LCR Cal Test PCB

[KungFuJosh]

I want to reproduce this as soon as possible.

You'll hopefully receive your PCB next week or the week after. Mike should have his (and my tested board) on Saturday.

[mawyatt]

Here's the results comparing my ST2832 and ST42 on V5 of the test board. Most tests were matched at 360mVrms (ST42 tested at about 358mVrms).

The DC bias values are based on 1kHz or 10kHz frequency for AC. It matters for some reason. 🤷

I think I did step DF test at 1Vrms. I forget, that was way back to yesterday.

Thanks,
Josh

Use your SDM3065X to measure the Thermisistor (2 wire), it should yield ~78F or ~25.6C.

Best,

[KungFuJosh]

Use your SDM3065X to measure the Thermisistor (2 wire), it should yield ~78F or ~25.6C.

Oops. A little late now, it's on the way to you. I'll try it out next time. 😉

[mawyatt]

No problem, we'll measure it when the PCB arrives!!

Best,

[Martin72]

You'll hopefully receive your PCB next week or the week after. Mike should have his (and my tested board) on Saturday.

It's not ready yet, but Mouser delivered today and everything has arrived.
I ordered everything from your parts list 2 times, you never know how stupid you can be when soldering.

[KungFuJosh]

It's not ready yet, but Mouser delivered today and everything has arrived.
I ordered everything from your parts list 2 times, you never know how stupid you can be when soldering.

I always get at least 3 for the same reason...except even stupider: sometimes I don't pay enough attention when I test SMT stuff, and the parts might become airborne. 🤣 ...only lost 2 or 3 items on the last board I made.

[mawyatt]

PCB received, and testing underway. PCB looks nice

See overhear for some discussions on LCR sensitive measurements.

https://www.eevblog.com/forum/testgear/sensitive-lcr-meter-measurements/

Best,

[KungFuJosh]

Man, I want one of those Hioki LCR meters. Gimme that 8MHz! 😉

[mawyatt]

They are good, real good in every way (except fan noise). Why we have one

They operate differently than most tho, and takes awhile to get used to.

Instead on bashing the operation we spent time learning to use it, and discovered that the way it works is to maximize measurement accuracy and speed. When you get to doing lots of measurements, you begin to appreciate the thought that went into this fine instrument, and really like the ability to select 4 parameters for display and easily change each without fiddling thru a bunch of menus. In general you don't need to navigate a bunch of menus, most things you want to do are just one touch away.

Now that we've become somewhat proficient in use, it's value has increased

However, the fan noise is horrific

Best,

[mawyatt]

Should add that the IM3536 can deliver 5VRMS max and 0.1ARMS max, both verified by DMM and DSO, of course not simultaneously and over certain frequency ranges

Best,

[Martin72]

Yes, I've also had my eye on your Hioki.
But when the time comes, I'll opt for an LCX100 from R&S.

[mawyatt]

We looked at R&S, Keysight, Tonghui and Instek, but decided on the Hioki IM3536 based upon what we needed and the $ value. Only serious issue was with the lack of internal graphing display capability, we solved that by creating our own custom RPi/PC resident plotting routines which have served well.

Of course YMMV,

Best,

[KungFuJosh]

The LCX100 might be comparable cash value if it can be improved the fun way.

[Martin72]

Only serious issue was with the lack of internal graphing display capability,

The LCX has this feature (optional), but whether this is “fun” given the screen size is another question.
In contrast to the Hioki and R&S, the ET35 has a “huge” 7” screen - but only displays numbers on it....
But we digress.

[mawyatt]

Only serious issue was with the lack of internal graphing display capability,

The LCX has this feature (optional), but whether this is “fun” given the screen size is another question.
In contrast to the Hioki and R&S, the ET35 has a “huge” 7” screen - but only displays numbers on it....
But we digress.

Maybe someone can conjure up some code for the display on the ET35

Best,

[Martin72]

Look what the post has brought...
I have most of the components, now it's time to solder - and then measure...

[KungFuJosh]

Look what the post has brought...
I have most of the components, now it's time to solder - and then measure...

Nice! Are you going to use the terminal pins? I'm curious how that works out if you do.

[Martin72]

You've left enough space for me to have both options open, so yes, it will be fitted and I'm curious to see if there are any differences later on.

