Measurements in the nanohenry range. Tweezer FNIRSI ST2 and others.

[iet]

An interesting and demanding range. It's also largely unstable, depending on a number of factors.
After several months of working with the ST2 tweezers, I've accumulated some questions about measurements in this range. Perhaps someone already has answers to my questions.
I noticed significant instability in LCR mode when measuring L at 100 kHz.
1.Is this a flaw in the measurement circuit for this range? Or is there another reason?
In SCAN mode, measurements are much more stable, and the results are generally consistent with other LCR meters at 100 kHz.
2.What could be the difference between the two measurement modes? What determines this?

[shabaz]

I think you wasted several months of effort. What was the point? Was it to design nH inductors to a specific specification? If so, then you were using the wrong tool for the job. A VNA would have given you a more useful set of results, and they don't cost a lot nowadays.
Was it to sort inductors based on value? If so, that seems a waste of time because one can just buy inductors with a tight tolerance (e.g. 2.5%).

Was it for hobby circuit construction? Again those months of effort make no sense because one can self assemble nH inductors based on dimensions, and/or wound with very accurate calculator tools, if a ready-made inductor is not available. Furthermore, precise values are often not required, since circuits are tuned and the output is then observed with other instruments (like an oscilloscope).

[mtwieg]

Are you trying to measure inductances in the nanohenry range (i.e. differentiate a 5.6nH part from 6.8nH), or are you trying to measure inductors with nanohenry precision (i.e. differentiate 150nH from 160nH)?

Doing either with a test frequency of only 100kHz would be quite difficult. The reactance of the DUT would be on the order of milliohms, likely far less than the ESR in the part and the test leads.

[iet]

What was the point?

Inductors in this range can be found on various boards, but they are not marked. Therefore, these inductors must be measured before desoldering them from the boards.

[iet]

Are you trying to measure inductances in the nanohenry range (i.e. differentiate a 5.6nH part from 6.8nH), or are you trying to measure inductors with nanohenry precision (i.e. differentiate 150nH from 160nH)?

The range of <100 nH is of interest to me.

[chilternview]

The range of <100 nH is of interest to me.

Do the arithmetic. At 100kHz, an inductor of 100nH will have an impedance of jwL = 0.0628 ohms.
You'll be measuring noise with your tweezers.

[shabaz]

What was the point?

Inductors in this range can be found on various boards, but they are not marked. Therefore, these inductors must be measured before desoldering them from the boards.

I believe it may be worth digging into your actual end goal, to see if your actual requirement is something else.

For what end reason are you doing this? Board repair for unknown boards? The applications a lot of those nanohenry inductors are used for, are extremely low current and high frequency. They are unlikely to burn out. What boards are you repairing where these are possibly cracked and need replacing, in a circuit where you cannot find a datasheet to see the required ballpark inductance? Is is RF related (e.g. antenna matching?). I just don't get why you'd spend months on this problem that appears to have little practical purpose.

[iet]

Do the arithmetic.

I use a calculator when I need to calculate, for example, the inductance of a piece of wire. Where inductance can be measured with a device, a calculator is not needed.

[mtwieg]

Inductors in this range can be found on various boards, but they are not marked. Therefore, these inductors must be measured before desoldering them from the boards.

The range of <100 nH is of interest to me.

I'm wondering if these "inductors" are actually ferrite beads (I would not expect to find actual inductors in that range on most boards, unless the board has RF filters or something).

If so, I don't think measuring their impedance at 100kHz is sufficient to identify specific parts for repair or reverse engineering purposes. The most important behavior is usually at frequencies between 1MHz-1GHz. Two ferrite beads might have identical impedance at 100kHz yet be completely different at 100MHz, which may be critical for the circuits using them. Then there's the issue of impedance vs DC current, which the tweezers cannot tell you...

[iet]

I just don't get why you'd spend months on this problem that appears to have little practical purpose.

On the contrary, it makes a lot of practical sense. The ST2 tweezers were very helpful when winding transformers for my ESR meters.

[shabaz]

I just don't get why you'd spend months on this problem that appears to have little practical purpose.

On the contrary, it makes a lot of practical sense. The ST2 tweezers were very helpful when winding transformers for my ESR meters.

Does that mean your goal is not measuring nH inductances for board repair, but creating hobby circuits such as https://www.eevblog.com/forum/projects/super-simple-esr-meter-with-a-resolution-of-10-mohms/ ?

Do you think there could have been a better way to approach that (especially given how frustrated people were that you didn't provide any inductances or core specifics for a month).

How do you think you'd construct such inductances, if say the tool you had didn't measure below 100 nH (or didn't measure with any accuracy there)?

[iet]

I'm wondering if these "inductors" are actually ferrite beads (I would not expect to find actual inductors in that range on most boards, unless the board has RF filters or something).

