Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

I-V characterizer schematic

Status
Not open for further replies.

elektrovoz

Newbie level 5
Joined
Aug 17, 2017
Messages
9
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
60
Hi edaboard,

How I developing hardware for characterizing 2port devices, primary special kind of diodes. It needs to be accurate enough and works at high speed, more than 10Msamples/s. Please, review my schematic :)

View attachment iv_plotter.v1.pdf
 

If you have a specific question, I suggest you ask that rather than expecting someone to review your 6 page schematic with no description and no requirements.
 

Understandable. My hope was that somebody will quickly looks at the schematic and find errors at first glance.
 

Many 'scopes only have ~ 8 bit vertical resolution.
But if this is acceptable then a simple ramp generator,
sense resistor (maybe w/ log amp, depending on how
much low end detail you care about) and 2 probes,
dump the data across whatever back panel interface
the 'scope has (may be possible to control all of this
from Excel or Matlab or Labview) and you could skip
some serious hardware build-debug-rip_up cycles.
 
why is this better than scope in XY mode?
My apologize, but I have lost some details on the topic start. My "diode" is not very simple. It is a non-silicon hetero structure which need some kind of electrical training. I need to form fast train pulses with programmly specified form and measure I-V characteristic after that. Next step is the C-V measurement.

The idea is simple - waveform is recorded to memory and is played by DAC, forming device internal charge. After that I-V is measured using ramp voltage. There are two ADC channels - low side I->V converter and high side ADC connected using Kelvin probe.
 

Many 'scopes only have ~ 8 bit vertical resolution.
But if this is acceptable then a simple ramp generator,
sense resistor (maybe w/ log amp, depending on how
much low end detail you care about) and 2 probes,
dump the data across whatever back panel interface
the 'scope has (may be possible to control all of this
from Excel or Matlab or Labview) and you could skip
some serious hardware build-debug-rip_up cycles.

You are right, I can use standart laboratory equipment, but I don't have scope and generator with fast external control. And one more thing - I need to control overcurrent during I-V measure. I think standard equipment is not so fast to provide feedback from scope to disable generator in case of overcurrent.
 

The circuit won't work exactly as is. There are e.g. unreasonable capacitor values in OP feedback paths. I also wonder if the low generator output impedance is a good idea.
The amplifier circuit details should be checked in a simulation or may be adjusted empirically.
 
There are e.g. unreasonable capacitor values in OP feedback paths.
Thanks, I forgot to change.

The amplifier circuit details should be checked in a simulation or may be adjusted empirically.
OK, thanks.

I also wonder if the low generator output impedance is a good idea.
Hmmm... Why not? I need to form fast voltage pulses about 10V/10ns, so i need enough current to recharge all parasitic capacitance.
 

I would advise you use a CC sine source and fine tune offset DC. for both IV and CV tests.

Perhaps a Howland Converter works best up to 10MHz with high GBW OA.
as below
http://tinyurl.com/yc25349c


The advantage is inherent current and voltage limiting as well as phase can be measured with simple mixer vs f input.

Sine wave sweep gen, should not be an issue.

If you need 12 Bit resolution with high BW, how will you achieve better than 60dB SNR when le Croy and Agilent cannot achieve this in 12 bit high speed DSO's unless you average.
https://www.tek.com/blog/not-so-high-high-resolution
 
Last edited:
thus using my simple cct and sig gen, you can get results immediately now with any 8 bit DSO and then use XY mode with phase and get speed limit or C vs V bias and on/off R ratio for different pre-programmed parameters and then decide what if any enhancements you need with custom PCB.

If you. need higher current limits than an OA can give , using complementary emitter followers inside the loop output with smaller R values is possible or anything, if , you specify the interface, V,I,t constraints.

After all, you can scale 8 bits over any range you need, which ought to be sufficient for testing a binary memristor
 
Last edited:
Regarding output driver, the present circuit doesn't look well suited for driving capacitive loads. Did you notice that AD810 is a current feedback amplifier? Not sure if the intended variable gain circuit is stable at all.

Using a current source as suggest by SunnySkyguy, or a moderate (e.g. 10 to 100 ohms) driver impedance seems more appropriate to me.
 
Regarding output driver, the present circuit doesn't look well suited for driving capacitive loads. Did you notice that AD810 is a current feedback amplifier? Not sure if the intended variable gain circuit is stable at all.
Using voltage source and check the current is a necessity for me because I need electrical forming before measurement.
How to check amplifier stability? According to data-sheet it works well on 75Ohm cable... May be I need some kind of filter on output?
 

This sounds like a nonvolatile memory type test. If so
I think you probably need to step away from DC tests
(which "I-V" usually implies to people and work in the
pulsed time domain that represents your write cycle
timing goals (and lets you characterize things like
retention and wearout vs write cycle attributes).

You might not need as much measurement finesse
as you think. Sometimes "cal-mapping" an open loop
(or tight local closed loop) stimulus and response is
better for "getting it done" than developing the ability
to measure coincident Iin, Vin, Iout, Vout.

We do not know enough about the element under test
to say how simply a test jig could be done.

There could be interest in "fine structure" of the
programming (or whatever) event but this may be a
different interest than the broad based mapping of
programming pulse I, V, t to "bit strength" and "bit
degradation" - hardware, stimulus and recording /
analysis might differ a whole lot.

Some thinking about the test plan, what you will
actually analyze and respond to, might be good now.
A common desire to make "do-all" hardware exists,
but maybe elaborateness only puts you off the path.
 
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top