TH2830 vs TH2832

[KungFuJosh]

That's a cool idea, but this board suits my needs just fine. I think if I was going to do that, rather than bulky BNCs, I would put top / bottom layer pads that kelvin clips would work with. Then the single component boards could be much smaller. Wouldn't that be fine with 2 layers?

Did you see my results PDF on reply 235? I didn't include the 10nH because nothing could read it anyway. 100nH was fine, 10pF was fine, everything else worked in the expected ranges.

The biggest issues I saw were: ~700mH inductor at 10kHz; 100Ω DCR; and 100kΩ at 200kHz. 100kΩ at 100kHz wasn't great either, but not as bad.

A smaller issue was maybe 100mH at 10kHz, but that and other smaller variances aren't as meaningful without a real calibration done beforehand.

[coromonadalix]

i think this thread is derailing a bit,  we talk about lcr meter mods between 2 models, for sure test(s) need to be done

for the test jigs and discutions,  they should be posted in the relevant thread(s)

[zrq]

Briefly looked at the strings in the firmware, unfortunately I found no reference to the factory calibration routine.
Another question: when you were testing the resistor measurements and found they are off, did you do the full 201 point correction before?
BTW: on Tonghui's Chinese website, there is a new model TH2832D that goes up to 300 kHz listed.

[mawyatt]

Think the idea of putting anything like a small SMD cap, resistor or inductor on a FR4/Epoxy PCB to use as some "Reference" is unwise. PCBs of FR4/Epoxy or similar material change with humidity, temperature, time, handling and so on, some PCB materials are better tho, and maybe those would be acceptable.

Maybe others can comment about the different PCB materials, and the relative behavior for supporting precision SMD devices.

As mentioned above every mm counts when measuring nH, and that any equipment/cabling moves can cause measurement uncertainty, including Kelvin cables. This is why we don't use Kelvin cables for precision measurements of SMDs, even the Tweezer Kelvin types. We go so far as to do a Open Cal with an LCR SMD fixture by placing the plunger opening distance equal to the expected DUT SMD spacing, this helps capture the fringe capacitance with the plungers at the expected spacing, thus minimizing the fringe capacitance effects on the DUT measurements.

Along this line, for the Short Cal, we often use a Zero Ohm SMD of the same size/type as the expected DUT, altho not certain this is beneficial in all cases. The thinking here is that the Zero Ohm device helps compensate for the expected contact area of the SDM DUT as well as the plunger spacing and tension.

The SMD fixture plunger spacing is an interesting thought regarding inductance measurements (as well as the mentioned capacitance measurements). Since inductance is associated with length and the fixture plunger has a diameter of ~1mm (fixtures we have) and varies in length when engaging a SMD. Then a 1206 SMD that was considered a round "wire" has a self inductance of ~1.14nH (Grover's) (1.0938nH), an 0603 0.40nH (0.4983nH), an 0402 0.22nH (0.3059nH) (Edit: found a flat wire inductance estimate). So it seems that when measuring small SMD inductances doing the Short Cal with a Zero Ohm SMD might be useful.

Disclaimer we have no way of knowing if these inductors are accurate per OEMs data sheet, and did a very crude test with a small SMD inductor, 0603 10nH +-0.2nH (Wurth WE-MK type). We see a difference in the ball park as shown by Grover's estimate above relative to if the Short Cal is performed with a Zero Ohm SMD device or not.

Later if time permits we may expand on this crude test, but seems that the concept of utilizing a similar size Zero Ohm SMD for the Short Cal with a SMD fixture is worthwhile for small inductors, and spacing the plungers at the proper spacing for the Open Cal is also beneficial for small capacitances. Does anyone know if "Zero Capacitance" SMD devices exist? We've thought about just using very high resistance SMDs of various sizes for the "Zero Capacitance" equivalent, but haven't do so yet.

Best,

[KungFuJosh]

Briefly looked at the strings in the firmware, unfortunately I found no reference to the factory calibration routine.
Another question: when you were testing the resistor measurements and found they are off, did you do the full 201 point correction before?
BTW: on Tonghui's Chinese website, there is a new model TH2832D that goes up to 300 kHz listed.

I did the regular open/close + DCR corrections. There's no option for how many points except if you add manual spot correction.

2832D? Hmm, I wonder if that hardware is any different. 😉🤣

[luudee]

I did the regular open/close + DCR corrections. There's no option for how many points except if you add manual spot correction.

2832D? Hmm, I wonder if that hardware is any different. 😉🤣

Not in the office at the moment, but I am pretty sure that I saw something like "correction lists" ....

Will check in 12h !


luudee

[mawyatt]

Think the idea of putting anything like a small SMD cap, resistor or inductor on a FR4/Epoxy PCB to use as some "Reference" is unwise. PCBs of FR4/Epoxy or similar material change with humidity, temperature, time, handling and so on, some PCB materials are better tho, and maybe those would be acceptable.