These are standard SMD inductors of the corresponding dimensions. The maximum measurement frequency of the tweezers is 100 kHz. Therefore, all measurements of such inductors are performed at this frequency.
Inductor 5 nH-0402.

[iet]

Do you think there could have been a better way to approach that (especially given how frustrated people were that you didn't provide any inductances or core specifics for a month).

The thread contains all the necessary information about the transformer. You can build and configure an ESR meter in half a day. And an inductance meter significantly speeds up all this work.

[shabaz]

..an inductance meter significantly speeds up all this work.

No it didn't. After two weeks you stated that a gentleman who applied a little math to work out the actual inductance, was "a little ahead of you".

Those EMI rings are either iron powder or ferrite ones.

With OP's 110turns resonating at ~100kHz with 330pF the AL of the material should be around 640nH/turns^2 (that is close to the ferrite material 43).

L = 110turns^2 * 640nH/turns^2 = aprox 7.7mH
7.7mH in parallel with 330pF = aprox 100kHz (not counting the parasitics)

PS: those oscillators have got the cold side winding (III) much lower number of turns than the hot side one (II - where the base is connected) usually. Not sure with the above one (it has 100turns on the cold side), it could be for some special purpose, however.

You are a little ahead of me. The information on the transformer is not yet fully ready. It is necessary to find time for additional tests, and there is very little of it.
Here the frequency turned out to be a little higher - 150 kHz. The inductance of the winding III-3.4 mH.
On the second board I used a dark gray ferrite ring 12x4x6. Information about it was lost, now I am trying to restore it.

[iet]

You can write all this in that thread and explore the entire chronology of events over the past year. I'll answer everything you're interested in there. Although this could have already happened during the year, strangely enough, for some reason it didn't.

[shabaz]

You can write all this in that thread and explore the entire chronology of events over the past year. I'll answer everything you're interested in there. Although this could have already happened during the year, strangely enough, for some reason it didn't.

You brought it up, not me. You picked the example in this thread - it's not my fault that your example demonstrates the opposite to what you intended.

The coil specifics were brought up last year many times.. can you not see? People there were wondering how to assemble it for a month or more, because the transformer details from you were scant, and it eventually took someone to work it out for themselves.. surprise surprise, without tweezers.

[iet]

You've confused yourself and others. This example doesn't relate to the topic you mentioned.
And to finish discussing other topics here.
The quoted participant in that thread, after many calculations, didn't get the desired results.
And I think we can end the discussion of other topics here.

[shabaz]

You've confused yourself and others. This example doesn't relate to the topic you mentioned.

In that case, you have my apologies, and go on then, the floor is all yours, explain why nH measurements from those tweezers were "very helpful" for the transformer in the photo, which you say was used for a different ESR meter?

On the contrary, it makes a lot of practical sense. The ST2 tweezers were very helpful when winding transformers for my ESR meters.

[iet]

When measuring such low values, I try to verify the measurement results with other devices, if possible. Then such values ​​can be trusted.

[shabaz]

When measuring such low values, I try to verify the measurement results with other devices, if possible. Then such values ​​can be trusted.

Actually, sticking with your example that you raised, you just added a pointless measurement, because you had to measure the frequency of oscillation in-circuit anyway, with an oscilloscope.

Even if the tweezers provided any accuracy at such low nH values, you would have still had to measure in-circuit, to include all parasitics. Only then could your circuit have been trusted to operate at the frequency you required for ESR measurement.

I should also add, that you would have got a result precise down to the minimum individual turn that you could have wound, if you'd done the calculations on the core, so any frequency adjustments would have been by slight movement of the turns, or trimming the capacitor; again with an oscilloscope, not nH out-of-circuit measurements.

[iet]

I tested this with another device, using a full four-wire Kelvin circuit. Everything is fine. These three results are reliable.
The inductance of this inductor is 5 nH.

[iet]

At first, I thought the reason for the different tweezers behavior in L mode was due to the measurement signal. But then a good oscilloscope arrived. It turned out that in LCR and SCAN modes, everything is super stable—100 kHz sine.

[mtwieg]

A 5nH measurement isn't really believable with such an instrument, at any test frequency.

Consider that 5nH is the inductance of a centimeter-long straight wire in free space, or a 5mm diameter loop of wire. The path of the test leads (whether using kelvin clips or the tweezer arms) would cause far more variation than 5nH. Only way to get such repeatable measurements is for the connections to be completely rigid during both DUT measurement and SOL calibration.

[iet]

A 5nH measurement isn't really believable with such an instrument, at any test frequency.

And what's the minimum inductance that LCR tweezers can or should measure?

[Grandchuck]

Here is the error/accuracy chart for the East Tester ET4401 LCR meter.