Design a New Precision LCR Tweezers with Much Lower Cost 【Shannon Tweezers ST42】

[SilverSolder]

I guess, you've seen that the pretty model ST5S/LCR Pro1 is 2 or 3 times expensive. 

Well, I don't want to talk much about those product's shortcomings, they are more like DMM, here is a test and analysis post, you can get more details.

Have you ever seen an LCR Tweezers that could achieve 0.01% noise performance at that speed?
And look at the test image, we used 0.01% 2ppm precision resistors for the DUT.

We will post more details about the performance and structure design, it used as real precision tweezers, very easy to hold 0402 SMD components.
It is much more convenient to use than a stationary LCR meter, it could support SMD(independent or soldered on PCB) and through-hole components at the same time.
We have great confidence in it, you'll want to have a try

The DT71 looks much more elegant and comfortable, the ST5S is unsightly. To use these tools as high-precision seems to me unwise. You can easily calibrate a demo sample and lose accuracy in mass production. If you ensure accuracy in mass production, it will be good. But we don't know that yet. 

I didn't buy a DT71 for $50 and I won't buy yours for $100, it's expensive and more convenient for me to use a stationary device.

It is often very inconvenient to look at the tweezer readings when checking on the board due to the forced position of the tweezers. However, I suggest that you, as a friend function that will cure this deficiency: automatic locking readings. It works like this: if the measured value is stored for more than, for example, 2 seconds, issue a beep and record the readings so that you can turn the tweezers and look at the screen. Of course, the time and on/off of this function must be added to the settings. 

What stationary device do you use?

I like the idea of the reading becoming "sticky" on the display, so you can review it.  Why not keep the previous reading until the user takes a new measurement...

[Shannon]

I guess, you've seen that the pretty model ST5S/LCR Pro1 is 2 or 3 times expensive. 

Well, I don't want to talk much about those product's shortcomings, they are more like DMM, here is a test and analysis post, you can get more details.

Have you ever seen an LCR Tweezers that could achieve 0.01% noise performance at that speed?
And look at the test image, we used 0.01% 2ppm precision resistors for the DUT.

We will post more details about the performance and structure design, it used as real precision tweezers, very easy to hold 0402 SMD components.
It is much more convenient to use than a stationary LCR meter, it could support SMD(independent or soldered on PCB) and through-hole components at the same time.
We have great confidence in it, you'll want to have a try

The DT71 looks much more elegant and comfortable, the ST5S is unsightly. To use these tools as high-precision seems to me unwise. You can easily calibrate a demo sample and lose accuracy in mass production. If you ensure accuracy in mass production, it will be good. But we don't know that yet. 

I didn't buy a DT71 for $50 and I won't buy yours for $100, it's expensive and more convenient for me to use a stationary device.

It is often very inconvenient to look at the tweezer readings when checking on the board due to the forced position of the tweezers. However, I suggest that you, as a friend function that will cure this deficiency: automatic locking readings. It works like this: if the measured value is stored for more than, for example, 2 seconds, issue a beep and record the readings so that you can turn the tweezers and look at the screen. Of course, the time and on/off of this function must be added to the settings. 

What stationary device do you use?

I like the idea of the reading becoming "sticky" on the display, so you can review it.  Why not keep the previous reading until the user takes a new measurement...

Hi SilverSolder,

we have implemented hold function.
if hold mode was enabled, when the tips of tweezers leave the DUT, the test result would keep display.   

Why not keep the previous reading until the user takes a new measurement
The reason is that it's not easy to decide when the tips are leaving the DUT
For example, assume the DUT is a precision resistor, the test value is stable before we try to release the tips, and then we loosen the arms of tweezers, contact states are changing, milliohm contact resistance is introduced, but the test is still running, and a new test result is displayed on the screen which has some uncertainty.
So hold function is better to be an optional mode when you can accept some uncertain, even though these uncertainties are usually small

There is a way to hold the precision result is we ask this tweezers "HOLD" the value during stable measurement, and tweezers stop the measurement and keep the stable value until next measurement.

