ET432 LCR Meter - Wild Results - Range Function?

[killingtime]

Hello,

I was attempting to take some transformer winding resistance readings with an East Tester ET432 LCR meter this evening, and was getting some very odd results.

This meter has a DCR function meaning no AC is used. This is necessary when measuring the winding resistance of an inductor as the reactance alters the measurement.

There's no series or parallel equivalent circuit in this mode, and no frequency setting. The only parameter you can alter is the RANGE, which cycles through 100Ω, 1KΩ, 10KΩ, 100KΩ and AUTO. AUTO normally gets it right, but on some occasions it gets it wrong and you have to manually pick a setting.

So far so good. It can't be that different to a multi-meter right? A DMM has different resistance ranges. You pick the one that gives the most precise reading. 3.001 KΩ and not 0.003 MΩ. Should be easy.

Cycling through the range options gives wildly different results.

For a 10KΩ 5% resistor in DCR mode after performing an open & short user calibration (2 wire):

100Ω - 921.6 KΩ
1KΩ  - 98.92 KΩ
10KΩ - 9.953 KΩ
100KΩ - -9.259 MΩ (minus!)
AUTO - 9.953 KΩ

I'd expect some loss of precision as you move around the ranges, but the above are at least an order of magnitude out. There's no warning it's out by such a margin either. You just have to know.

To confuse matters a bit more, the manual lists a bunch of 'ranges' for the DCR mode that are not available on the unit. Extract from the manual attached. Manual lists 0.4Ω, 40Ω, 400Ω, 4KΩ, 40KΩ, 400KΩ, 4MΩ, 10MΩ and 20MΩ.

So either:

1) The ranges in the manual never made it into the production model or,
2) They've just copied stuff from another manual (Keysight) or,
2) I've misunderstood how ranging works on an LCR meter.

Meter has the same issues on L or C ranges as well. Exact same options for RANGE. Changing measurement speed to slow doesn't improve matters either. Using 4 wire kelvin probes has little effect.

Do any of the Keysight_U1731C / U1732C / U1733C work the same way, or have the same issues? I wouldn't expect the values to be this far out. If anyone has a U17xxx Keysight LCR I'd be interested to know what you see when manually ranging.

Thanks.

[csuhi17]

If I understand the description correctly.
there is a physical range which is 100, 1k, 20k, 100k. It has to do with the resistors in it.

And there is what you shared from the description, the value that can be read from the display. for example, if you start adjusting a 0-220 ohm pot, you will see a change in resolution at 40 ohms. From 39.999 ohms to 40.00 ohms.
The accuracy associated with these has been described.

[killingtime]

>>The accuracy associated with these has been described.

It has been described in the manual, but it's not meeting those accuracy specs at all, apart from on one range, and you have to guess the range correctly, which makes the whole exercise a bit pointless. The whole point of a measuring device is to take the guess work out of it.

To be clear, a 10K resistor displays as 921.6K on the 100 Ohm range, and 98.92K on the 1K range, which isn't 0.x% + x digits is it? That's out by a factor of 9116% on the 100 Ohm range. That and the fact there isn't a 100 Ohm range in the manual, only 40 Ohms and 400 Ohms...

I wonder if the guy that wrote the manual actually had this unit in his hands to test with, or perhaps he just borrowed something else.

Edit:
Manual is here in case anyone else has this meter.

https://www.easttester-cn.com/uploads/ET430-HANDHELD-NEW-COMPACT-LCR-METER-USER-MANUAL.pdf

[Tycho_Brahe]

I have a slightly different model (ET431).   So, I used a 10K carbon film resistor 5%.   This thing has kelvin clips, so I just used those right off the bat.   I 'think' I got results similar to you.   Used DCR mode and got:

Range	Ohms Measured
100	908 K
1 K	98 K
10 K	9.848 K
100 K	negative 2.1 M
Auto	9.849 K

I guess I never noticed this before because I just rely on Auto.  Thank goodness Auto usually picks the magic range.   But what the heck?  I'm also wondering if I'm just using the thing incorrectly (I wouldn't be surprised as I really don't know my around this device yet). 

So you say it also happens when measuring capacitors and inductors (I didn't have time to test that tonight)?    Hey, so when you said on rare occasions the Auto setting gets it wrong and you have to resort to a manual range, how wrong was Auto?  Was it as crazy wrong as some of our numbers (like orders of magnitude and sometimes even crazy negative values)?

[killingtime]

... how wrong was Auto?

