DSO & AWG Based Curve Tracer

[mawyatt]

While riding hurricane Ian out we decided to prepare a more versatile and complete Curve Tracer concept based upon a DSO and AWG, see Fool'n around thread.

https://www.eevblog.com/forum/testgear/fooln-around-with-dso-awg/msg4421209/#msg4421209

This concept worked so well thought maybe a setup fixture which included the ability to do A/B comparisons, deal with SMD as well as leaded DUT devices would be in order. Also, having a higher voltage/current range would be beneficial without the need for an additional buffer amplifier, and using a virtual ground transimpedance amplifier to "sense" the DUT current without creating an offset voltage.

With this in mind we set out to utilize a standard DSO and AWG to display and create the waveforms respectively, but also utilizing external PS inputs to power the fixture in an effort to KISS. Along this thinking we wanted to utilize some available cheap components (well, some components were available before the crisis!) to keep the cost down and adaptable to "what's" available.

Here's the preliminary schematic.

This utilizes a transimpedance amplifier (LM358 for U4) for DUT current sensing with scale factors of 1u, 1m, 1 amp/volt. The amp is buffered with a pair of TO-220 NPN/PNP (TIP122/TIP127). SBD D11 and D12 assure the input doesn't get out of hand above +- ~1V in fault conditions until PTC fuse FB9 opens, under normal closed loop conditions the input is forced to virtual ground by negative feedback formed by U4 and the senses scaling resistors shown with selection switch S/E (Source/Emitter).

For the DUT Base current drive a modified Howland Current Source/Sink is employed around U3 (also a LM358), resistors R21,20 & 15 and switch BASE set the current range from 1u, 1m, 1 amp/volt AWG staircased input shown as STEP IN. U3 is also buffered with NPN/PNP TIP41 & 42 for higher current source/sink capability as shown. Note the Howland Current Source/Sink can be converted into a simple voltage amplifier with switch I/V, this allows use with voltage controlled MOS devices.

For the Drain or Collector sweep (D/C) we decided to use a simple audio IC amp like the LM1875 or LM3886 setup with a voltage gain of 5 V/V. Switch OUT-Z and attached resistors allow the Sweep D/C Output Impedance to be set from 1, 1K and 100K Ohms.

Input power is nominally +-30V (+HV and -HV) as normally available from popular Lab Supplies, but can go higher (LM3886) if required. Standard 7815 and 7915 Linear regulators are employed to supply the regulated VCC and VEE voltages, duals +15V and -15V were selected to distribute the power dissipation rather than utilize a higher capacity single regulator (we had the standard 78 and 79 series in house).

The design is based to allow a upgrade to higher voltage ranges like +-150V for the Drain/Collector (D/C) sweep, by only replacing the Sweep Amplifier (PA441).

Anyway, here's a quick look at the project. We will follow up with additional details if folks are interested!!

Winds are picking up and heavy rain, Ian is getting near

Edit: Updated schematic.

Best,

[mawyatt]

The PCB is designed to fit within an available Aluminum 130mm 110mm 50mm case, same case type we utilized with the LCR Meter DC Bias Adapter. The output is by means of the 16 pin DUT connector shown in the above schematic.

For the actual test fixture a concept will utilized a top side PCB which will host a ZIP connector for leaded DUT, a pair of standard 3 pin connectors L-C/B/E & R-C/B/E which will allow wire & clip leads, and a pair of "Toggle Clamps" for SMD DUT types. A pair of Switches L/R-1 & L/R-2 allow A/B comparisons for all configurations and Jumper Setups for D/C (Drain/Collector), G/B (Gate/Base) and S/E (Source/Emitter) which allows routing of the 3 test signals for various DUT arrangements and pinouts.

The concept is designed to allow all sorts of either polarity DUT to be evaluated, from Bipolar, MOSFET, JFET, Diodes, Regulators, & LEDs, Triacs to name a few.

Best,

[RoGeorge]

Wow, looks great! 

