New LCR tweezer, Fnirsi LCR-ST2 (November 2025)

[indman]

Normal for Fnirsi 

[mawyatt]

When comparing a capacitor measurement with a DMM and LCR meter, the techniques utilized are different, so one shouldn't expect the results to agree to a high degree. The better capacitors for comparisons between like instruments are Mica, Polystyrene, Polypropylene, and NPO ceramic types.

All capacitors capacitance varies with many parameters, including frequency (temp, voltage, history, age and so on), and electrolytics are high on the list. Most well designed circuits don't utilize electrolytics or other non-precision types for precision stable use.

Generally as stated, it's best to measure at the frequency of intended use, and why 100/120Hz for AC Mains filtering, and 10/100kHz for SMPS filtering are good choices for capacitors intended for this type of use (reflected in the data sheet of quality sources).

A proper LCR meter has an advantage as the frequency can be selected whereas the DMM doesn't have this capability, so preference should lean towards the LCR meter for accurate/repeatable/comparison measurements. Just like resistors (and inductors) different capacitor types are available, intended for different applications and one size does not fit all!!

Best

 

[indman]

so preference should lean towards the LCR meter for accurate/repeatable/comparison measurements.

A small addition to the correct message is that preference should be given to an LCR meter that you can trust.
FNIRSI also positions its tweezers as an LCR meter.

[dobsonr741]

Is it able to measure below 100nH and 10pF with at least 10% accuracy?

[KungFuJosh]

Is it able to measure below 100nH and 10pF with at least 10% accuracy?

Doubtful, and yes. I can check later or tomorrow and report back to confirm nH.

With the kelvin clips, 1pF came back as 3pF, but 10pF was ~12pF, so the latter is within your margin- though it's worth noting that both were 2pF too high, for whatever that's worth. With the tweezer tips instead, I got 2.2uF and 10.4uF, so the latter there is much more reasonable. The kelvin clips appear to be a poorly executed gimmick...and yes, they were calibrated (in case somebody doesn't read above where I already stated that 😉). 100uF was 5% low,  and 1mF @100Hz was exactly 1.0, with no additional digits.

100nH and lower is a difficult task for LCR meters. Only my IM3570 has accurately measured at that level of all the meters I've tested, including other not cheap options.

[indman]

Is it able to measure below 100nH and 10pF with at least 10% accuracy?

For such measurements it is not necessary to buy tweezers. A regular tester transistors with firmware 1.13k can do a good job with this task.
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4596946/#msg4596946
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg3732526/#msg3732526

[KungFuJosh]

Is it able to measure below 100nH and 10pF with at least 10% accuracy?

For such measurements it is not necessary to buy tweezers. A regular tester transistors with firmware 1.13k can do a good job with this task.

Whether that's reliable or not, it's gonna be better than the Fnirsi.

[indman]

Whether that's reliable or not, it's gonna be better than the Fnirsi.

It is undoubtedly better and more reliable, because the software authors have spent a lot of time (years) debugging the code.

[KungFuJosh]

Whether that's reliable or not, it's gonna be better than the Fnirsi.

It is undoubtedly better and more reliable, because the software authors have spent a lot of time (years) debugging the code.

Conversely, it's undoubtedly better because Fnirsi hasn't.

[Paul T]

Received my unit today and put it through some paces.  You can see my results graphed below. All the accessories including the kelvin leads fit inside the slim case.

I felt the USB switchover to Kelvin leads was very sturdy and so was the tweezer alignment when squeezed.  However, the tips no longer have the internal serrations that I liked so much for axial lead components with the ST1.  When using the kelvin leads, readings tended to be a bit more noisy.  Therefore I concentrated on using the tweezers for the shown measurements.

Interestingly the scanned values differed regularly (but not severely) from the main display.  100 KHz frequency did not help with pF precision or accuracy as well as I had hoped.  It helped resolve nH inductance values better.  However, my ET431 in 10 KHz mode was way more accurate than the LCR-ST2 in 100 KHz.  Changing voltage level to 1.0V instead of 0.6V helped with milliohm range readings, but they were still not good.

Voltage has +0.01 offset which shows up at zero and also measuring a calibrated source (10.00 vs 9.99 the other way).  The voltage indicator is not bright enough or clear enough.  They show a rectangular battery to indicate polarity.  I'd rather see large +/- markings.  Calibrating the voltage took multiple efforts.  The cal duration is not long enough for the amount of time the CPU needs to average the value.  I did not calibrate resistances, but tried SC and OPEN calibrations to see if that would cure the voltage offset.