From the testing I've done, it depends on where the value you're measuring is compared to the nearest range. After the 10K I tried a 5.6K as it's half way between the 1K and 10K ranges. I got 5.6K and 7K on different ranges. The other values were out by a factor of at least 10.

If your DUT is near a range, one range reading will be accurate, and all the others will be out by a factor of 10 at least.... which is fine if you already know the value of your DUT, but if you don't then it's all guesswork. I'm thinking of smd capacitors and toroidal core inductors, both of which are usually unmarked.

In my case I was measuring the primary inductance of a 3KVA line transformer with the output unloaded (so just a large inductor). AUTO was giving me 0.7H which I thought was low, as transformers like this normally measure above 8 Henries (depends on the core type, number of windings and other things). The next range gave me the correct reading (9.9H).

If you're getting similar readings to me, then I didn't get the 'Friday afternoon model' and this is either just the way this meter has been designed, or it's how all LCR meters work.

If anyone has a different brand LCR hand held (Keysight or BK Precision) I'd be grateful if you would stick a 10K metal film on it and play with the range setting. That will tell us if it's worth spending about 3x for a different design.

Many thanks.

[Veteran68]

I'll try to remember to check with my Hantek 1833C tonight.

[csuhi17]

ET432 blue
I had time to test a few resistors.

In normal mode, when the frequency is not 0Hz.
Range: Auto <- All OK, writes the real value.
Range: 100;1k;10k;100k <- The value read rarely matches the value read in AUTO mode.
For example: 10k resistor tested at 1kHz frequency.
Auto: 9,974k
100: ~10,100k The value varies a bit
1k: 9.974k The value is stable
10k: 9.976k is also stable
100k: 18,641k Also stable
In DRC mode:
Auto: 9,979k
100: 9.520M changes by 0.4M Ohm
1k: 987.6k changes with 0.2k ohms
10k: 365.12k changes
100k: -60.00M changes

What I can't process:
4.99k resistance in DRC auto mode 4.980k
However
100: 4.760M
1k: 492.8k
10k: 1.4530M
100k: -71.80M
Range: auto mode measures well again.

Use only AUTO mode for DRC measurement!
For non-DRC measurements, use only in AUTO mode!

I don't know what causes it, but I don't think this phenomenon occurs with other brands of devices, at least I haven't read about it.

This meter has many bugs that the manufacturer has not fixed and probably will not.

[killingtime]

This meter has many bugs ....

The ones I have found so far;

1) In slow mode the meter can lock up. Keypad unresponsive. Power button unresponsive. You have to remove the bumper, unscrew the battery hatch and disconnect the battery to reboot. If you can wait 10 minutes after pressing the 'speed' button (seems to have a keypad buffer) it eventually jumps into 'medium' speed mode and things speed up. Press the speed button again to get into fast mode and you're back in business. Got this to repeat by just pressing buttons too quickly. Not had it lock up in fast mode. Thought this might be a dry solder joint initially but I can't find any on the PCB. I think it's a software issue.

2) The user calibration is either open or short, not both. Doesn't appear to be any way to do open and short at the same time. It only seems to remember the last calibration you did from my tests. If that was open, then short readings are out by quite a bit (they come back negative). If you cal on short then shorts read 0.000 but high values are then out. So it looks like you have to know which range the meter is going to use and calibrate accordingly. Doesn't help that in auto mode it doesn't tell you which range it selected. Guessing 100Ω and 1K use short cal, and 10K, 100k use open. Nothing in the manual about this.

3) After a user cal the meter sometimes jumps back into slow mode when moving between measurement types; LCR, DCR etc.

[csuhi17]

1) So far, it has not produced such an error for me.

2,) I didn't notice, it's true that you have to calibrate the open and closed states separately, but for me, if I do it in the order 1. closed 2. open, the resistance of the measuring cable remains 0 in DRC mode.
If I measure with an alternator, i.e. in non-DRC mode, 100-200mOhm remains, I think it might be a property of the measuring cable...
The factory Kelvin clip is not the best.

3) I don't have this error either.
In the case of calibration, if I calibrate in non-DRC mode, it lasts for 30 seconds and resets the frequency to 100Hz, all the others remain at the settings I left.
If I calibrate in DRC mode, it is done in about 4 seconds.
if I change the Range to another value, e.g.: 100k and calibrate one, it resets, i.e. it shows 0 Ohms, but when I change the Range to 10k and then back to 100k, it shows the wrong value again.

What I wrote as an error is the incorrect measurement of high-value resistors in non-DRC mode.
At 10kHz and 100kHz, it incorrectly measures high-value resistance.

