How can i make the output of the multiplier follow the output of the filter in the below circuit?

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
Hello,
I am supposed to design a circuit where i can determine the change in Variable capacitance Cx by a change in DC output voltage at the output by keeping the reference capacitance constant. I tried using a lot of different filters but the output of the filter does not follow the output of the multiplier I am using. I am posting my circuit file (created in LTSpice) along with the paper i am referring. Any help is much appreciated.

Thank you.
 

Attachments

radiohead

Joined May 28, 2009
508
It appears, from the paper, that the sensor is based on sand flow through a PVC pipe (maybe for use in a cement manufacturing plant). The sand flow will generate a static electricity on the PVC pipe. If the flow of sand is interrupted, the capacitance on the sensor will change. It would be used to measure how much (gravity-fed) sand is being delivered. The author has done the math to determine a two inch pipe can allow only so much sand to flow per second.
I cannot see your circuit as I do not have LTSpice on my computer. Can you post a PDF version?
What kind of sensor are you using? Are you trying to filter voltage or frequency?
 

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
The green waveform represents the filter output and the blue waveform represents the multiplier AD633 output. It is not following each other and also if the Cx is increased or decreased more from the Cref then i am having issues with the time of simulation as well.
Screenshot (90).png Screenshot (91).png
 

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
I am trying to convert AC to DC basically by varying Cx as you can see from the circuit picture. I also maintain an inverted version of the input wave and feed it to the comparator as i need a constant waveform to compare the variation. This is then multiplied by the multiplier and demodulated by the low pass filter. So I am supposed to get a sine wave and square wave at the input of the multiplier.
 

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
I do not see a similar circuit in that referenced paper.

Post a screen-dump picture of the schematic.
I have not used a transformer. I have a used a normal input and inverted the input signal in order to avoid using the transformer. That is the only change i have done to the circuit on the paper.
 

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
It appears, from the paper, that the sensor is based on sand flow through a PVC pipe (maybe for use in a cement manufacturing plant). The sand flow will generate a static electricity on the PVC pipe. If the flow of sand is interrupted, the capacitance on the sensor will change. It would be used to measure how much (gravity-fed) sand is being delivered. The author has done the math to determine a two inch pipe can allow only so much sand to flow per second.
I cannot see your circuit as I do not have LTSpice on my computer. Can you post a PDF version?
What kind of sensor are you using? Are you trying to filter voltage or frequency?
I am trying to convert AC to DC basically by varying Cx as you can see from the circuit picture. I also maintain an inverted version of the input wave and feed it to the comparator as i need a constant waveform to compare the variation. This is then multiplied by the multiplier and demodulated by the low pass filter. So I am supposed to get a sine wave and square wave at the input of the multiplier.
 

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
It appears, from the paper, that the sensor is based on sand flow through a PVC pipe (maybe for use in a cement manufacturing plant). The sand flow will generate a static electricity on the PVC pipe. If the flow of sand is interrupted, the capacitance on the sensor will change. It would be used to measure how much (gravity-fed) sand is being delivered. The author has done the math to determine a two inch pipe can allow only so much sand to flow per second.
I cannot see your circuit as I do not have LTSpice on my computer. Can you post a PDF version?
What kind of sensor are you using? Are you trying to filter voltage or frequency?
The green waveform represents the filter output and the blue waveform represents the multiplier AD633 output. It is not following each other and also if the Cx is increased or decreased more from the Cref then i am having issues with the time of simulation as well.
 

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
I do not see a similar circuit in that referenced paper.

Post a screen-dump picture of the schematic.
I have not used a transformer. I have a used a normal input and inverted the input signal in order to avoid using the transformer. That is the only change i have done to the circuit on the paper.Screenshot (90).png Screenshot (91).png
 

crutschow

Joined Mar 14, 2008
29,791
I missed the multiplier function in the reference schematic. :oops:

Here's my simulation of your circuit.
My only change was to rearrange the multiplier connections to correspond to the model symbol I had, and eliminate all the duplicate power supplies (node labels can really help simplify a schematic).
The filter output signal looks okay to me.

upload_2017-3-16_23-5-54.png

Don't understand why yours doesn't.
Where exactly where you looking at its output?
Your filter appears to be not functioning and seems to have a misconnection of some sort.
 

Attachments

Last edited:

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
I missed the multiplier function in the reference schematic. :oops:

Here's my simulation of your circuit.
My only change was to rearrange the multiplier connections to correspond to the model symbol I had, and eliminate all the duplicate power supplies (node labels can really help simplify a schematic).
The filter output signal looks okay to me.

View attachment 122651

Don't understand why yours doesn't.
Where exactly where you looking at its output?
Your filter appears to be not functioning and seems to have a misconnection of some sort.
The circuit works fine now. Thank you but there is still a little time delay. How can i eliminate it?
 

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
So what was the problem?

You need to be more specific with your questions.
Where are you seeing this time delay?
Sorry. I meant the output of the multiplier must follow the output of the filter. So when the waveform of the multiplier is high, the waveform of the filter should also be high. But here i am observing that it is the opposite and i need to rectify it. I tried a lot of trial and error methods but nothing really helped. Screenshot (92).png
 

crutschow

Joined Mar 14, 2008
29,791
So when the waveform of the multiplier is high, the waveform of the filter should also be high.
You are observing the normal delay of the filter.
Why do you need the filter output high when the multiplier output is high? :confused:
Aren't you just interested in the average DC value?
 

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
You are observing the normal delay of the filter.
Why do you need the filter output high when the multiplier output is high? :confused:
Aren't you just interested in the average DC value?
That is exactly what i told my professor but he is adamant on having the filter output follow the multiplier output.
 

crutschow

Joined Mar 14, 2008
29,791
That is exactly what i told my professor but he is adamant on having the filter output follow the multiplier output.
Then your professor doesn't understand how filters operate.
What is his training?

If you lower the output filter frequency, or add additional filtering then the output will be a constant DC with no significant ripple and he won't be able to have that complaint. (see attached).

upload_2017-3-17_13-26-40.png
 

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
Then your professor doesn't understand how filters operate.
What is his training?

If you lower the output filter frequency, or add additional filtering then the output will be a constant DC with no significant ripple and he won't be able to have that complaint. (see attached).

View attachment 122718
Thank you. I tried using AD734 multiplier as it has an error percentage of 0.1%. The multiplier connections are done as given in the data sheet but it seems that it is taking forever for simulation of the results. Any ways to speed up the process?
 

crutschow

Joined Mar 14, 2008
29,791
Any ways to speed up the process?
If you have a Measurement Timestep specified in the Transient Analysis options, remove it.

You could try going to the Tools/Control Panel/Spice and changing the Integration Method to one of the options as see if that helps.

You can also try reducing the precision of the various parameters (Gmin, etc.) by a factor of 10 to see if that makes a difference.
 

Thread Starter

Nanaiah

Joined Mar 16, 2017
25
If you have a Measurement Timestep specified in the Transient Analysis options, remove it.

You could try going to the Tools/Control Panel/Spice and changing the Integration Method to one of the options as see if that helps.

You can also try reducing the precision of the various parameters (Gmin, etc.) by a factor of 10 to see if that makes a difference.
Nothing seems to be working. The best i could get is this.Screenshot (93).png Screenshot (94).png
 

Attachments

Top