Overall I'm not impressed because the new features did not materialize into extended accuracy.  I am more likely to grab the older ZT-MD1 or my full-sized LCR if I need the best result.  Hoping for firmware updates to address the shortfalls.

The diode mode is combined with continuity. There is no beep on diode. The mode is 0.6V only according to the screen, but will measure silicon diodes as 0.64V.

I also noticed 100KHz enforces fast mode even if the device is set for slow.

[habsrepariert]

I've tested it as well - not in regards of accuricy but in regards of practical use for repairs. In my opinion for this money it is a useful tool to find defective caps in circuits - and to judge if the capacity and ESR are ok or not. This you can't find with a DMM. To test in a component tester you have to solder the part out.
My experience is here:
https://youtu.be/ox7PW-9L6dg

[habsrepariert]

I've tested it as well - not in regards of accuricy but in regards of practical use for repairs. In my opinion for this money it is a useful tool to find defective caps on circuits - and to judge if the capacity and ESR are ok or not. This you can't find with a DMM. To test in a component tester you have to solder the part out.
My experience is here:
https://youtu.be/ox7PW-9L6dg

[Paul T]

I modified the resistor board they provided. Because the board itself has resistance, it made no sense to use R0 for calibrating a short. Instead mine now has a 100 milliohm total resistance where the R0 used to be. I also found some 0.05% resistors that are more accurate than the ones they supply. Still, I wish they had the slots open to the end of the PCB so that the kelvin clips can be used.

[BillyO]

Interesting.

I see the eevblog habit of comparing $40 thingy's to instruments costing 6 to 160 times as much is still alive and well.   

I just got one and it compares to other $40 things just fine.  It indeed has some additional features that put it ahead of other thingy's in this price category like .1V to 1V signal amplitude, up to 100kHz frequency, Kelvin clips, etc..

I did find it is a bit wonky in capacitance measurement and pretty sensitive to frequency.  As for L and R, it's pretty good.  Maybe the capacitance measurements can be improved at some point, but firmware updates are hit and miss with FNIRSI.

I found ESR measurements to be on par with others in this price category.

As for dispersion factor, hmmm, in 10 years of professional experience in electronics (earlier in my working life) and 50 years of hobbyist, f'n around, dispersion factor has never been a factor.   

I think it amounts to a reasonable choice for someone looking for something in this price range.  The only caveat I have, is that people will not bother to re-calibrate when moving from tweezer to Kelvin clips.  But heck, it's a $40 class device.  The error introduced by going back and forth is not likely to make any difference to the person looking for a $40 device.

It's fine for what it is.  It's not going to kill anyone, or ruin anyone's day by being horribly out to lunch.  It will measure a 1uF capacitor around 1uF, it will measure resistors, most capacitors and inductors well within 1%, but it will not compete with a Shannon ST42 or a Hioki IM3536, and no one should expect it to.

[KungFuJosh]

It's fine for what it is.  It's not going to kill anyone, or ruin anyone's day by being horribly out to lunch.  It will measure a 1uF capacitor around 1uF, it will measure resistors, most capacitors and inductors well within 1%, but it will not compete with a Shannon ST42 or a Hioki IM3536, and no one should expect it to.

I think with some firmware updates it could be fine for what it is. I compare it to higher quality devices to see if it's telling the truth at all. IMO, they should remove the features that lie if they can't manage them in a $40 product. Then it really would be fine. For nominal value checks, it's not bad.

And for the record, the IM3570 is 275 times more expensive.

When I got my ST42, it was less than 3 times more, and still is a great value at its higher price now.

[BillyO]

And for the record, the IM3570 is 275 times more expensive.

Oops, sorry.  As usual I was rummaging in the bargain bin. 

[KungFuJosh]

And for the record, the IM3570 is 275 times more expensive.

Oops, sorry.  As usual I was rummaging in the bargain bin. 

That's the only place to shop.

[BillyO]

That's the only place to shop.

It's my home!   

Irish.  'Nuff said.

[indman]

Interestingly the scanned values differed regularly (but not severely) from the main display.  100 KHz frequency did not help with pF precision or accuracy as well as I had hoped.  It helped resolve nH inductance values better.  However, my ET431 in 10 KHz mode was way more accurate than the LCR-ST2 in 100 KHz.  Changing voltage level to 1.0V instead of 0.6V helped with milliohm range readings, but they were still not good.