Maybe others can comment about the different PCB materials, and the relative behavior for supporting precision SMD devices.

As mentioned above every mm counts when measuring nH, and that any equipment/cabling moves can cause measurement uncertainty, including Kelvin cables. This is why we don't use Kelvin cables for precision measurements of SMDs, even the Tweezer Kelvin types. We go so far as to do a Open Cal with an LCR SMD fixture by placing the plunger opening distance equal to the expected DUT SMD spacing, this helps capture the fringe capacitance with the plungers at the expected spacing, thus minimizing the fringe capacitance effects on the DUT measurements.

Along this line, for the Short Cal, we often use a Zero Ohm SMD of the same size/type as the expected DUT, altho not certain this is beneficial in all cases. The thinking here is that the Zero Ohm device helps compensate for the expected contact area of the SDM DUT as well as the plunger spacing and tension.

The SMD fixture plunger spacing is an interesting thought regarding inductance measurements (as well as the mentioned capacitance measurements). Since inductance is associated with length and the fixture plunger has a diameter of ~1mm (fixtures we have) and varies in length when engaging a SMD. Then a 1206 SMD that was considered a round "wire" has a self inductance of ~1.14nH (Grover's), an 0603 0.40nH, an 0402 0.22nH. So it seems that when measuring small SMD inductances doing the Short Cal with a Zero Ohm SMD might be useful.

Disclaimer we have no way of knowing if these inductors are accurate per OEMs data sheet, and did a very crude test with a small SMD inductor, 0603 10nH +-0.2nH (Wurth WE-MK type). We see a difference in the ball park as shown by Grover's estimate above relative to if the Short Cal is performed with a Zero Ohm SMD device or not.

Later if time permits we may expand on this crude test, but seems that the concept of utilizing a similar size Zero Ohm SMD for the Short Cal with a SMD fixture is worthwhile for small inductors, and spacing the plungers at the proper spacing for the Open Cal is also beneficial for small capacitances. Does anyone know if "Zero Capacitance" SMD devices exist? We've thought about just using very high resistance SMDs of various sizes for the "Zero Capacitance" equivalent, but haven't do so yet.

Best,

We just measured (Hioki IM3536) an 0603 10nH Wurth WE-MK mentioned above, the measured valued was 11.71nH when using the standard Short Cal without Zero Ohm Shunt with the SMD fixture, and with a Zero Ohm Shunt device 10.71nH. These were measured at 1MHz, so we could compare with data sheet.

Also at 100KHz so we could use the TH2830, the IM3536 at 100KHz 11.71nH without and 10.75nH with Zero Ohm Shunt used for Short Cal, the TH2830 at 100KHz produced 11.88nH without and 10.82 with Shunt.

So looks as if the Short Cal use of a Zero Ohm Shunt SMD of the same size as the inductor DUT for small inductance values has merit 

Best

[KungFuJosh]

Now you're making me want to try the TH2830 on the board and see if the failures in that respect are the same or not. 🤔

[KungFuJosh]

I did some comparison tests with the TH2830, and the ST2832. Except for a couple failures from the ST2832, it actually outperformed the TH2830 firmware.

The largest failure with ST2832 was the 100Ω DCR measurement:


The second biggest fail was the 700mH inductor. According to their accuracy chart, between 100mH to nearly 1H should have excellent accuracy between 1k to 10kHz.

[KungFuJosh]

If somebody with a 2830 wants to test, the failures on mine were:
100uF was lower than expected (100uF read as 83uF; 2832 saw it as 90uF).
100nH read as 100.045 on the ST2832, and read as 180.321 on the TH2830.

[Martin72]

Hi,

I still can´t read this kind of diagrams, but measuring hundreds of milli-henry with a frequency between 1 and 10khz ?

If somebody with a 2830 wants to test

100µF I could test on thursday with our ST2830, nH I don´t know if we have something like this on stock.

[mawyatt]

Hi,

I still can´t read this kind of diagrams, but measuring hundreds of milli-henry with a frequency between 1 and 10khz ?

If somebody with a 2830 wants to test

100µF I could test on thursday with our ST2830, nH I don´t know if we have something like this on stock.

If referring to the Tonghui LCR diagrams shown, yes they can be confusing, most all LCR meters use them tho. You should see the entire chapter devoted to this in the Hioki IM3536 user manual, it has those graphs and page after page on getting an estimate of the relative measurement accuracy 

One must remember that these LCR meters measure the VECTOR voltage across the DUT and the VECTOR current thru the DUT, all the various selectable parameters are derived from these Vector measurements.

If you have an LCR meter that displays the actual DUT measured voltage and current (ours display only magnitudes), this can help in judging if the measurement at hand is within acceptable ranges, some meters also display a message when the measurement parameters are in question. Basically if the voltage measurements are very low then the measurement should be questioned, same for DUT current.

