NI solution to measure Min 10nA@25V to Max 5mA@2000V

[OkiSumo]

Hi

i am looking for NI DAQ solution that is able to measure semiconductor devices with this leakage values Min 10nA@25V to Max 5mA@2000V. Total 700 DUTs at a fixed voltage, 60 seconds total cycle from 1st to 700th DUTs. Cost is a factor. Please advice how I can achieve to measure this leakage range. Thanks

[santo_13]

Multiple questions,

• Does it have to only measure current or supply power too?
  
  • Means, you have an external power supply but you need a better current measurement instrument alone?
• Are you looking for single-channel or multi-channel
  • Will affect your budget and measurement duration
  • You've 700 DUTs to be measured within 60seconds => you've to measure on average 85ms/DUT including any relay switching required
• Do you have a current shunt for each DUT?
  • Or you expect the DAQ to have the current shunt?
• What is the resolution and accuracy you're looking for in the measurement?
• You mentioned cost is a factor - but how much?
• Does the DAQ have to be USB or PCIe or PXI?

We can provide guidance and suggestion, still, if you reach out to NI Sales, they would be able to analyze your requirement and provide the best configuration.

Ideally, I would recommend you to go with PXI Analog Input modules - https://www.ni.com/pdf/product-flyers/pxi-analog-io.pdf 

https://www.ni.com/en-us/shop/category/voltage.html?productId=118726

Almost all high-end DAQ systems are PXI based, only the simpler general-purpose DAQs are USB based (there are some exceptions too)

[OkiSumo]

Thanks Santhosh

Multiple questions,

• Does it have to only measure current or supply power too?
  
  • Means, you have an external power supply but you need a better current measurement instrument alone?
  • Answer: Need to measure both the Vcc to the DUT (can be 1000V), and leakage of the DUT (10nA to 5mA range)
• Are you looking for single-channel or multi-channel
  • Will affect your budget and measurement duration
  • You've 700 DUTs to be measured within 60seconds => you've to measure on average 85ms/DUT including any relay switching required
  • Answer: Multi channels. Switches can be included
• Do you have a current shunt for each DUT?
  • Or you expect the DAQ to have the current shunt?
  • Answer: DAQ
• What is the resolution and accuracy you're looking for in the measurement?
  • Answer: 1%
• You mentioned cost is a factor - but how much?
  • Answer: More important is the solution at the moment.
• Does the DAQ have to be USB or PCIe or PXI?
  • Answer: PXI 
• I would recommend you to go with PXI Analog Input modules
  • Answer: We have checked and founds that NI DMM is able to measure the 10nA not AI. However, we loss speed like 1 sec per DUT.
  • Present HW list;

  • Item	Description	Qty
    1	784058-01 PXIe-1084 18-slot (17 hybrid slots), up to 4 GB/s PXI chassis	1
    2	763064-01 Power cord, 240V, 10A, U.K.	1
    3	779701-03 NI PXIe-PCIe8361, X1 MXI-Express for PXI express, 1 port, 3 m cable	1
    4	781348-01 PXIe-4353 32-channel, 24-bit PXI temperature input module	1
    5	781349-01 NI TB-4353 Isothermal terminal block for PXIe-4353	1
    6	783634-01 PXIe-6375 208 AI (16-Bit, 3.8 MS/s), 2 AO, 24 DIO, PXI multifunction I/O module	4
    7	192061-01 SHC68-68-EPM shielded cable, 68 D-Type to 68 VHDCI offset, 1 m	4
    8	191945-01 SHC68-68, Twisted pair cable with basic shielding, 1 m	12
    9	783130-01 PXIe-4081 7½-Digit, ±1,000 V, onboard 1.8 MS/s isolated digitizer, PXI digital multimeter	1
    10	780587-75 PXIe-2575 196‐Channel, 1-Wire PXI Multiplexer Switch Module	4
    11	779038-03 LFH200 to 4x50-pin Dsub switch cable (CH-CH Twisted) 1 m	4
    12	DAQ & RC module board	2

[santo_13]

Thanks for the quick reply. Looks like your application requires a detailed analysis based on top-level specifications.

In short, there is no straight solution for your requirement and a lot has to be taken into account and many compromises.

Your application deals with high voltage and low current. Typically for 1000V, DMM is the best choice, none other instrumentation (DAQ or AI) can measure such high voltage (of course with attenuator and safety mechanisms)

Good thing that you have an existing PXI solution, the ideal solution will be a switched system with multiple DMMs for high channel count parallel test.

This is a good solution for high channel count flexible resolution voltage/current measurement (with external current shunt) - https://www.ni.com/en/solutions/semiconductor/mixed-signal-ic-validation/power-validation.html

One possible solution for current measurement would be,

• Use current shunt per DUT on the low side so that the common-mode voltage is well within DAQ limits
  • Mindful of the burden voltage due to the shunt
• Ensure that the absolute maximum ratings for instrument inputs are maintained - add TVS diodes and protection devices

For the voltage measurement,

• Use the existing DMM, use the multiplexer to switch and measure each DUT
• Implement an accurate voltage divider or attenuator - connect to the DAQ AI

If you're fine, these two operations (measure voltage and measure current) could be separate functionalities, this way, the DMM has to switch to 700+ DUTs whereas the existing DAQ can switch DUTs in groups of 104 and happen in parallel.

I can think of many challenges that need to be addressed,

• Accuracy, resolution?
• Does it require range change, if so, auto-range change?
  • Hardware-based auto-range or software-based?
  • Fast range-change to measure within the 60s for 700DUT time
• Dynamic range of the ADC (AI) if there will be no range change
  • If you need to measure 10nA to 5mA in a single range, you need a 114dB Dynamic Range, which is quite high for a DAQ (typically 24-bit and higher)
• Calibration for circuitry outside instrument (current shunt) etc.,
• Relay/multiplexer rating & lifetime?

Thinking about all these, this feasibility & instrument selection is a large project by itself and cannot be just solved by forum members sharing suggestions.

[OkiSumo]

Thank you Santosh

One of the idea in improving sampling time is to increase the # of DMM and switches. This is huge cost factor. 

I have alternative solution using embedded circuit to perform all these tasks but it is always a closed system and proprietary which is hard to manage and to expand without paying more to the Integrator. The accuracy and calibration of such system is also questionable. 

Btw, the nA, uA, mA DUTs are not running at the same time, it depends on the device technology when we load it up. The leakage distribution is pretty tight per batch and per technology.  The bias voltage also varies per device technology, eg. 20V, 600V, 1000V. So I need solution that can do all of this. Maybe my limited by the cards that are available today.

Thanks again. It is a challenge.