It happens that I'm experimenting with something similar this week, an AWG buffer.  One of the intents is to use it for curve tracing, except it's not a dedicated tool, for now it's a rat's nest build.  Having fun with current-feedback, while using whatever parts happen to find through the scrap boxes.  No longer than this afternoon I was burning my fingertips to harvest bigger radiators from defective PC power supplies. 

I didn't know there's a hurricane there, take care.

[mawyatt]

Current feedback is a great type of amp, amazing almost constant BW regardless of gain

Here's what the Curve Tracer Concept should look like?

Best,

[RoGeorge]

For those who don't have an AWG for the curve tracer, the AWG can be replaced with a 555 timer. 

Not kidding, W2AEW made a video about such a circuit.  Very ingenious design, entirely analog:
#231: Circuit Fun: Stairstep generator using 555 and op amps  (schematic shown at minute 1:05)
then later he added a ramp generator with a 3rd opamp, like seen here at minute 4:05
#232: More Circuit Fun: Simple transistor curve tracer using Stairstep generator circuit

[RoGeorge]

Took a closer look this afternoon, and made a block diagram so to not get lost myself in all the switches and protections.  This is what I understood, and some questions:



- 1. the labels for the current range and the functional description in the OP says 1A, 1mA, 1uA, yet the corresponding resistors are 1M/10k/10 and 100k/10k/1.  I guess the lowest range is 10uA, not 1uA, is it?

- 2. the emitter (in a BJT the DUT) would be kept at a virtual ground by U4.1, but then it will measure Ie, while the curves are usually traced against Ic, not Ie.  For high Beta transistors Ic and Ie are about the same, but for power transistors beta can be as low as 10-20 or so.  I wonder, would it worth to subtract the Ib from the measured Ie (for example with a resistor from the Howland pump to the U4.2 opamp), so the Y axis would be proportional with Ic?

- 3. why the GND points at some pins of the DUT connector, is that for shielding in case of external probing wires?

- 4. just got an idea while typing this, would it be possible to use a normal ground instead of the virtual ground of U4.1 and the two Darlingtons? 

I mean, to tie R29, 31 and 22 to GND, and the other end will go to the emitter of the DUT.  A normal voltage amplifier opamp can read the voltage drop from R29/31/22.  The Howland pump can be floated instead, by referencing it to the DUT's emitter pin (the R10 now tied to ground, would be tied to the DUT's emitter).  In the end, the R29/31/22 voltage drop will be given by the Ic alone (without including the Ib), and the floating of a low current (the Howlan pump) should be easier then floating the emitter, which is a high current, no Darlingtons would be needed.  Don't know if the text description makes sense, will scribble a schematic.

[mawyatt]

1) Good catch, yes this should be 100ma and not 1a, just changed the text to 100ma. The ranges are in volts per amp, so 100ma across 10 ohms produces 1V. This is for Base current, so many amps should not be required!!

2) This is emitter or source current, and since the base current is from the Howland Source, one could adjust the emitter current to subtract the base current if desired. As shown the basic plot is Emitter Current vs Collector Voltage at Base Current Steps. Had thought about putting the current sense in the Collector but this would require the sensing circuit to move around with the collector sweep voltage and decided to use the transimpedance approach. One could certainly include the Collector side current sensing, but likely will be more involved with scaling and such and not as easy with simple components as this approach. This circuit should allow easy sensing of sub-microamp to amp currents without much trouble. If one wanted to get precise results then the a OP-07 or similar OP-Amp could be employed, but decided to just use the LM358 and realize the plot is Emitter current.

3) The Grounds are there for shielding the cable and DUT Test fixture if necessary.

4) Yes that would work, this is basically what we did with the original DSO & AWG setup. Use various resistor values to sense the emitter current with DSO Ch2 since this is ground referenced, but the collector voltage Vc sensing by the DSO Ch1 is also ground referenced and thus not collector to emitter voltage Vce, but actually collector plus emitter voltage. This works OK if the emitter current sense resistor is small (small emitter sense voltage) and thus Vc is ~ Vce. The virtual ground approach keeps the DSO Ch2 as Vce for the DSO XY plot, since the emitter is forced to virtual ground.