There was another one that I thought was a bug, but it wasn't.
The RECord function does not work if the measured value has a negative sign.
I don't even understand :Đ

However, looking at the pictures, your version and mine are different.
I have the Blue, you have the yellow.
Things on the display look a little different to me.
It looks like this

https://www.aliexpress.com/item/1005004656845323.html

My Version : V1.00.2048.018

[killingtime]

My Version : V1.00.2053.019

[Tycho_Brahe]

I see you guys are posting your version numbers.  Does this mean we can update the firmware or we're just at the mercy of whenever we purchased the product?    I'm not finding any update files here:  https://www.easttester-cn.com/et430-handheld-new-compact-lcr-meter-product/

[Veteran68]

If anyone has a different brand LCR hand held (Keysight or BK Precision) I'd be grateful if you would stick a 10K metal film on it and play with the range setting. That will tell us if it's worth spending about 3x for a different design.

I'll try to remember to check with my Hantek 1833C tonight.

Results below in both 2W and 4W modes using the Hantek-supplied shorty croc leads and Kelvin clips with the slot connector. I recalibrated (open and short) between each probe type. Regardless of the manual range selected, the Hantek always displayed correctly as K ohms. The test frequency was at left 100Hz for these results, but I did try all the supported frequencies to see if it made a difference -- the reading didn't change between 100Hz and 10KHz, then creeped up ~10mOhm 10 Ohm between 40Khz and 100KHz. Overall, the 1833C tested as expected, with only the 100K manual range being significantly off. And nowhere near the cost of a Keysight or BK.

I also checked against my Agilent 34401A and Siglent SDM3065X 6.5 digit meters using the same ProbeMaster 8000 probes.

10K 1% Metal Film Resistor

Hantek 1833C - Croc leads 2W
AUTO 10.052K (stable)
10R 10.154K (moves)
100R 10.141K (moves)
1K 10.055K (stable)
10K 10.052K (stable)
100K 11.257K (stable)

Hantek 1833C - Kelvin leads 4W
AUTO 10.053K (stable)
10R 10.153K (moves)
100R 10.120K (moves)
1K 10.053K (stable)
10K 10.053K (stable)
100K 11.250K (stable)

Agilent 34401A - Probemaster 8000 2W
AUTO 010.0524K

Siglent SDM3065X - Probemaster 8000 2W
AUTO 10.05281K

EDIT: Corrected resistance creep value at high test frequencies

[killingtime]

Thanks for the test results Veteran68.

So it looks like it's just an EastTester thing then.

If you look at the official factory page for this model below, they seem quite happy to showcase the meter displaying negative capacitances. That should have been a warning that there's still some work to do on the firmware. Sigh.

https://www.aliexpress.com/item/1005004656845323.html

[Veteran68]

I'm generally not a fan of Hantek gear, but I have to say this 1833C LCR meter has been a pretty good one for my needs so far. I've been saying if I were buying another handheld I'd consider the Uni-T 622E or the Sanwa LCR700, but not really sure I'd need to.

I've been considering a step up to a bench LCR meter, but not one of the super expensive pro level ones. So that means the likes of the EastTester 44xx series and the RuoShui/Victor 40xx series in the $300-400 range. Based on what I've been hearing from EastTester reports lately, I'll probably steer clear of them. The RuoShui/Victor line looks like much the same hardware but seems to get better reviews, probably due to better QC and/or firmware.

Would love to have a Tonghui/Sourcetronic but can't justify spending that money for my hobby. It's not like a DMM that I'd use most every day.

[killingtime]

Pulled out another line transformer to see if it's just inductors that the meter doesn't like.

Hooked the Xfmr up to mains and measured the primary current with the secondary unloaded. 0.4A rms, which yields 1.95H using XL=2(pi)FL @245V AC rms 50Hz. Measured using a Watt meter and my DDM on the 10A range. Both agreed.

ET432 readings below. AUTO gave the wrong reading at 0.528H. That's out by a factor of 3 at least The ESR looks way out as well. I did a DCR measurement and got 2.3 Ohms which agrees with my Fluke DMM. The 10K range is closest at 1.6H, but that's still a big error.

Hmm.

[csuhi17]

Thanks, Then according to your measurements, this is one of my worst devices. I'll keep that in mind...
That way I won't fall in love

[Tycho_Brahe]

This is depressing.   I thought I had a decent brand/model picked out.   Oh well.   I think I still have until Feb 1 at the latest to start a return back to Amazon.   Wife bought me one for Christmas.