Edit: Note, this is where using the appropriate LCR drive impedance and output drive voltage level can help, the objective is to produce a high enough voltage level across the DUT for good measurement and have enough current flow for a good measurement, but not so high the voltage of the DUT is almost as high as the drive voltage, nor so low as to susceptible to noise. Best to just play around with the various available settings to get familiar with how they influence readings.


Anyway, we find this displayed voltage and current levels very helpful when making measurements and not having to revert to those confusing graphs

Best, 

[Martin72]

You should see the entire chapter devoted to this in the Hioki IM3536 user manual, it has those graphs and page after page on getting an estimate of the relative measurement accuracy

Same here with my hioki....

If you have an LCR meter that displays the actual DUT measured voltage and current

Voltage I must have a look again, but I know my hioki shows the actual current flowing.
Oh I wish to have it fixed (Backlight, Touchpanel)so I can use it more for comparison in threads like this one here.

[KungFuJosh]

I still can´t read this kind of diagrams, but measuring hundreds of milli-henry with a frequency between 1 and 10khz ?

Vertical lines go from the bottom to the top to tell you the test frequency capability.

The dark outline boxes in the middle are percentage accuracy points. The center box is the highest accuracy, the sweet spot, giving 0.05% to 0.1% accuracy. Every step out from the center is less and less accurate, the worst being from 0.65%  to 1%.

All of the diagonal lines represent the DUT value. I highlighted the value range of 100mH to 1H, and crossed over it with the test frequency of 1kHz to 10kHz. This shows the DUTs in that range, tested at that frequency range, should expect the highest level of accuracy the meter is capable of.

[KungFuJosh]

Anyway, we find this displayed voltage and current levels very helpful when making measurements and not having to revert to those confusing graphs

Per your advice, I always leave the Vm/Im displayed on, but that doesn't display on DCR tests.

I only learned to read those charts a couple weeks ago, but they're pretty darn helpful now that I understand. 😅

[mawyatt]

Looks as if the Impedance Magnitude sweep sweet spot is between 15 ohms and 300K from 100Hz to 2KHz, then tapers from 300K to 2K ohms at 30KHz. So one should try and get Vm/Im within the range if possible for best results.

BTW the chart has an error, the upper right hand corner should read 10fF, not 10pF!!

Should note that these LCR meters have a 1 ohm DCR and 3 ohm ACR ranges, and can deliver 66.7ma of DUT current, very handy for low Z value measurements.

Best,

[luudee]

Hi Guys,

as promised, I redid the resistor tests with wire-wound 0.1% resistors.

The table should be self-explanatory.

The LCR meter and my DMM, are usually capable of showing "many" digits
after the comma. If I show a limited number of digits, it's because the next
digit was jumping back and forth and could not decide on it's value.



I used the TH26048A test fixture that came with the meter, not the cables ...

Cheers,
luudee

[KungFuJosh]

HOW DO I MAKE THE IMAGE APPEAR IN MY MESSAGE

Nice results. I hope mine can do that soon. 😉

For images, it's annoying, but here's what you have to do:
Post your message, then expand the image below the posted message.
Copy the URL of the expanded image (or open the image in a new tab, and copy the URL from there).
Then Modify your message, and click the "Insert Image" button that's probably directly beneath the Bold button, on the left. Then paste the copied image URL between the image tags, and save your changes.

[luudee]

Fixed a few typos in my table above ... should be all correct now

If you compare it to the accuracy chart, it matched quite well ....

Cheers,
luudee

[mawyatt]

One of the benefits (not usually mentioned) of these LCR meters is the lower Z ranges, like DCR at 1 Ohm and ACR 3 Ohm ranges, where the DUT current can be as high as ~67ma. These ranges are below most quality DMMs like the KS34465A (100 ohms) and DMM6500 (10 ohms).

Best,

[KungFuJosh]

If you compare it to the accuracy chart, it matched quite well ....

Yes, it would be nice if you could acquire the firmware so others could try it. 😉

I'm curious, can you try inductance test in the 600mH to 1H range? I'd like to see how accurate/stable it is compared to my not-so-great results. If you could compare at least 1kHz and 10kHz results, that would be awesome.

Thanks,
Josh

[KungFuJosh]

I pulled a 10nH inductor off a 2nd test board, and tested it in my SMT fixture. It worked pretty well! 12nH is pretty close considering the generic test fixture, and I don't know the actual value. 😉

I had an 0603 10nH inductor also, but it was too small for my fixture to work well.

[Martin72]

Hi,

Bias on ? Why ?

[KungFuJosh]

Bias on ? Why ?

It's not on. It's set at 1.5V, but it's only on when the external BIAS button is lit up red...which you can't see in a screenshot. 😉

[Martin72]

Ahhh...

12 instead of 10nH is good, better as I expected with such a "low" testfrequency.
jXL is about 6mOhm - What Rdc have the coil ?