[Hydron]

If the auto-hold function looks for a long enough period of stable readings before updating the measured value then it should ignore a few milliseconds of instability as the pressure is released.
Would likely be a bit of a compromise between speed and stability though, and there might need to be a dead-band around open-circuit that would mean very small capacitors need to be measured without auto-hold.

[SilverSolder]

[...]
Why not keep the previous reading until the user takes a new measurement
The reason is that it's not easy to decide when the tips are leaving the DUT
For example, assume the DUT is a precision resistor, the test value is stable before we try to release the tips, and then we loosen the arms of tweezers, contact states are changing, milliohm contact resistance is introduced, but the test is still running, and a new test result is displayed on the screen which has some uncertainty.
So hold function is better to be an optional mode when you can accept some uncertain, even though these uncertainties are usually small

There is a way to hold the precision result is we ask this tweezers "HOLD" the value during stable measurement, and tweezers stop the measurement and keep the stable value until next measurement.

The algorithm could be:   When reliably detecting the tweezers have "let go" of the DUT,  show and hold the statistical MODE (the most frequently occurring value) of the previous 300ms of readings before the tweezers were "let go". 

That way, you would always "hold" a stable and good reading, completely filtering out any glitches during the process of "letting go" of the DUT.

[Shannon]

If the auto-hold function looks for a long enough period of stable readings before updating the measured value then it should ignore a few milliseconds of instability as the pressure is released.
Would likely be a bit of a compromise between speed and stability though, and there might need to be a dead-band around open-circuit that would mean very small capacitors need to be measured without auto-hold.

RIGHT!
We could try to optimize the hold logic to get more accurate results.

But to be honest, there are many cases leading to the noisy result.
MLCC value would be influenced by vibration, temperature, even the excitation signal due to the piezoelectric effect
The impedance of any DUT that is too small would be noisy as well

[SilverSolder]

If the auto-hold function looks for a long enough period of stable readings before updating the measured value then it should ignore a few milliseconds of instability as the pressure is released.
Would likely be a bit of a compromise between speed and stability though, and there might need to be a dead-band around open-circuit that would mean very small capacitors need to be measured without auto-hold.

RIGHT!
We could try to optimize the hold logic to get more accurate results.

But to be honest, there are many cases leading to the noisy result.
MLCC value would be influenced by vibration, temperature, even the excitation signal due to the piezoelectric effect
The impedance of any DUT that is too small would be noisy as well

In addition to taking the MODE of the last few hundred ms of readings, you could calculate the standard deviation and indicate to the user (on the display) a little icon that indicates a "good" reading with a tight spread?   Always ignoring the last few ms of "letting go".    Additionally, during the measurement, generate a (quiet!) beep when a stable reading is obtained, so user knows when to let go the DUT... and then of course hold that value in the display?

This is the kind of stuff that would make your tweezers worth $100 to someone, over a cheaper one.

[Shannon]

If the auto-hold function looks for a long enough period of stable readings before updating the measured value then it should ignore a few milliseconds of instability as the pressure is released.
Would likely be a bit of a compromise between speed and stability though, and there might need to be a dead-band around open-circuit that would mean very small capacitors need to be measured without auto-hold.

RIGHT!
We could try to optimize the hold logic to get more accurate results.

But to be honest, there are many cases leading to the noisy result.
MLCC value would be influenced by vibration, temperature, even the excitation signal due to the piezoelectric effect
The impedance of any DUT that is too small would be noisy as well

In addition to taking the MODE of the last few hundred ms of readings, you could calculate the standard deviation and indicate to the user (on the display) a little icon that indicates a "good" reading with a tight spread?   Always ignoring the last few ms of "letting go".    Additionally, during the measurement, generate a (quiet!) beep when a stable reading is obtained, so user knows when to let go the DUT... and then of course hold that value in the display?

Thanks SilverSolder, it's a good idea!