Anyway, thanks for spotting the test error and the dialog, we'll post the corrected schematic!!

Best,

[RoGeorge]

This is the DUT emitter-reference Ib and Vce described at point 4.  Though, the biggest DUT CE resistors are 100k, that might be big enough that the casual bias current of an average opamp will influence the measurement, not sure, I should look up the datasheet and/or simulate.



Looking into alternative arrangements for the DUT to settle what circuitry to add to my AWG current-feedback buffers, so to get from them, apart from other usages, the curve-tracing too.

I've stolen already your Howland pump and the idea to turn it into a voltage source for MOSFETs, that's neat! 

Just for the docs (for myself), adding two links about the Howland pump, one of which is from "The Lightning Empiricist" Vol.12 No.1 1964-Jan-01 from Philbrick Research, and has very clean and intuitive explanations:
Impeadance & Admitance Transformations using Operational Amplifiers - D.H. Sheingold

the other one is "sboa437", from Texas Instruments, ca.2020 colorized  :
Analysis of Improved Howland Current Pump Configurations - Ignacio Vazquez Lam

[mawyatt]

That should work

 The emitter sense amp could be scaled up with a single resistor from op-amp - to ground, this would help keep the sense voltage low if a good op amp is used (OP-07).

However the Howland setup may not work as intended.

With the - resistor returned to the emitter junction the supplied current will be:

  Ibase = (Vawg - Ve)/Re, with Ve as Ie/Re or (Vawg/Rsense)/[1+{Re/Rsense}(Beta +1)]

So if Re/Rsense is small then Ibase ~ Vawg/Rsense which is desired.

Also, with the switch open now the supplied voltage from the Howland source becomes 2*Vawg - Ve which is not desired, but if Ve is low then Vb ~ 2*Vawg. Noting that Vbase become Vgate for FET use, and Ve is Vsource.

This little analysis made me realize that using the other switch contact returned to ground give the result Vgate = Vawg which is what we want, and no need to mentally scale by 2!!

Anyway, interesting discussions and keep up the good work

Edit: Updated latest version of schematic to reflect mentioned change.

Best,

[mawyatt]

Here's a HV version capable for supplying +-200V for the D/C Sweep. It's based upon the available inexpensive ($20) APEX PA441.

Now to see if we can create a PCB the same size to fit, with the included internal HV +- supply. Input power is just ~ +-15VDC and the +-HV is created from the +15VDC Input.

Best,

[RoGeorge]

However the Howland setup may not work as intended.

Outch, I've implemented wrong the proposed change. 

The intent was to float the ground for the current source, and substitute its GND with a voltage potential copied from the emitter of the DUT.   I've missed to shift the AWG stairs voltage, too, with the same Ve.  Might require a voltage follower for the Ve, too, not sure right now, it's 5 minutes to AM hours here.  Feels like it should work but I'll think about it tomorrow, and redraw the schematic, then simulate it just to be sure.

Thank you for taking the time to point why it won't work correctly.

[moffy]

Just a thought, but I would make allowance for a small cap directly from U4.1 output to its inverting input to provide some more control over stability especially considering that a 1Meg resistor can feed back into the 100pf cap. I see that you have an RC feedback TBD but that might not be enough or in the wrong place.
Trying to get my head around your circuit, is it a bit like a network analyser? Sorry it is my first encounter with such a device

P.S. Have you SPICE modelled the transimpedance amplifier?

[RoGeorge]

is it a bit like a network analyser?

The circuit makes a Curve Tracer, together with a signal generator and an oscilloscope in XY mode, the schematic will plot how current varies with voltage for a tested component (DUT stands for Device Under Test, can be a transistor, diode, capacitor, resistor, etc.).  Displays the I/V curves for the tested components.

Something like shown/explained in this video by W2AEW:
#290: Vintage Tech: Tektronix 576 Curve Tracer

[moffy]

is it a bit like a network analyser?