Here is a competent analysis of the shortcomings of the hardware of these cheap tweezers from a person who well understands the circuitry of such devices:
"1. There is no quartz on the board. Internal stability 0.5%.
2. The frequency at 100 kHz dances. It is very difficult to reliably record this frequency on the oscilloscope screen. Only with the STOP button. From this, a reliable lower limit of measurement is 1 μH, and not nH.
3. There is absolutely no calibration for the 2nd parameter!! D and Q shows nonsense at frequencies from 1 kHz. In scanning mode it is clearly visible. It looks like only the active resistance is calibrated, but not the reactive resistance!! That is, there is no LOAD calibration.
4. PGA112 is installed on the board. Physical ranges are switched with a factor of 2, and calibration resistors have a step of 10."

Of course, from a device costing $30-40 it would be stupid to demand the hardware of devices costing an order of magnitude higher, but this once again confirms that you are getting a device only for a quick check of a part, nothing more.

[BillyO]

Interestingly the scanned values differed regularly (but not severely) from the main display.  100 KHz frequency did not help with pF precision or accuracy as well as I had hoped.  It helped resolve nH inductance values better.  However, my ET431 in 10 KHz mode was way more accurate than the LCR-ST2 in 100 KHz.  Changing voltage level to 1.0V instead of 0.6V helped with milliohm range readings, but they were still not good.

Here is a competent analysis of the shortcomings of the hardware of these cheap tweezers from a person who well understands the circuitry of such devices:
"1. There is no quartz on the board. Internal stability 0.5%.
2. The frequency at 100 kHz dances. It is very difficult to reliably record this frequency on the oscilloscope screen. Only with the STOP button. From this, a reliable lower limit of measurement is 1 μH, and not nH.
3. There is absolutely no calibration for the 2nd parameter!! D and Q shows nonsense at frequencies from 1 kHz. In scanning mode it is clearly visible. It looks like only the active resistance is calibrated, but not the reactive resistance!! That is, there is no LOAD calibration.
4. PGA112 is installed on the board. Physical ranges are switched with a factor of 2, and calibration resistors have a step of 10."

Of course, from a device costing $30-40 it would be stupid to demand the hardware of devices costing an order of magnitude higher, but this once again confirms that you are getting a device only for a quick check of a part, nothing more.

There must be something wrong with yours.  On mine, the100kHz has a standard deviation of less than 63Hz.

Even still, the measurement is not hugely dependent on frequency.  Measuring a capacitor at 1000Hz or 1005Hz should return essentially the same value.  At least within the basic accuracy of these devices.

I might bet you a year's wine supply that the secondary parameters, and possibly even some of the the 'primary' parameters, on these cheap meters are calculated  and not measured.

I agree 100% with your last statement.

[KungFuJosh]

a device only for a quick check of a part, nothing more.

I think this is a best-case scenario for anything Fnirsi sells. Worst case, would be dangerous items like their garbage stud finder that detected electrical wires as a stud. 

[BillyO]

Worst case, would be dangerous items like their garbage stud finder that detected electrical wires as a stud. 

Are you serious?

Holy F!

That needs to be withdrawn from the market.

[KungFuJosh]

Worst case, would be dangerous items like their garbage stud finder that detected electrical wires as a stud. 

Are you serious?

Holy F!

That needs to be withdrawn from the market.

Agreed! It probably won't be though. I shared a review of it on Amazon, and I showed a photo of the wall before the drywall was installed, and a video where it detects the wires as a stud. You can clearly see where the stud and wires are. The piece of shit went right over the stud and then said it was a stud center on the electrical wires. Fnirsi tried to make some BS excuses. They're a trash company. 

[BillyO]

Worst case, would be dangerous items like their garbage stud finder that detected electrical wires as a stud. 

Are you serious?

Holy F!

That needs to be withdrawn from the market.

Agreed! It probably won't be though. I shared a review of it on Amazon, and I showed a photo of the wall before the drywall was installed, and a video where it detects the wires as a stud. You can clearly see where the stud and wires are. The piece of shit went right over the stud and then said it was a stud center on the electrical wires. Fnirsi tried to make some BS excuses. They're a trash company. 

They've had hits an misses along the way, but this could kill someone.  I know I sometimes come across as a supporter of some of their stuff .. but not when lives are in the balance.  I mean, it's one thing to get some cheap thing that just isn't very good, it's another thing entirely to get something that suggests you to drive a nail into Romex!!!!

[KungFuJosh]

Exactly. But they don't care. They made excuses and the product still lives.