We have integrated BEEP in the circuit which could inform the user "letting go" signal.
The threshold of standard deviation could be a reasonable value.

[Shannon]

Mechanical Design

I have used some SMD tweezers, although they are much more convenient than using a multimeter or Kelvin clip, but there is still a lot of gap between the handle a real tweezer, so we try to optimize mechanical design to improve the user experience. I've seen the mechanical design of the SMD tweezers on the market, such as directly soldered tweezers‘arm on PCB, hinge structure, etc., which are not reliable and easy to make SMD devices fly away.
Tweezers mechanical analysis:
The figure attached is an ordinary commercial tweezer. By analyzing the movement of tweezers tip in three axes, we can judge what precision tweezers are like

A criterion to judge whether a tweezer is easy to use is that the movement in the y-axis direction must be very small, which is the key to stability. When clamping the SMD device, there must be no shaking in the y-direction, otherwise, the device will fly off

[tautech]

A criterion to judge whether a tweezer is easy to use is that the movement in the y-axis direction must be very small, which is the key to stability. When clamping the SMD device, there must be no shaking in the y-direction, otherwise, the device will fly off

Correct however after some experience with SMD tweezers the correct 'feel' develops and components are then rarely lost. Hold any component too firmly and the risk of losing it is great.

Like any tool a little patience is required for your own skills to develop......tweezers are very different to pliers. 

[SilverSolder]

For sure, no sloppy hinges etc., that would definitely spoil it.  It doesn't have to be a surgical tool, but it should feel good to use... 

People have different sized hands, and different strength fingers, and different preferences with how hard they like to pinch the tweezers.  What if it used a separate spring between the two halves, and you could choose between a "stiff" and a "soft" spring?  The customer can then make it perfect for how they like to use it...   and it is an ergonomic selling point that nobody else has...

[Shannon]

A criterion to judge whether a tweezer is easy to use is that the movement in the y-axis direction must be very small, which is the key to stability. When clamping the SMD device, there must be no shaking in the y-direction, otherwise, the device will fly off

Correct however after some experience with SMD tweezers the correct 'feel' develops and components are then rarely lost. Hold any component too firmly and the risk of losing it is great.

Like any tool a little patience is required for your own skills to develop......tweezers are very different to pliers. 

If possible, we'd rather spend the same money on more stable tweezers rather than unstable ones, even for soldering
Making it easier to use is one of our goals

For the LCR tweezers, a better mechanical design may not cost too much to produce, but it needs more design effort and will make the user feel good.

[Shannon]

People have different sized hands, and different strength fingers, and different preferences with how hard they like to pinch the tweezers.  What if it used a separate spring between the two halves, and you could choose between a "stiff" and a "soft" spring?  The customer can then make it perfect for how they like to use it...   and it is an ergonomic selling point that nobody else has...

Thanks SilverSolder,
another good idea, we will try to add a spring, and to feel if it's different

By the way, if it is mass-produced, can we send you a pair of tweezers to thank you for your suggestions, because it has your contribution.

To be honest, one of the reasons we're publishing these ideas here is to make it better, and we will try to appreciate those help.

[indman]

Del

[S. Petrukhin]

People have different sized hands, and different strength fingers, and different preferences with how hard they like to pinch the tweezers.  What if it used a separate spring between the two halves, and you could choose between a "stiff" and a "soft" spring?  The customer can then make it perfect for how they like to use it...   and it is an ergonomic selling point that nobody else has...

Thanks SilverSolder,
another good idea, we will try to add a spring, and to feel if it's different

By the way, if it is mass-produced, can we send you a pair of tweezers to thank you for your suggestions, because it has your contribution.

To be honest, one of the reasons we're publishing these ideas here is to make it better, and we will try to appreciate those help.

If possible, send it to me for testing. Using different springs is an interesting idea, but changing them in a miniature device will not be very convenient, especially if it is a twisted coil spring. I think U-shaped springs from the plate are most applicable, such springs can be folded several pieces for reinforcement.