I emailed East Tester last night for their opinion, but haven't heard back (and may never, who knows).    I guess a Hantek could be in my future.   

[Randy222]

Hello,

I was attempting to take some transformer winding resistance readings with an East Tester ET432 LCR meter this evening, and was getting some very odd results.


Cycling through the range options gives wildly different results.

For a 10KΩ 5% resistor in DCR mode after performing an open & short user calibration (2 wire):

100Ω - 921.6 KΩ
1KΩ  - 98.92 KΩ
10KΩ - 9.953 KΩ
100KΩ - -9.259 MΩ (minus!)
AUTO - 9.953 KΩ

I'd expect some loss of precision as you move around the ranges, but the above are at least an order of magnitude out. There's no warning it's out by such a margin either. You just have to know.

To confuse matters a bit more, the manual lists a bunch of 'ranges' for the DCR mode that are not available on the unit. Extract from the manual attached. Manual lists 0.4Ω, 40Ω, 400Ω, 4KΩ, 40KΩ, 400KΩ, 4MΩ, 10MΩ and 20MΩ.

So either:

1) The ranges in the manual never made it into the production model or,
2) They've just copied stuff from another manual (Keysight) or,
2) I've misunderstood how ranging works on an LCR meter.

Meter has the same issues on L or C ranges as well. Exact same options for RANGE. Changing measurement speed to slow doesn't improve matters either. Using 4 wire kelvin probes has little effect.

Do any of the Keysight_U1731C / U1732C / U1733C work the same way, or have the same issues? I wouldn't expect the values to be this far out. If anyone has a U17xxx Keysight LCR I'd be interested to know what you see when manually ranging.

Thanks.

Did you scope it in DCR mode just to verify it's really DC ?

One of my old LCR meters says on the front "L C R", but R is really a Z mode, runs a fixed 1kHz. So that meter is 100% useless if I just want real component value.

I'm generally not a fan of Hantek gear, but I have to say this 1833C LCR meter has been a pretty good one for my needs so far. I've been saying if I were buying another handheld I'd consider the Uni-T 622E or the Sanwa LCR700, but not really sure I'd need to.

I've been considering a step up to a bench LCR meter, but not one of the super expensive pro level ones. So that means the likes of the EastTester 44xx series and the RuoShui/Victor 40xx series in the $300-400 range. Based on what I've been hearing from EastTester reports lately, I'll probably steer clear of them. The RuoShui/Victor line looks like much the same hardware but seems to get better reviews, probably due to better QC and/or firmware.

Would love to have a Tonghui/Sourcetronic but can't justify spending that money for my hobby. It's not like a DMM that I'd use most every day.

What FW version is your 1833C?

[Veteran68]

What FW version is your 1833C?

20210615AM
HW V1

[Randy222]

What FW version is your 1833C?

20210615AM
HW V1

I won't hijack this ET thread, but interesting, some are reporting a FW of "2023xxxxxxAM".
Thanks for noting it.

[Veteran68]

What FW version is your 1833C?

20210615AM
HW V1

I won't hijack this ET thread, but interesting, some are reporting a FW of "2023xxxxxxAM".
Thanks for noting it.

We can take the discussion over to the other thread, but I just looked and 20210615AM is still the latest version Hantek offers on their website. New units may be shipping with something newer -- isn't yours a brand new unit? What version is it?

[Randy222]

We can take the discussion over to the other thread, but I just looked and 20210615AM is still the latest version Hantek offers on their website. New units may be shipping with something newer -- isn't yours a brand new unit? What version is it?

I took this to the Hantek unlock thread.

[killingtime]

Did you scope it in DCR mode just to verify it's really DC ?

Good question. Hooked it up to a scope and switched it into DCR mode.

The unit outputs a square wave. Period 2 seconds, 1.16 pk-pk voltage. Minimum voltage is 234mV. I don't think this would cause an issue though.

In LCR mode it outputs a sine wave. The default 600mV reading on the ET432 display is RMS according to my scope measurements. See scope cursors.

Attached.

[Randy222]

0.5Hz 50% duty square, fo a Rdc measurement?

Just makes no sense to me. If the DUT is a large inductor then the hard rise and fall of sqaure wave "is" an issue.

Would really need to know how the meter runs the math to obtain Rdc from 0.5Hz 50%

[mawyatt]

This is often done to reduce thermal EMF effects.

Math is simple, Thermal EMF effects don't change with test voltage polarity reversal. So subtracting the negative test voltage from the positive test voltage yields twice the peak voltage value while eliminating the thermal EMF effects as they subtract out 

Best,