fluke 87V.

Some interesting insights:

Since the Rigol is 8 bit, using 1V led to larger voltage measurement errors than using larger voltages, 5Vp-p minimum was better, I settled on 10Vp-p.

I used the Rigol's automated measurement function to measure Vp-p voltages. I simply spun the frequency down or up until I met the cutoff voltage ~ 7.07V then plugged the frequency into the equation.




 

[TechGuy]

Oops!  Safri you should make it your mission to get a function generator when you can

I think he already has one. I believe he wanted to build an LC meter that did all of the calculation using a MCU, not to manually do it with a signal generator and DMM.

[DJPhil]

I got a model 1003 for $US150 delivered, from tequipment.net.

I've been eyeballing that, and the version that's ~$20US more expensive with 'voltage display'. It's good to hear positive things from a pro, as I feel I'm looking in the right direction. It's now officially next on my list after a scope.

[saturation]

Its worth it even if you already own an FG, to get into DDS, as Dave discussed in one eevblog, its a large leap in stability and reduced distortion over an analog FG, for <= price! 

If you know of other makes other than Instek, please link so I can review them.

I do not know of any FG near this price range for a DDS FG, from a reputable manufacturer, as this Instek.  Its like the 1062 scope versus the Rigol, the entry level is very competitive and affordable for it class of instrument.

For +$20 for the voltage readout I pondered that too, but a lot depends on how accurate it is.  Strictly speaking you can easily check it with a DMM, which is what I do.

FWIW the box is essentially empty, it weighs very little.  I bet the design could be far smaller, but they are saving money by reusing the casing for the standard range of FG.

It comes with a cal certificate too, which surprised me, using GW and Fluke instruments for traceability.

I got a model 1003 for $US150 delivered, from tequipment.net.

I've been eyeballing that, and the version that's ~$20US more expensive with 'voltage display'. It's good to hear positive things from a pro, as I feel I'm looking in the right direction. It's now officially next on my list after a scope.

[Mechatrommer]

Oops!  Safri you should make it your mission to get a function generator when you can

I think he already has one. I believe he wanted to build an LC meter that did all of the calculation using a MCU, not to manually do it with a signal generator and DMM.

in fact, i dont have any. my tools for ee are just...
1) iron n gun n acc
2) DMM
3) Rigol
4) and lastly, a nicely modded cheapy PC PSU.

i still dont have a justification to get one FG. if the reason is only to build this rig, then i think i better off with the ebay diy ready made thing... or even the more serious and professional brand LCR meter. i'm not expert and hobbiest enough for this. but the need to measure L and C is there for me, esp the smaller one, since i'm working/learning on RF system that needs these L and C stuffs.
ps: i wish i have a strong reason to get one FG.
btw: any suggestion for a good but "programmable arbitrary anything shaped" function generator, in the range of less than $100-$300? if in case this is the final outcome for me? to buy a FG. or maybe i'm just dreaming for that price?

[alm]

btw: any suggestion for a good but "programmable arbitrary anything shaped" function generator, in the range of less than $100-$300? if in case this is the final outcome for me? to buy a FG. or maybe i'm just dreaming for that price?

I think you're dreaming if you expect a real new arbitrary waveform generator for that price. Even the Rigol model with crappy software/support was something like $600+ last time I checked.

[saturation]

I've seen USB based AWG at hobbyist level prices.

A reason for getting the Instek is the quality of the generated output, but who knows what these Hantek/Riko can do? On paper, this cheapo seem as good if not better.   If you get one review it for us.

A post on eevblog:

https://www.eevblog.com/forum/index.php?action=search2

Alas, the OP has not reposted any further review.

For $180 US from China:

http://www.youroscilloscope.com/usb-pc-functionarbitrary-waveform-generator-dds3x25-p-18.html

On ebay for $150:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=150439450643&hlp=false&rvr_id=178392101396&crlp=1_263602_304652&UA=WXF%3F&GUID=ad8419371280a02681b38503ffc8ed2c&itemid=150439450643&ff4=263602_304652

[saturation]

sorry, try this:

https://www.eevblog.com/forum/index.php?topic=197.msg1892#msg1892

A post on eevblog:
https://www.eevblog.com/forum/index.php?action=search2

didnt work.

the dds3x25 is Amplitude ±3.5V Max. by that alone, its not tempting

[Mechatrommer]

this thread is not so new... so let me gather some older discussions here...
https://www.eevblog.com/forum/index.php?topic=293.0
https://www.eevblog.com/forum/index.php?topic=335.0
https://www.eevblog.com/forum/index.php?topic=945.0
https://www.eevblog.com/forum/index.php?topic=478.0 (i was in there fooling around)
https://www.eevblog.com/forum/index.php?topic=785.0 (esr discussion)
https://www.eevblog.com/forum/index.php?topic=1177.0 (while searching LM311)

and some useful links from there:
http://www.rfcandy.biz/communication/imp_lc.html (detailed explanation, very good reference, yet to read and register )
http://ironbark.bendigo.latrobe.edu.au/~rice/lc/index2.html (another variant, with HEX code)
http://www.todopic.com.ar/foros/index.php?topic=26840.0 (foreign language)
http://octopus.freeyellow.com/99.html (99cent esr)
http://electronics-diy.com/lc_meter.php (how many times do i have to look at this?)

[Mechatrommer]

sorry, try this:
https://www.eevblog.com/forum/index.php?topic=197.msg1892#msg1892

i think the guy clearly indicated his problem with
1) winXP only software
2) ni-visa
3) aborting installer
4) support service
5) and loads of complaints

i still in "reasonable doubt" that maybe the guy is not competent enuf to handle such situation?, or is this problem still relevant today? as he never mentioned the real working experience and crappiness with the software gui and features etc, but... well, i'll dig more deeper if i have those $600.

[Mechatrommer]

looking at the part LM311 datasheet, shattered my hope since i only have LM393. turned out LM311 to be something else.

[Mechatrommer]

ok. after making an exhaustive DG1022 search in eevblog. i come out with another issue, ie load impedance. it is said any FG/AWG will default to 50ohm load impedance (whatever it is means), so... the problem is... say i have the FG and feed it to the Vi RL/RC filter, how do i know if the Vi is actually feed by Vpp thats is display in the FG monitor? it could be 2xVdd? or something else that will make 0.707Vdd observation near impossible to get the correct L or C value. its explained in https://www.eevblog.com/forum/index.php?topic=1613.msg21950#msg21950 by BoredAtWork.

[tyblu]

The whole point of filters is to change its impedance at different frequencies. Attaching a 'scope probe at the input and output of your filter provides perfectly valid readings (for sub-RF waves; for sub-mm, you'll need to pay attention to measurement impedance). If you need to rely on the voltage output display on the FG to determine your load voltage, you must know the load at that frequency. 50-ohm impedance is a standardized number used across many high frequency devices (due to material constraints and history). It is possible to make matching networks for any load such that the FG output voltage will appear at the input of your circuit, but likely not necessary, as it is really the change in voltage from your circuit input to output one is interested in. If k-times the FG voltage appears at your circuit input and you need a 1:1 ratio and don't care about phase, then you can just dial the FG output knob to 1/k.

[Bored@Work]

or something else that will make 0.707Vdd observation near impossible to get the correct L or C value.

You make a relative measurement, by dividing the output voltage by the input voltage, whatever the input voltage is. If you are capable of measuring the output voltage you are for sure also capable of measuring the input voltage.

But still, these DIY meters with a 311 comperator as oscillator circuits are a viable alternative. All common parts and easy to build.