[indman]

Del

[Shannon]

By the way, if it is mass-produced, can we send you a pair of tweezers to thank you for your suggestions, because it has your contribution.

I would also be grateful for one tweezers you sent me to test. If possible.

Thanks indman and S. Petrukhin

If crowdfunding is successful, we can make mass-produced to be true, and we will send some to those who have a good idea to help us optimize the design for appreciation

[tautech]

By the way, if it is mass-produced, can we send you a pair of tweezers to thank you for your suggestions, because it has your contribution.

I would also be grateful for one tweezers you sent me to test. If possible.

Thanks indman and S. Petrukhin

If crowdfunding is successful, we will send some to those who have a good idea to help us optimize the design for appreciation

Did you get my PM a day or two ago ?

[Shannon]

Mechanical Design

A criterion to judge whether a tweezer is easy to use is that the movement in the y-axis direction must be very small, which is the key to stability. When clamping the SMD device, there must be no shaking in the y-direction, otherwise, the device will fly off

We used this criterion to test the performance of SMD tweezers and ordinary tweezers.
I tried to use my finger to separate the tip of the tweezers gently in the Y-axis direction (it's difficult to quantify the strength, but I try to be fair, sorry for there's no more accurate and convenient method). Attached is the test photo, and the results are as follows
Precision Tweezers > ESD tweezers > ST5S > DT71 > HP-990B

[Shannon]

Shannon, what frequency limit of the test signal can your tweezers provide?
Is it possible to raise it to 50-100 kHz?
And one more question - what resolution and accuracy can you provide for the secondary parameter "D"and "Q"?
You will have to compete with the rather successful and inexpensive MS8911 tweezers.

Currently, we focus on 10kHz, but we are still striving for the support of IC with higher performance, and we think there is a high possibility to expand excitation frequency to 100kHz in the future.

For the secondary parameter "D"and "Q", it's very hard to specified, the main parameter has a huge impact on the secondary parameter, and The range of main parameters is very wide, so it's hard to determine the boundary of the secondary parameter

[SilverSolder]

I don't know if this is just me, but rounded ends on the tweezers makes life much more difficult than pointed ends...  Maybe I'm poking around inside too much old junk, and actually taking core samples rather than measuring electric parameters? 

[tautech]

I don't know if this is just me, but rounded ends on the tweezers makes life much more difficult than pointed ends...  Maybe I'm poking around inside too much old junk, and actually taking core samples rather than measuring electric parameters? 

To suit todays miniscule SMD points need be some 20 thou (0.5mm) or finer and for rework some sharpness can be useful for in-circuit measurements if needing to get through oxides.

[thm_w]

What is the excitation voltage for Diode/LED mode? Will it be the same 1.0V AC signal, or do you have some extra hardware range to bump it up to battery voltage (~3V).
I would definitely pay $100 for these over the DT71's, from what you've posted here and in the other thread.

[Shannon]

Shannon,
You will have to compete with the rather successful and inexpensive MS8911 tweezers.

Ms8911 and our LCR tweezers are not at the same level, at least in terms of performance.

[Shannon]

What is the excitation voltage for Diode/LED mode? Will it be the same 1.0V AC signal, or do you have some extra hardware range to bump it up to battery voltage (~3V).
I would definitely pay $100 for these over the DT71's, from what you've posted here and in the other thread.

Hi thm_w,
Thanks for your support.

We use DC excitation for Diode/LED mode, and use pure AC signal for impedance measurement.
In theory, Diode/LED are non-linear devices, the relationship between current and voltage is exponential and R/L/C are linear devices, the current is proportional to the voltage.

So the test mode for Diode/LED is a constant current source which supply excitation current to these components, generally 1mA/2mA, and then measure the voltage drop on LED/Diode.

[SilverSolder]

For the diode test, it would be cool if it just "blinked" the diode long enough to make a measurement, instead of keeping it constantly lit and wasting battery.  As with the other components, it should just beep when it has the reading, and keep it on the display...