The circuit makes a Curve Tracer, together with a signal generator and an oscilloscope in XY mode, the schematic will plot how current varies with voltage for a tested component (DUT stands for Device Under Test, can be a transistor, diode, capacitor, resistor, etc.).  Displays the I/V curves for the tested components.

Something like shown/explained in this video by W2AEW:
#290: Vintage Tech: Tektronix 576 Curve Tracer

Thank you, gotcha. So it does a series of DC voltage steps, or a repetitive linear ramp, measures the current and as you stated plots I/V.

[kripton2035]

suscribed.

[mawyatt]

Just a thought, but I would make allowance for a small cap directly from U4.1 output to its inverting input to provide some more control over stability especially considering that a 1Meg resistor can feed back into the 100pf cap. I see that you have an RC feedback TBD but that might not be enough or in the wrong place.
Trying to get my head around your circuit, is it a bit like a network analyser? Sorry it is my first encounter with such a device

P.S. Have you SPICE modelled the transimpedance amplifier?

Yes we have done some analysis with LTspice. The feedback TBD components should provide the ability to maintain stable operation under the expected conditions since the output buffer has ~1X gain and little additional delay relative to the slow op amp, good point tho and why we included those components!!

The buffer is designed to provide added current capability and the cross over should not be a factor since the waveform speed is slow and the added resistor (1K) from bases to emitters allows the op amp to drive the output current until the magnitude is above ~1ma, then the bipolar buffers begin supplementing the required load current.

As RoGeorge indicated, this is a device curve tracer for use with a DSO and AWG.

Will post some LTspice plots if you are interested.

Best,

[kripton2035]

I don't see the interest in using an external AWG for this. we only need a simple sinus voltage at different frequencies, this could easily be implemented in the
curve tracer itself ?

[moffy]

Yes we have done some analysis with LTspice. The feedback TBD components should provide the ability to maintain stable operation under the expected conditions since the output buffer has ~1X gain and little additional delay relative to the slow op amp, good point tho and why we included those components!!

The buffer is designed to provide added current capability and the cross over should not be a factor since the waveform speed is slow and the added resistor (1K) from bases to emitters allows the op amp to drive the output current until the magnitude is above ~1ma, then the bipolar buffers begin supplementing the required load current.

As RoGeorge indicated, this is a device curve tracer for use with a DSO and AWG.

Will post some LTspice plots if you are interested.

Best,

Did a quick sim myself, looks stable as long as you don't mind a limited BW, which really shouldn't be an issue but I would still provide a place for a cap as mentioned just in case. Real world can sometimes be a little more complex.
By the way would it be possible to sweep current and just measure voltage? That way you only have to produce the current once without the need to actively sink it. Not sure that would work with DIACs etc.

P.S. Not so good for reverse biased diodes either! Guess it's not such a good idea.

[mawyatt]

I don't see the interest in using an external AWG for this. we only need a simple sinus voltage at different frequencies, this could easily be implemented in the
curve tracer itself ?

How are you going to produce the Stair-Step Base current (for Bipolar) or Gate Voltage (MOS or JFET) for Transistors and Fets?

Sure you could design this and include it, but with the AWG you have the ability to create custom sweep types, ranges, offsets, then the Stair Step waveforms, number of steps, step level, polarity and so on. That's what they are good at, creating Arbitrary Waveforms, and with modern entry level AWGs you get two independent channels!!

Best,

[RoGeorge]

I wonder if it would be a good idea to simplify the current measurement by using the R_load also as a shunt to read Ic, like in the draft schematic (R8/R9/R10).

No protections have been included in this draft.  Y axis in the captured plot is 10mA/V.

[mawyatt]

Did a quick sim myself, looks stable as long as you don't mind a limited BW, which really shouldn't be an issue but I would still provide a place for a cap as mentioned just in case. Real world can sometimes be a little more complex.
By the way would it be possible to sweep current and just measure voltage? That way you only have to produce the current once without the need to actively sink it. Not sure that would work with DIACs etc.