[Martin72]

I had soldered the coils on at work today, then quickly measured them with the calibrated ST2830.
The results were rather mixed, especially with the smaller values.
At home, a quick check with the ST42 was the same.
I've just got the datasheets available, which show the frequencies used to determine the values.

10nH   450Mhz
100nH 25Mhz
1µH  10Mhz
10µH 2Mhz
100µH 1Mhz
And always with an SMD test fixture and either a network or impedance analyzer.
We are miles away from that with our LCR meters.
I would leave out the 10nH and 100nH and, if possible, not select any HF components for the values above 1µH.

[KungFuJosh]

I'll try finding parts that used lower frequencies to test. At least under, let's a pick a random number, and say 5MHz. 😉

[mawyatt]

Small inductances at low frequencies will be very difficult to get reliable and repeatable results, reason is impedance Z is j*omega*L, where omega is 2*pi*F. So 10nH at 100KHz is only ~6.3 milliohms Z, at 1MHz a little better at ~63 milliohms and at 5MHz ~0.314 ohms.

For an instrument with 100KHz limit, seems 0.1 or 1uH is a reasonable lower limit with a Z of 0.0628 or 0.628 ohms respectively, which should be measured with some fidelity with a quality LCR meter.

One must remember that the LCR meter doesn't directly measure Inductance (or Capacitance for that matter), it measures the voltage across and the current thru the DUT, then computes the impedance Z at frequency, then the equivalent Inductance (or Capacitance) to produce that impedance.

With low impedances the setup and fixture as well as DUT contact impedance play a significant roll in the readings, the meter's Short Calibration attempts to remove these parasitic series impedance effects, but never quite prefect. Such a low DUT Z, the meter will supply a large current (usually the larger the better) to produce a higher DUT voltage reading, but limits come into play, especially with battery powered Handhelds (Tweezers).

Likewise high DUT impedances can be troublesome, such as small capacitances. Here parasitic shunt capacitances becomes the culprit and the meter attempts to remove this effect with an Open Calibration, effective but never perfect. Operating with a higher voltage and frequency increases the DUT current which helps but meter limits also come into play, usually favoring bench types over Handhelds (Tweezers).

Kelvin Clips just complicates things with the long cables, which can influence readings, favoring lower frequencies, and why direct connected DUT fixtures are generally preferred for both low and high Z components at moderate to higher frequencies.

Anyway, the trials and tribulations one goes thru for accurate repeatable DUT measurements with LCR Meters (any meters for that matter). As can be seen one Meter does not serve all DUTs, why we have the TH2830, IM3536, DER-5000 and 3 VNAs

Not to mention the various DUT fixtures, both OEM type and DIY including modded OEM types, all for the sake of mostly passive component measurements, of course YMMV   

Best,

[KungFuJosh]

I wish I looked at the test frequencies before making the BOM.

It looks like 450MHz is pretty common for testing 10nH, so I'll just skip that.

However, it looks like there are some options for 100nH at 100kHz. For example: FP0805R1-R10-R.

[Martin72]

@mawyatt:
It is exactly as you write it and I really hope that the people who are interested in it are aware that in the end only current and voltage are measured.
For example, we manufacture most of the required chokes/transformers ourselves.
They are not only measured but also calibrated (air gap).
This is solved in various ways, not only or rather rarely with a “simple” measurement with an LCR.
Mostly by means of a resonant circuit or with the voltage-time surface method, where real current can then flow (measuring an output filter choke with a few µ...mA makes almost no sense, so it is better to send a pulse with a rated current of up to 200A through the choke).
If a test board is now developed, as here, which is intended for normal LCR measuring bridges in the affordable range, then the references on the board must first be measurable with such devices.
If it is only about the pure L-value, things like the quality Q or the RDC hardly play a role.
And since the measuring current is small anyway, the core can also be “crap”, you won't reach saturation anyway.

[Martin72]

I ordered the components from Mouser according to your BOM.
And Mouser has a datasheet for each component, which also shows how the values were determined.

[KungFuJosh]

I ordered the components from Mouser according to your BOM.
And Mouser has a datasheet for each component, which also shows how the values were determined.

I checked the datasheets on Mouser, up to 1uF and they didn't show any test frequencies. The inductors were easy to find.