East Tester ET3502 LCR-Bridge

[w.v.s.]

Hmm, I would also have expected something around 200mA, if these are the levels directly at the DUT.

[Martin72]

Hi,
I can't understand it at the moment either, maybe I can think of something else.
Or I'll just ask ET.

[KungFuJosh]

Nah, just more proof ST283X > ET35. 😉😉

[Martin72]

Hm?
I don't really understand that one now.

[egonotto]

Hm?
I don't really understand that one now.

Hello,

if you measure the voltage at the DUT, you would at least know what is really going on.

Best regards
egonotto

[mawyatt]

I took a few more pictures.
It doesn't look too bad.

Martin,

That's the kind of SMD Fixture that can cause issues!

It has no internal shield partition between the H and L sections. This will cause errors with small Cap measurements (high Z) as the internal fields change with plunger location on the moving end. This will be aggravated by the ET35 since its H and L sections are reversed which means the moving plunger end is the "hot" H end whereas the normal LCR Meter orientation has the moving plunger on the "cold" L side.

Also we found this type with the two end screws on the cover use a thinner aluminum case/cover and powder coat paint the entire case and cover without any masking. This may cause less than ideal shielding between the cover and case which makes contact thru just the 2 end screws.

The other (better) type SMD Fixture we have has 4 screws to hold the cover and uses threaded inserts in the cover.

The cover you have looks like it might be sturdy enough tho, might want to ream the holes in the cover to remove the paint coating and get good contact with the 2 screws to case.

The case looks as if it's only relying on the BNC washers to puncture the paint coating to tie to the cover, this may require removing the BNCs and cleaning away the paint coating around the BNC holes for better contact.

Also check the Plunger outer cases, these form a somewhat coaxial conductor with the plunger, you want these to be firmly connected to the cover which must make a good connection the case to get back to the BNCs.

Remember to check the brass threaded rods that make contact with the plungers, ours arrived loose and it took awhile before we discovered this "surprise"!!!

Anyway, looks like you might have some reworking ahead to get this SMD Fixture up to standards for precision LCR measurements with the ET35!!! After getting things in order our modified SMD fixture has good repeatability even up to 8MHz, so these can be "coached" with a little effort into a good fixture for precision measurements.

Keep us posted on your progress.

BTW are you sure you have a cold, or just "skipping work" to have time to play around with this stuff

Best,

[Martin72]

BTW are you sure you have a cold, or just "skipping work" to have time to play around with this stuff

Unfortunately not, but if at all possible tomorrow I want to go to work - just to measure the ST2830...

[Martin72]

This doesn't seem correct. The DUT Voltage and Current imply ~90nH not ~10nH.

Wonder if the fixture is causing problems??

Best,

After egonotto noticed this, I looked into it again.
I'm still not entirely convinced that the values are wrong, I'll hopefully take the same choke with the fixture to work tomorrow and connect it to the ST2830.
Then record the values at the maximum possible frequency of 100kHz and measure them at home on the ET35.
I also measured a capacitor earlier, like you did on your TH2830 a page earlier.
However, 10pF, because I didn't have a 4.7pF.
The result itself is plausible, but you can't derive it from the monitor voltage/current.
I'll measure that tomorrow too, if I'm fit enough.
And it's supposed to get really warm tomorrow of all days...

[Martin72]

That's the kind of SMD Fixture that can cause issues!

The only question is how large these can be with a (current) maximum frequency of 300 kHz.
That's almost audio...
From an HF point of view, this fixation is simply garbage.
And not only this fixation, others too.
They are all not "HF-tight" ....
I already have ideas on how to change this, I will also order the ST fixture for our test field as announced - and "of course" take it apart for optical inspection....

[mawyatt]

Got a Wurth RF 0603 Chip 10nH Inductor & Pulse Electronics with Standard SMD fixture and measured with IM3536 at 300KHz and 100KHz. Readings are in nH (Vdut mv/ Idut ma).

DUT Device                     100KHz                         300KHz                        1MHz

Wurth 10nH                     11.86                            11.83                          11.79
Pulse Electronics 10nH      10.34                            10.38                          10.37

With TH2830   

Wurth                              11.878
Pulse                               10.574

This is with our mentioned modified Standard SMD Fixture, swapped and use on both the IM3536 and TH2830.

Note how small the Pulse 10nH is compared to the Wurth 0603 10nH!!

Think your fixture is causing the issues you are experiencing. Just tried 8MHz and getting 11.83nH with Wurth, so the fixture is holding up well to 8MHz.

BTW Tonghui specs thier SMD fixture to 120MHz

Wouldn't rely too much for low Z work tho, and why we developed the Special Split Kelvin Fixture


Edit: Went back and reran with Wurth (Pulse is too small and difficult to handle) and added a couple images.

2nd Edit Removed Vdut and Idut readings, see post 137 below.

Best,                                           

[egonotto]

Got a Wurth RF 0603 Chip 10nH Inductor & Pulse Electronics with Standard SMD fixture and measured with IM3536 at 300KHz and 100KHz. Readings are in nH (Vdut mv/ Idut ma).

