No, despite my last thread, resistance sweep really isn't working correctly

[Euler's_Identity]

Oops!  Look at this capture instead, as the last one didn't clearly show the resistance being swept from 0 to 10k.

If I think of anything else I can capture to better demonstrate the problem, I will.  But for now that's all I can think...

Ah wait a minute! I'll give you the Excel captures, of the isource and vsource runs!

The second attachment is the isource run showing good data, and third is the vsource run showing bad data.  Note that I must use a vsource for this sweep.  I can't use an isource, because the current will not increase as necessary when I add the parallel resistors.  However, using the vsource produces bad numbers.    

[Euler's_Identity]

I think the key to solving this is figuring out why Multisim's changing the dc supply during the sweep.   Notice how the sweep ends at half the supply and, also, that a linear circuit is producing non-linear results.   I conclude that the supply is changing during the sweep for some reason, causing the final wiper position (100%) to tie the wiper to a supply that is 1/2 the supply of the circuit.  Yes?

Note: I have found that changing the supply to any other value causes the same result: the sweep still ends at 1/2 the supply.

More and more this is looking like a bug in Multisim's sweep function.

[lacy]

I have taken a look and I am not going to be able to come up with a solution for this. I do not use these analysis very much, so my experience level with them is minimal. I do see the same results that you are getting, which I know that doesn't solve the problem, but since 2 people are geting the same results is indication of a problem either with the software or the set-up that we both are using.

Maybe NI can read this and chime in and let both of us know if we are setting this up wrong or that there is a problem with the sweep function.

Sorry I could be of much help here, but I gave it my best shot.

[Euler's_Identity]

So how does one get NI to chime in?

[Euler's_Identity]

I just used my service contract's "ask an engineer" option.  But rather than go through this problem's description again in an email, I posted the URL to this thread.  Hopefully I can get this sweep issue resolved now. 

[OStep]

Hello Euler,

The potentiometer consists of two resistors, as you probably know. Parameter sweep only sweeps one variable at a time; thus you're only sweeping one of the resistors - r1:xr1 - which happens to be connected to ground in your circuit (you can see this if you export the netlist). As you sweep it from 0ohms to 10k, you're seeing the output V(vwiper) =1mA * swept resistance, and there's nothing wrong with those values (a 0Ohm resistor is actually limited to a very small value ~6uOhm). This is with respect to the circuit in the attached screen shot.

However, I would highly not recommend sweeping variables inside interactive components such as the potentiometer because it is hard to see exactly what it is that you're sweeping.

Let me know if I answered your question.

[Euler's_Identity]

Max,

I could be wrong, but the evidence says you did not read the thread as requested.  I have no problems doing what you just described; however, it must be done with a 1mA = 10v/10k current source.  Where you would be helping me is if you would show me how to do it with a 10v/(10k||R)=I current source, where R is unknown. I > 1mA if R<infinity, and THAT's where you'll find the problem arises clear as day.

What I'm saying (and I'm pressed for time right now, which is why this is short) is, in order to pull 1mA through the 10k potentiometer and also have a shunting unknown R, you must be able to draw more current from the voltage source.  In order to draw more current than 1mA from the voltage source, the voltage source must adapt to the lower load resistance by decreasing its source resistance.  Multisim's sweep isn't adapting to the lower load resistance.  Specifically, when a resistance sweep is made within a DC operating point analysis (one of the options available in the resistance sweep setup), the voltage source changes and causes inaccurate results.

True, if you use a 1mA current source and a 10k pot, you can sweep the pot and get accurate Vo results, but that isn't a suitable option for me, for I need shunting R.  10k(1mA) = 10V, but that 10V falls apart when you try to add any resistance in parallel (shunt) with the voltage source and sweep the pot.  Try it and you'll see what I mean.

[Euler's_Identity]

Max,

I tried to edit my previous post so it doesn't sound so rude -- I speak very matter of factly, and it often comes across as rude to people.  However, I couldn't edit it.  Mainly, I just wanted to be sure you understand the problem you're working on for me.  Telling you that what I saw showed me you didn't read the thread was probably not the best way to start the ball rolling, and I'm sorry for that.