P.S. Not so good for reverse biased diodes either! Guess it's not such a good idea.

You basically have a Voltage and Current Source you can use as desired, so you could use the Howland Current Source to sweep if desired, and use either the AWG drive, or the Output for the DSO Ch1. We tried to create lots of flexibility for various uses.

JFETs can be traced by Sweeping the D/C and monitoring the S/E current and using the G/B Current Source switched to a Voltage Source with a a negative StairStep waveform for the Gate Voltage, same for DMOS. TRIACs and DIACs should also be available to be measured.

Thought it was interesting in this thread where we were able to show the negative incremental resistance of the reversed biased BE junction of a ordinary 2N3904 NPN in breakdown. Confirmed the incremental negative resistance by adding a simple shunt capacitor which created a relaxation oscillator!!

https://www.eevblog.com/forum/testgear/fooln-around-with-dso-awg/msg4422982/#msg4422982

Best,

[mawyatt]

Here's a HV version capable for supplying +-200V for the D/C Sweep. It's based upon the available inexpensive ($20) APEX PA441.

Now to see if we can create a PCB the same size to fit, with the included internal HV +- supply. Input power is just ~ +-15VDC and the +-HV is created from the +15VDC Input.

Best,

Updated version with PCB layout. HV driver can be a PA441 or a OPA462.

Best,

[mawyatt]

I wonder if it would be a good idea to simplify the current measurement by using the R_load also as a shunt to read Ic, like in the draft schematic (R8/R9/R10).

No protections have been included in this draft.  Y axis in the captured plot is 10mA/V.

Yes, you can do this (use R Shunt) and it should work fine. Certainly simpler than the Transimpedance approach. May require a better differential amp than a LM358 creates though. If you can keep the impedance very high then it may be possible to get to very low current levels which is something we were interested in, with the Transimpedance approach this is possible without much effort (or expense) as the limits are just the Op-Amp bias current as diode leakage which should be very low. The input protection SBD have ~ 0 volts across them under normal operation, so leakage is low and the clamp diodes at the op-amp input aren't even necessary, so they could be eliminated all together since the op-amp input can only "see" the SBD diode drop thru 1K worst case. Output leakage isn't an issue either unless it finds it's way to the input, thru the switch or other paths, but this should be manageable.

Anyway, many ways are possible and keep up the great work and dialog!!

BTW nice setup & plots with Spice!!!

Best,

[RoGeorge]

Glad you liked the plots, thank you. 

If you can keep the impedance very high then it may be possible to get to very low current levels which is something we were interested in, with the Transimpedance approach this is possible without much effort

That's a strong argument.

I would be very interested, too, in measuring small leakage currents, especially since the same schematic can also be used for DC measurements.  Now the HV opamp makes even more sense.  First that came to mind would be to measure capacitors' leakage at DC.  Or the dark current in a photodiode.  Or any other application for a TIA.

Wow, I'm fully sold to the idea now, will add the TIA back, that's a great feature, thanks for the idea! 

[mawyatt]

Glad you liked the plots, thank you. 

If you can keep the impedance very high then it may be possible to get to very low current levels which is something we were interested in, with the Transimpedance approach this is possible without much effort

That's a strong argument.

I would be very interested, too, in measuring small leakage currents, especially since the same schematic can also be used for DC measurements.  Now the HV opamp makes even more sense.  First that came to mind would be to measure capacitors' leakage at DC.  Or the dark current in a photodiode.  Or any other application for a TIA.

Wow, I'm fully sold to the idea now, will add the TIA back, that's a great feature, thanks for the idea! 

This discussion has created some "thinking out loud" so to speak. Wonder if a better overall approach would be a pair of Howland Sources, each configurable as a bipolar Current or Voltage source and range scaling. One would be a lower voltage type but higher current, say to >|1a|, the other a high voltage but low current type <|50ma|. Add to this the transimpedance amp and you have a very flexible test function capable of covering a wide range of 2 and 3 terminal components over a wide range of currents and voltages!!

Best,