DUT Device                        100KHz                      300KHz                        1MHz

Wurth 10nH                     11.86 (20.04/102.2)     11.83 (22.95/101.7)      11.79 (41.35/98.22) 
Pulse Electronics 10nH      10.34 (28.05/101.2)     10.38                           10.37

With TH2830   

Wurth                              11.878 (5.22/33.16)
Pulse                               10.574 (7.97/33.07)

.....                               

Hello,

I have calculated the impedance. I must have made a mistake somewhere, because it doesn't add up.

Best regards
egonotto

[mawyatt]

....
From an HF point of view, this fixation is simply garbage.
And not only this fixation, others too.
They are all not "HF-tight" ....

Since these fixtures are somewhat direct clones of the Tonghui TH26008A which is rated to 120MHz and states, "Applicable to: SMD devices, high frequency small capacitors and small inductances." are you implying that their TH26008A is garbage??

https://www.tonghui.com.cn/en/product_144.html

BTW what is "HF-tight"?

Honestly think Tonghui knows what they are doing and we've certainly got excellent results after modifying the clones but only to 8MHz tho

Not too impressed (exposed blades in handle) with the Tonghui Kelvin Clips tho, but haven't seen any others that appear better.

Hope you get over your cold soon!!

Best

[mawyatt]

Got a Wurth RF 0603 Chip 10nH Inductor & Pulse Electronics with Standard SMD fixture and measured with IM3536 at 300KHz and 100KHz. Readings are in nH (Vdut mv/ Idut ma).

DUT Device                        100KHz                      300KHz                        1MHz

Wurth 10nH                     11.86 (20.04/102.2)     11.83 (22.95/101.7)      11.79 (41.35/98.22) 
Pulse Electronics 10nH      10.34 (28.05/101.2)     10.38                           10.37

With TH2830   

Wurth                              11.878 (5.22/33.16)
Pulse                               10.574 (7.97/33.07)

.....                               

Hello,

I have calculated the impedance. I must have made a mistake somewhere, because it doesn't add up.

Best regards
egonotto

You didn't make a mistake, we did and in hind sight we shouldn't have posted the Voltage and Current values

The impedance calculated from these values must include the LCR Open and Short Calibration coefficients to be correct for the DUT, these coefficients are not directly available.

We should have realized this before posted about Martin's readings 

Thanks for pointing this out, we'll remove those for clarity!!

Best,

[egonotto]

Hello,

thank you for your answer.

The voltage at the DUT should be easy to measure using the 4-wire method. The current is distorted by the shunt.

You should be able to estimate the impedance of the shunt from the values given. The phase shift must be taken into account.

I still have to think about whether it is possible to calculate the shunt from the given values, i.e. measured inductance, frequency, measured voltage and measured current. The ohmic resistance of the coil should be negligible in this case.

Best regards
egonotto
PS: I don't think that's possible. If you look at the series connection of shunt and DUT, you have two unknowns, the shunt and the voltage that is applied to the series connection.

[mawyatt]

The problem with these type SMD fixtures (why we developed the Split-Kelvin type) is the 4 wire Kelvin connection is made at the entry to the plunger bottom, not at the Plunger tip which makes contact with the SMD DUT.

If you look at Martin images you can see where the Force and Sense (Hcur-Hpot and Lpot-Lcur) wire pairs enter the case top cover to the plunger, this is where contact is made, not at the DUT. The LCR meter will "attempt" to remove this unmeasured connection from the Force Sense connection to the DUT by means of the Open and Short Calibration, however this is far from ideal.

The method we developed attempts to remove this uncertainty by keeping the Force and Sense wire separate until they are connected by the SMD DUT end terminals, thus removing most of the uncertainly in measurement. This is a little tricky to get all 4 connections to the SMD DUT, and why we created the rough contact surface texture, but with a little patience results are quite repeatable down to milliohm Z levels!! Of course this assumes that you are using an LCR meter capable of resolving this low a level!!

Anyway, thanks again for pointing out our improper numeric posting

Best,

[Martin72]

Since these fixtures are somewhat direct clones of the Tonghui TH26008A which is rated to 120MHz and states, "Applicable to: SMD devices, high frequency small capacitors and small inductances." are you implying that their TH26008A is garbage??

https://www.tonghui.com.cn/en/product_144.html

BTW what is "HF-tight"?

Honestly think Tonghui knows what they are doing and we've certainly got excellent results after modifying the clones but only to 8MHz tho

Not too impressed (exposed blades in handle) with the Tonghui Kelvin Clips tho, but haven't seen any others that appear better.

Hope you get over your cold soon!!

Best

Thankyou, I hope tomorow it´s going better...
I didn't say that Tonghui's fixtures are garbage, but from an HF point of view everything is still far too "open".
I could imagine at least "sealing" the inside of the housing with HF sheet metal, perhaps even constructing a cover that can be placed on the fixture during the test.
But these are just thoughts for now.

[mawyatt]

Thankyou, I hope tomorow it´s going better...
I didn't say that Tonghui's fixtures are garbage, but from an HF point of view everything is still far too "open".
I could imagine at least "sealing" the inside of the housing with HF sheet metal, perhaps even constructing a cover that can be placed on the fixture during the test.
But these are just thoughts for now.