Here's what I'm wanting, in a nutshell.

1.  Place a dc voltage source and a ground.

2.  Connect a potentiometer to ground across the dc voltage source.

3.  Assign the wiper of the pot the name Vo(wiper) (or something syntactically correct which identifies the wiper) and then sweep the pot producing many Vo(wiper) values that can be easily ported to Excel.  (No, I'm not asking for you to work with Excel, for your business is Multisim -- Multisim already does fine with porting the results to Excel.)

4.  Now place another arbitary-valued resistor across the wiper and tied to the dc voltage source and repeat steps 1-3.

You can see that step four requires the dc supply to supply more current to the load than is necessary when the load is only the 10k pot.  Hence, the source resistance will have to decrease to be able to source the heavier load brought on by the added shunting resistance of step four.  However, as the dc source is an ideal voltage source (no source resistance,) its output voltage should be remain constant throughout the sweep; I found that this is not the case with Multisim.  Instead, I found that, for some reason, Multisim only gives accurate Vo(wiper) values for a current source input to the pot.  But, because I must have shunting resistances, I cannot use a current source, for a 1mA current source into a 10k pot load will only keep the voltage drop across the load at 10v if there are no resistors shunting the pot.  

To summarize, as I see it, Multisim's resistance sweep, in conjunction with a dc operating point analysis that runs with each iteration of the sweep, will not work output accurate results using a dc voltage source, but, per Ohm's Law (not Multisim's fault in other words,) a current source cannot be used either unless the current source can increase to account for the heavier loading brought on by adding shunting resistance.  Incidentally, to make the current-source circuit work (to maintain the 10v across the pot) would require prior calculation of the loading resistance in order to set the current-source to produce the 10v drop (keep that 10v constant) across the load.  A 1mA current source will always create a 1mA current flow, whether the load is a 10k pot or a 10k pot with a shunting 10k resistance.  But you can see that, by adding that shunting resistance, the source voltage is halved, because the load resistance has been halved.

And now I have no more time.

[OStep]

Hi Euler,

I have followed your steps and found there to be nothing wrong with the results.

"However, as the dc source is an ideal voltage source (no source resistance,) its output voltage should be remain constant throughout the sweep; I found that this is not the case with Multisim"

-The Voltage source output is always the programmed value throughout the sweep. It is the case.

In the attached files, I have setup a circuit using 2 resistors equivalently representing the pot model and a circuit with the pot, both with a shunt resistor Rshunt. As you sweep the lower resistor in either circuit (r1:xr4 for the pot component) you get identical outputs, with expected values. For example, with resistor @ 5k, you have a resistor divider with v(vwiper)=10*(5k||Rshunt)/(5k+5k||Rshunt)) =~ 4.87V which is correct for Rshunt=100k

[Euler's_Identity]

Max,

You really got my hopes up as I read your latest post, but I tried your circuit after I made it work the way I've been asking it work, and it still failed to produce accurate results.

Try looking at it this way.  If you sweep the bottom of the pot, the wiper to ground, and your sweep ranges from 0 to 10k ohms, you're sweeping the full range of the pot.  Hence, Vo, as it is at the wiper, should end up at 10v, the dc source voltage.  Here's what I'm saying.  The sweep begins with Vo=0v, as the wiper is at ground, and then the sweep ends with the wiper at the dc source, which should be, still, 10v.

Your circuit, when placed under these conditions, fails to do this.  Both a capture of the circuit and results and, as well, your circuit producing those results are attached below. 

Notice that I changed a total of three things about your circuit:

1. I changed the initial setting of the pot to 0%.

2. I changed the output variable to monitor the wiper to ground.  (Therefore, it shows the initial value as near zero ohms, because of change #1.)

3. I changed the sweep to the full range of the pot, 0-10k. 

Did I do something wrong in any of the above three changes?