Questions for a project.

[Scigatt]

I'm planning to design a pcb for the Chua oscillator, a circuit which I have been interested in for a while. I'm completely new at this and I don't know much about electronics so before I go off to get parts and design the layout, I have a few questions about what components to use.

-Under operation the circuit would be intended to be hooked up to power and an oscilloscope, the user would turn some knobs, and observe the results. For the oscilloscope hookup, should I use a BNC connector, or just some pin headers for probes to hook onto?
-One of the voltages I'd like to measure is the floating voltage across a resistor. I'm planning to use an instrumentation amp IC to get this voltage, but if I just took the inputs across the resistor, I'd have to run a very low gain(<10 by voltage) to prevent the output from clipping. There is a considerable amount of common-mode signal(2-3 times the differential, voltage-wise) here, though, and from browsing relevant component datasheets(including the one I've settled on for now, INA217), CMRR seems to improve with higher gain. Should I divide the voltage across the resistor so I can use a higher gain, or just settle for the low gain CMRR?

[ebclr]

Scope have A-B function, this means you can measure  the voltage on the floating resistor using 2 channels and subtracting

[rstofer]

It would seem that, in normal operation, you would have 3 scope connections:  X, Y & Z.  For durability, I might be inclined to use 3 BNC connectors.  This does imply that your scope will be operating in X1 vs X10 mode.

It might be worth reviewing Dave's video on x1 probes and how they differ from coax.  There is much less of an issue with coax.

If I were going to install permanent scope probe locations, I would want to use real test point hardware.  Here are a number of alternatives.  Again, I want something where the probe doesn't fall off all the time:

http://www.mouser.com/search/Refine.aspx?Ntt=Keystone%20Test%20Points

I'm very interested in your project.  I hope you post something when you get it working.  Looking at the 3 differential equations, I'm tempted to integrate both side of all 3 and stuff the results into an analog computer.  My personal AC is too small but I was thinking that Simulink under Matlab might work.  Hm...

[rstofer]

And of course, the Matlab thing has already been done and by people far smarter than me.
http://www.chuacircuits.com/matlabsim.php

I haven't found an example with knobs and dials.  I like to be able to set parameters with what looks like a human interface.  It's much more satisfying than burying constants in the blocks or using a table of values.

[Scigatt]

Scope have A-B function, this means you can measure  the voltage on the floating resistor using 2 channels and subtracting

-One of the most important ways to view the outputs is in x-y mode.
-Some pairs of measured voltages do not share any potential at all.
-My oscilloscope has only two inputs.

Also, here's a vid I shot of the oscilloscope plot of a breadboarded circuit.


-I fix the dim trace about a minute into the vid.
-The cinematography, sadly, does not improve during the vid.
-Some of the erratic changes in the trace are due to my use of breadboard trimmers as pots. If and when I do make a functional pcb for this, I will use proper pots for these, probably with fine and coarse adjust.

Also, I'm a bit unsure as to how to use a BNC cable with an oscilloscope, whether to use a terminator or what end of the cable to put it on.

[rstofer]

I wouldn't think you would want or need a terminator unless your circuit is designed with 50 Ohms of output impedance.
That's a very nice demonstration.  I think I would be inclined to use 10 turn pots and dial indicating knobs.  That makes it easy to come back to known settings or at least get in the neighborhood.  The idea of using coarse and fine pots also seems like a good solution.

The main reason I am thinking about BNC is that I would want to box it up and have it laying around ready to play.

[Scigatt]

Well, I'm not planning to design for any particular box. Currently I plan to pcb mount everything.

[bson]

Well, since you can probe your circuit with an analog probe without affecting its behavior it can't be particularly sensitive, so INA offset and leakage currents likely don't matter.  So just scope out the peak voltages of the points that will form the amplifier inputs to make sure they don't exceed the maximum specs - common voltage, voltage to ground, or get too close to rail.  Give the amplifier enough gain to get say 4V peak to peak on the output.  If you use 50ohm source termination (which will need to be matched at the load - on the scope end) add 3dB to the amplifier gain.  I'd go with 50ohm BNC personally, it's much easier to work with and won't get knocked over or come loose and short something.  If your scope doesn't have 50ohm input termination on a switch just get a cheapo inline terminator on eBay.  You'll also need a 50ohm coax cable.

[tautech]

This thread may interest the OP:
https://www.eevblog.com/forum/projects/t20347/

[Scigatt]

Would it be advisable to buffer the output voltages not passing through the instrumentation amplifier?

[Scigatt]

Well, since you can probe your circuit with an analog probe without affecting its behavior it can't be particularly sensitive, so INA offset and leakage currents likely don't matter.  So just scope out the peak voltages of the points that will form the amplifier inputs to make sure they don't exceed the maximum specs - common voltage, voltage to ground, or get too close to rail.  Give the amplifier enough gain to get say 4V peak to peak on the output.  If you use 50ohm source termination (which will need to be matched at the load - on the scope end) add 3dB to the amplifier gain.  I'd go with 50ohm BNC personally, it's much easier to work with and won't get knocked over or come loose and short something.  If your scope doesn't have 50ohm input termination on a switch just get a cheapo inline terminator on eBay.  You'll also need a 50ohm coax cable.

Would a 50 ohm BNC connector that you would solder to the pcb come with that internal resistance or would you need a separate resistor?

[rstofer]

I would expect the resistor to be external.  If the output impedance is to be 50 Ohms, that means the TOTAL impedance which is usually the output impedance of the amplifier/buffer plus a series resistor to make up 50 Ohms.  The resistor could be soldered to the center pin of the connector if desired.

Read the last 5 paragraphs here:
http://www.eetimes.com/author.asp?section_id=36&doc_id=1225569

The driving op amp has to be capable of driving the output into what will effectively be 100 Ohms.  According to the article, the op amp will require additional bandwidth but I need to think about that.  I'm not sure why...

[tautech]

Not sure why but the OP has started another thread on his Chua oscillator project:
https://www.eevblog.com/forum/projects/chua-oscillator/