The case and cover are sort of an RF enclosure. However, the cover must be well constructed, tight fit, and connected to the case for good RF shielding. Also applying to the BNC shell to case for a good RF connection. If you study the design the only RF exposure is with the two plunger tips, which has worked well with our experience.

You want the enclosure interior to provide space for the wire from the BNCs to the plungers, and also enough distance to each wire to have a small effective shunt parasitic capacitance. This becomes important for the movable plunger side as the wires move with the plunger the parasitic capacitance changes. If the space were smaller with wires closers to case/cover the parasitic capacitance would be higher and the capacitance change with plunger movement larger.

Note the need for a field shield between the H and L sides of the enclosure for High Z and/or Higher frequency work, you don't want the H and L wires "seeing" each other electrically. A simple "L" shaped aluminum bracket shield bolted to the case would work, or even a Cu or FR4 rectangular section soldered to the BNC ground washer assuming the washer is well connected to the case.

Suspect when you get the OEM Tonghui version at work, you'll see all the finer details they use. The clones obviously forgo many of these details to save cost, but one can recover the benefits with some additional tinkering and work

Best, 

[mawyatt]

@ Martin,

Get some good pictures of the SH26008A (Tonghui TH26008A) if you can. As most quality OEMs they pay attention to detail, altho one could question the Guard Terminal on the REAR of the TH2830 LCR Meter

Best,

[Martin72]

Shouldn't VAC be around 0.2 mV or IAC around 100 mA?

Thanks again for this post.
I simply asked ET how the VAC/IAC values come about and as always they responded promptly (which is really great).
They sent me a picture of a handwritten sketch, really cute.
But then you can also understand it, because in my calculations I neglected the output resistance (100 ohms).
So far so good, so understandable - but....
The engineer who drew this has made a small calculation error.
The nano value is one decimal place too low.
This has hardly any effect on the current display, but the voltage is displayed too high by a factor of 10.
So if this is the same engineer who is responsible for the software, then the same error will be found in the program.
I have therefore sent ET a picture with notes on the slip of paper.
Let's see what comes back.
Martin

[mawyatt]

100 ohm source impedance is too high to evaluate low Z components like 10nH at LCR Meter frequencies. Look at the ratio between the DUT Z and the source impedance!!

This is why the TH2830 has a lower 3 ohm range (30 ohm source) and the IM3536 is lower with a 0.1 ohm range (10 ohm source)!!

Best,

[Martin72]

Hi Mike,

This is not a problem, the output resistance can be selected for the ET35 (as well as for the small ET44/45).

This is why the TH2830 has a 33X lower 3 ohm range

As with the ET35, you can choose between 30 and 100 ohms for the TH2830/ST2830.

[mawyatt]

The TH2830 has a lower 3 ohm range with 30 Ohm source impedance (IM3536 is 0.1 ohm range with 10 ohm source impedance), what is the ET35 lower range?

Always use the lower range and source impedance except for higher Z components. The DUT current is mostly dictated by the source impedance for low Z DUT, so 30 ohms would give you ~3.3X the DUT current and thus ~3.3X the DUT voltage over the 100 ohm source, which makes things a little easier to resolve.

Of course this assumes the ET35 can deliver this level of current stability and accurately at frequency.

Best,

[mawyatt]

Shouldn't VAC be around 0.2 mV or IAC around 100 mA?

Thanks again for this post.
I simply asked ET how the VAC/IAC values come about and as always they responded promptly (which is really great).
They sent me a picture of a handwritten sketch, really cute.
But then you can also understand it, because in my calculations I neglected the output resistance (100 ohms).
So far so good, so understandable - but....
The engineer who drew this has made a small calculation error.
The nano value is one decimal place too low.
This has hardly any effect on the current display, but the voltage is displayed too high by a factor of 10.
So if this is the same engineer who is responsible for the software, then the same error will be found in the program.
I have therefore sent ET a picture with notes on the slip of paper.
Let's see what comes back.
Martin

Also ask, why would the ET engineer even suggest a 100 ohm source for the low Z 10nH inductor, when apparently 30 ohms is available

Best,

[Martin72]

He did not suggest anything.
He only calculated what I had asked.

[Martin72]

Hi,

Today I took the fixture and the choke/capacitor to work.
Afterwards I also retested at home with 100khz on the ET35.
I don't think you should pay too much attention to this Im/Vm thing.
While the values for the ST2830 are still correct for the choke in terms of voltage/current, they are higher for the ET35 and both are wrong for the capacitor.
I found an explanation in the ST2830 manual.
There it is expressly stated that these Im/Vm values are for information only, i.e. no accuracy is given, and they can deviate anyway if corrections (open/short/load) are used.
Presumably this is why they can also be switched off on the ST2830, so that no confusion arises.
For me this is the end of the matter, it would have been nice to find such an explanation in the ET35 manual too.
But now I know, and now I'm almost sorry that I was bitching at the guys from ET before I read it in the ST2830 manual.