LCR/LCZ or Z analyser suggestions for low R coils

[ElectronRob]

Hi all,

I wonder if anyone is able to suggest either an LCR/LCZ or impedance analyser which would be well suited to measuring the impedance of low resistance coils.

The company I work for makes an instrument which relies on very high current flat PCB coils, the coils are typically in the order of 1x2inches total size, 1-10 turns or 4oz copper. These have a DC resistance of less that 0.1ohms and operate between 10Khz and 1Mhz.

Ultimately I would like to know the inductance of the coil so I can calculate coil current for various frequencies, we adjust voltage to limit peak current in order to stay within SOA of drive mosfets, due to the relatively large currents we pulse at (500-1500A) its quite tough getting a real measurement of current without effecting it.

Secondly I would be keen to have a display if impedance across the frequency range stated - this would let me turn this mystery component into something far simpler for general simulations. Ie, its very convenient to drop a resistor into a simulation rather than the complex LCR which we have in reality.

Any suggestions would be brilliant, I am mostly looking at late 80's - late 90's instruments with a budget of under £3k, would be keen for something HP/Agilent. An HP 4284A looks like it might fit the bill but having Z displayed would be convenient.

All the best
Rob

[ElectronRob]

Does anyone have a 4284A and could give a a review?

[bob91343]

How about an old fashioned Q meter?  It can measure Q of coils down to fairly low frequencies.

[TimFox]

If you are able to measure these coils with a suitable LCR meter or Q meter, you will find that the ESR is not constant with respect to frequency.  Spice simulations can't use a frequency-dependent resistor, so far as I know, so you will need to break the simulation down into frequency bands.  For low-resistance coils, the fixturing may be critical with any good meter.
A Keysight document  https://www.keysight.com/us/en/assets/7018-06715/technical-overviews/5952-1430.pdf  gives a table of LCR meters and similar, including their resolution.  Some go down to 2 m

[TurboTom]

An inductor designed to operate a those mentioned high currents will usually be equipped with a soft saturating core, so a measurement with any LCR tester that doesn't operate at these extremely high currents, or permits the application of a DC current of the required order of magnitude superimposed to the measurement signal, won't produce useful results. My approach would probably be to find a well-behaved capacitor (array?) of suitable electrical values, charge it to the required voltage, and use an ultra-low resitance bi-directional MOSFET switch assembly to connect the cpacitor to the DUT inductance. Analyzing the resulting dampened oscillation would reveal the corresponding parameters of the inductor, if required even at full operating current. This may necessitate some intricate mathematical methods, though.

[TimFox]

The original message said nothing about a core--it sounded like an air-core inductor.
ElectronRob:  please clarify.

[1audio]

You should be able to find a suitable current monitor here: https://www.pearsonelectronics.com/products/wideband-current-monitors They do show up on eBay as well. 
You can calculate the inductance if its as clearly defined as you have indicated. However if there are magnetics nearby that will affect it. Even a Nanovna might work, they go down to 50 KHz. You could also measure with some of the speaker testing setups using a soundcard and an amp up to just shy of 100 KHz, again pretty inexpensive. An Omicron Bode 100 https://www.omicron-lab.com/products/vector-network-analysis/bode-100  is well suited to this but a lot more money,

[bob91343]

The newer nanoVNAs go as low as 10 kHz.

[The Electrician]

Hi all,

Any suggestions would be brilliant, I am mostly looking at late 80's - late 90's instruments with a budget of under £3k, would be keen for something HP/Agilent. An HP 4284A looks like it might fit the bill but having Z displayed would be convenient.

All the best
Rob

Rather than buy a unit, why not rent a modern high-end impedance analyzer (for example, Keysight E4980A) for a month?  You should be able to get a lot of measurements done in a month.

[ElectronRob]

Hi guys,

Sorry for the late reply! I did not get any email notifications.

Tim - yes this is air core. Just reading through other points now.

[ElectronRob]

Thanks all,

Well, part of the issue is that I really like old test equipment so this is a fantastic excuse to buy more.

More realistically, I am dubious of the cheap VNAs but will look into them as option, the low resistance of the coils I think might be an issue for them.

Renting a unit might well be the most sensible way forwards.

1audio, thank you for the link to nice current monitors, this is something I have looked at briefly before - the main issue is their size, this is a PCB coil, its physically small with fets within 10mm of the coil terminals and a capacitor bank across the fets. Any increase to length/inductance decreases the peak current quite substantially. 

I am able to solder a wire to the coil to divert a small proportion of the current which allows me to at least look at the current waveform using my trusty P6021.

Tim, thank you for note regarding ESR changing with frequency, my plan is to set up the sim for each frequency rather than to sweep through as you mention. This is one of the main reasons for using an LCZ as I can test the coil at say 100khz and compare the sim at 100khz.

I have one of these in a draw somewhere: http://www.cypherinstruments.co.uk but again would be dubious of the results. I'll give it a go and see if I get anything sensible from it.

Bob - I am not familiar with Q meters so will do some reading thank you.

Thank you for the thoughts all and apologies for the disjointed reply - hope it makes sense.

Rob

[TimFox]

A traditional Q meter has an excitation transformer to give an extremely low source resistance to the device under test, connected between the secondary and a high-Q variable capacitor.  Q and L are measured measured by tuning the capacitor for maximum voltage across it at a chosen fixed frequency.  The frequencies chosen for the original Boonton design live on as the standard frequencies for inductor specifications.

[1audio]

The currents you mention are really higher than a P6021 is linear for. However you may be able to carefully measure the resistance of the 10 mm of trace and calculate its inductance and then measure differentially across it to get a current reading. You would need a good differential probe. Maybe a P6046 would work and feed the vintage habit.

Let us all know what works to add to our bags of tricks.

[ElectronRob]

Thanks Tim,

Really interesting reading and have found a few units in the UK - I'm not convinced it will be as useful as an LCR/LCZ however they are cheap and impressive.

[ElectronRob]

The P6021 is indeed not at all suitable, it would likely be immediately damaged if it saw full coil current regardless of the low duty.

However - if you divert a tiny proportion of the current using a thin wire and place the clamp around that its fine - making sure you respect the limits of the probe.

Obviously you cannot make good judgements on absolute current but you can see relative changes and explore current shape.

Funnily enough I have a few diff probes including a P6046, and a number of nice agilent probes, the coil is connected to an Hbridge and its absolutely impossible in my opinion to work efficiently with an Hbridge without a at least two diff probes. I have a draw full of my own home brew 50Mhz probes too which are useful for this. 

[ElectronRob]

Thanks Tom,

However I'm not sure if this is better than simply placing a capacitor in parallel with the coil and sweeping with a frequency generator....

[The Electrician]

Thanks Tom,

However I'm not sure if this is better than simply placing a capacitor in parallel with the coil and sweeping with a frequency generator....

What is the cross sectional area of the trace that makes up the inductor?  You say the DC resistance is less than .1 ohm.  If you're passing ~1000 amps through it, the time of that pulse of current must be rather short to avoid blowing the trace off the board.  Just how long is it?